presentation of key frame

What is Keyframing? Let's Explore the Basics of Keyframe Animation

Keyframing can add energy and dynamic movement to your videos so that every element feels fast, snappy, and polished. Here's how to get started.

Keyframes are the starting or ending frames in a video that mark a change in appearance, position, or state. For example, the frame in which an object in the video begins to rotate is a keyframe.

Keyframe animation is the use of keyframes to move, change, or animate parts of your video by picking starting and ending positions or properties. For example, if you have an object scroll from the right side of your video to the left, that's keyframe animation in action.

With keyframing, you can add energy and dynamic movement to your videos so that every element—from intros to transitions—feels fast, snappy, and polished. You can also highlight important footage, animate text , or apply motion to visuals to make videos visually engaging while staying informative.

Put simply, keyframing is an end to boring, static videos. That's why we're so excited to launch our new Keyframe Animation Editor here at Kapwing.

Keyframing matters because engagement is critically important for modern marketing videos. And with keyframing, you can edit engaging videos that earn more engaged viewers. Let's dig in a bit more and explore what you can do with keyframing.

Benefits of keyframing

As we covered, keyframing allows you to set specific points in your video where you want movement or changes to occur. You can adjust the position, size, and opacity of elements in your video, and even add in text or images. The result is a video that moves and flows in a way that captures your audience's attention.

So how can you use keyframing to level up your videos? Here are just a few examples:

1. Highlight important information

Use Keyframing to zoom in on specific text or images in your video, drawing your viewer's attention to the most important information.

2. Create seamless transitions

Keyframing allows you to smoothly transition between different elements in your video, creating a cohesive and professional-looking final product.

3. Add movement and energy

Whether it's a subtle shake or a dramatic zoom, Keyframing can add movement and energy to your video, making it more engaging and visually interesting.

We've found that adding new overlays, transitions, or motion effects every 10-15 seconds to videos can increase watch time—one of the most important factors for the YouTube algorithm and other social platforms dominated by short-form video.

Types of keyframes

Keyframes have a number of uses, but it's helpful to know the standard keyframe interpolations available. Interpolations determine the rate of change between the keyframes you've marked.

Put another way, interpolations determine how an object will move or change between the keyframes you've set in your video. There are three standard keyframe interpolations:

Linear keyframes describe a uniform movement between keyframes that keep a consistent pace or motion between the start and finish—there's no acceleration or deceleration between keyframes. Linear interpolations are the most basic form of keyframing available and a good place to get started if you're new to keyframe animation.

A subset of linear interpolation is a continuous interpolation , which allows movement to multiple points without a stop and start in between.

Hold keyframes keep the same value for an object or attribute until the next keyframe. Since the hold keyframe keeps objects in place, it's more commonly known as a freeze frame.

Ease keyframes are what most people think of when hearing "motion effects." Ease keyframes allow for motion with acceleration and weight, which helps them feel more natural and adds a flair that's harder to achieve with linear keyframes.

Ease in and ease out are two common applications of ease keyframes. The former lets an object start with faster acceleration and slow down once it closes in on the ending keyframe, and the latter is the exact opposite—acceleration is slow in the beginning but speeds up closer to the ending keyframe.

Ways you can use keyframes

The number of basic keyframe interpolation belies the number of cool things you can do with keyframe animation! Now, let's cover some of the most common ways keyframes show up in video production.

1. Slow and fast zoom

Fast zooms are commonly used to add energy and emphasis to specific parts of a video, while slow zooms help build up excitement or lead into a reveal. Zooms are relatively easy to apply and involve setting two different scales between the starting and ending keyframes and then choosing your transition speed—this is an effect where you'll often use the ease keyframe.

2. Motion tracking

Motion tracking , which allows the "pinning" of an element to a specific space within your video, is a great way to communicate information to your viewers or make an edit that stays consistent throughout a scene, even when an object is moving. Common uses of motion tracking include:

• Pinning text to moving objects to label people, places, and things
• Pinning visuals and effects to objects in videos, like blurring someone's face

3. Animate text

Subtitles are at their best when auto-applied to your videos for accessibility and to let viewers watch your videos with the sound off. But that's not the only way to apply text in your videos: animating text by way of keyframing is great for scroll-stopping information that draws a viewer's attention or gives personality and style to slower parts of a video, like a transition or section title.

4. Moving effects

Effects that you apply to videos can also be keyframed to allow them to move across sections of your video. For example, you could keyframe a particular filter or effect, such as lighting effects, so that they move with a subject or a focal point in your video.

And that's just a sample of what's possible with keyframing. Check out Kapwing's keyframing features to learn even more about what's possible—without having to download any software.

Keyframing FAQ

What does keyframing mean.

Keyframing is the most straightforward way to animate an element in a video. Keyframing works by choosing specific starting and ending frames ("keyframes") and then choosing your interpolation, or how things will change or move between keyframes. Common forms of keyframes include Linear keyframes, Hold keyframes, and Ease keyframes.

What is an example of a keyframe?

A keyframe is the individual "frame" in a video where you change or apply an effect. For example, if you introduced pinned text to label someone who was on screen, the moment the text appears and then disappears would be the keyframes.

How do you keyframe something?

You can keyframe any video in Kapwing without downloading software. Here's how to keyframe:

• Select the layer you want to animate in the timeline within Kapwing.
• Move the seeker along the timeline to the point where you want to start your animation.
• Click on the diamond icon next to "Position" in the right sidebar to add a keyframe to that point on the layer.
• Move the seeker to the point where you want to continue your animation.
• Drag the layer to a new position on the canvas and release to add a keyframe.
• Continue adding keyframes and adjusting the layer position as needed to create the desired animation.

What Are the Best Fonts for Video Subtitles? Here’s a Research-Backed Answer.

How to Quickly Fix and Enhance the Audio in Any Video

How to Create a Video with Pictures

Video Game Design and Development

Encouragement, advice, and support for aspiring game designers.

• How to Start Keyframing in Animation

As you dip your toes in the industry, you will come across words like “keyframes,” which seemingly everyone talks about like they are immediately understandable.

Every video has a keyframe.

What exactly are they, and why are they so important to animators and video editors?

What Are Keyframes?

Every single shot in a video, whether live or animated, consists of individual frames. A keyframe is the most crucial part of each frame, as it defines the start and end points. It sets a precedent for all other transitional frames and indicates what the other frames would look like.

To put it simply, you can use keyframes to create dynamic videos and highlight essential elements. They make the production feel much more energized and look more polished and engaging.

The Origin of Keyframes

Keyframes were associated with animation long before digital video editing technology was developed. For decades, animators used techniques like cel animation or stop-motion to create videos.

Hand-drawing cartoons are rather tedious and time-consuming. So, to save time, lead artists drew the “keyframes” or essential frames and left the transitional or less important frames for junior artists.

Keyframes in the Digital Age

Even with the most cutting-edge digital animation tools , lead artists still often just complete the keyframe and go back and forth to create the in-between shots.

And keyframes have become a critical element of video editing, as they define the shots to follow; it still dictate the start and endpoints. In the digital age, modern-day video editing is simpler than ever, and creating transitional frames is much different from what it used to be.

What Is Keyframe Animation?

Keyframe animation is the transition between two crucial frames. Think of it as the effect you apply between slides when creating a PowerPoint presentation.

This change in keyframes can be automated through presets in most editing software or applications, or they can be manually added to create a seamless transition.

Some of the common techniques include changes in magnification or scale, framing (moving the image within the frame), rotating the image, the color of the image, fading in or fading out, or a change to the audio.

What Are the Benefits of Keyframing?

As mentioned earlier, keyframing makes videos flow in a way that grabs the audience’s attention and makes the final product much more engaging.

Some of the biggest benefits of keyframing include:

Highlights Important Content

Zooming in on essential text, objects, or characters is a great way to highlight important information.

Creates Seamless Transitions

It ensures smooth transitions between frames, objects, and texts, creating a professional and cohesive final product.

Keyframing polishes the animation, adds energy, and makes it look fresh and exciting. The dynamic transitions, motion effects , and overlays make the video much more fun to watch—something that is critical to a project’s success.

Moreover, the freedom to change effects every few seconds or for any duration you would like allows for a lot of creative flexibility.

Types of Keyframes

There are three standard keyframes:

This is a good place to start if you’re new to keyframing. Linear keyframing refers to a uniform or consistent pace of transitions between keyframes. There is neither an increase nor a decrease in speed.

Continuous keyframe, a linear sub-category, is also quite common. As the name suggests, this calls for constant interpolations or transitions without missing a beat.

This is also known as the freeze frame, a popular effect seen in many old-school movies. As many may know, this keeps an object still until the next keyframe.

Easy frames are what most refer to as “motion effects.” They add flair to the final product by allowing motion acceleration and creativity, unlike linear keyframes.

The two common options are ease-in and ease-out. Ease-in involves an object starting with faster acceleration and slowing down once it reaches the ending keyframe. Ease-out is the opposite of this.

Characteristics of Keyframes

Note that there are different names for keyframing, so you may have to look for these functions depending on the software you are using.

When looking for a keyframe in a video editing application, check out the “Frame View” or “Action View” option in the control panel.

A bar with many “objects” in the timeline indicates a keyframe. Depending on the object you have added, you can customize transitions.

Some of the characteristics include:

All of these objects are highly customizable to achieve your desired results. You can change background colors, place objects in different places, change orientation, increase or decrease opacity, size up the objects, and even add text.

This explains why keyframing is such an essential aspect of video editing and animation.

Keyframing Animation Techniques

Let’s take a closer look at some classic keyframe effects:

Slow and Fast Zoom

The process of zooming in or out is pretty much the same. The only thing that would change would be the effect’s time between the keyframes. Most experts recommend placing the second keyframe at least five seconds before the scene.

Using “Anchor Point” will dictate where the rotation should begin and end. Playing around with the numeric values in the software will help you understand it so you can achieve the exact effect you want.

You can also customize the title’s position, opacity, and rotation—there’s a lot of scope for creativity and fun.

Tracking Motion

When looking at tracking motion, a box will appear on the screen on the frame. You can put any object that needs to be tracked and adjust it accordingly. Then, move the tracker and object to where you want it in the timeline.

Tracking motion can be added to titles, video clips, images, or even an effect (like blur, mosaic, or spotlight). Your title or effect will follow the object you tracked.

Remember, keyframes are automatically created when tracking motion. Double-clicking on the timeline’s newly added tracked object will help you see the scale and position keyframes created.

Reveal Mask

Select the frame and where the title will mask the screen. Then, drag the tool to where the mask must split to reveal the frame. You can choose Object Settings to move the title anywhere on the masked area to hide or reveal the title as well. The options are endless.

Mask Object

Most editing software have a masking designer with its primary tools. You can choose a predesigned mask or customize one to suit the object you wish to mask. Play around with different styles to create the exact effect you envision.

The actual process is different depending on the software, so giving yourself enough wiggle room and time to practice masking objects in various ways will help you in all your future projects.

Moving Lens Flare

With the adjustment tool, the duration, timeline, and frame can be selected, rendering a moving lens flare effect that looks seamless. Experimenting with and adjusting the parameters will create different results every time.

Animated Light Rays

Click on the keyframe icon to adjust the duration and the timeframe for the light rays.

Can Keyframing Be Automated?

Keyframing can be a time-consuming affair. After all, getting the shot right and perfectly timing it to the frame that you want is no easy feat. A lot of technicality is involved, and some effects take a lot of practice to get right.

You can certainly automate keyframes with editing software, but it’s not that simple. The steps to automate keyframing may be quite complex, depending on the software.

Even then, you still need to specify the start and finish of the animation. These can be used in future projects as well. In some applications, these are known as animation presets.

How To Use Keyframing Animation Presets to Videos?

Presets are a list of existing effects, animations, and keyframes you can use in projects. These can be quickly implemented in a project, with a few personalized touches.

Applying keyframe animation presets is fairly straightforward. Some video editing software have these ready—all you have to do is apply them to the timeframe and make small modifications for a polished look.

Select the layer, frame, or property you want to animate to apply a preset to an existing video. Then click on the preset in the “Effects” and “Presets” panel; this will generally be under “Tools.”

You can also drag and drop the preset onto the layer or the timeline, but that would differ from software to software.

Finally, apply the keyframe preset and make small changes in settings to suit your needs and help you achieve the desired finish.

How To Create Personalized Keyframe Animations?

Most video editing software offer preset design templates and effects. All you have to do is adjust the duration of the effect within your frame. However, in some cases, a customized or personalized keyframe could be the solution to make a project interesting and unique.

Some of the software or applications that you can use to create personalized keyframe animation include:

• PowerDirector
• ColorDirector
• Director Suite

So, as you now know, the keyframe is the main scene or frame of a video, while keyframing animation is the transition between two important frames. A keyframe defines the “start” and “end” points.

The concept was first used by old-school cartoonists, and to this day, it is used in modern video editing software. Video editing software can seamlessly integrate transitions or keyframing animations in between frames for a professional and highly engaging product.

There are many types of keyframe animations, each providing highly personalized options to make your project better. Objects, lights, text, opacity, background—everything can be fine-tuned for a cohesive output.

We hope this article has helped you understand keyframes, keyframe animation, and all that it entails.

https://www.fullharbor.com/blog/keyframe-essentials

Popular Posts

About Dustin Tyler

Initially focused on web dev, Dustin was introduced to game design by a friend after college and was immediately attracted to the combination of technical skill and creativity required to make an awesome game.

He’s utilized his web dev skills and desire to help others by creating a resource where expert game designers can share what they’ve learned with the next generation of aspiring game makers.

Leave a Reply

Your email address will not be published. Required fields are marked *

Recent Posts

• How to Use Animation Storyboards
• What is Cel Shading & How to Use It?
• The Role of Computer Science in Video Games
• Cite this Website
• California Consumer Privacy Act
• Editorial Guidelines

• Game Design Education
• Game Design Careers
• Game Engines
• Learn Skills
• Ask an Expert
• Graphic Design
• Free Mini-Courses
• Our GameMaker Course

Send us mail:

Lake House Media 7910 4TH ST N STE 300 ST. PETERSBURG, FL 33702

Slider Revolution

More than just a WordPress slider

What Is a Keyframe in Animation? Explained Simply

Discover what a keyframe in animation is, its role in creating fluid motion graphics, and how to master keyframe techniques for stunning results..

Picture this: You’ve got a character ready to burst into life, but you need a blueprint to guide every movement. Enter the  keyframe , the cornerstone of  animation .

It’s the anchor point that defines pivotal moments in your  animation timeline , whether it’s the dramatic arch of a jump or the subtle raise of an eyebrow.

In this article, we’re diving deep into the essence of keyframes, exploring not just their definition but also their critical role in crafting compelling  motion graphics  and fluid  character animation .

By the end, you’ll grasp how to harness  keyframing processes  and utilize advanced tools like the  graph editor  and  control points .

Prepare to unlock the secrets behind seamless  frame sequencing , understand the magic of  tweening , and discover why mastering keyframes is vital for any animator. Ready to transform your animations from static to spectacular?

What Is a Keyframe in Animation?

A keyframe in animation is a specific frame that defines the starting and ending points of any smooth transition. It marks significant moments in the movement, with the in-between frames, or “tweens,” filled in to create fluid motion.

Keyframe Fundamentals

Basic concepts, keyframe characteristics.

Ah, keyframes—the heartbeat of animation, marking pivotal action points. Set in stone on the animation timeline, they hold the essence of motion. Just imagine them as the skeleton of your animated masterpiece.

They capture significant changes: the jump of a character, a sudden flash of light, the twist of an object. You move from one keyframe to another, painting a sequence that transitions seamlessly.

Yet, they aren’t just about change. They control the rhythm and pace, making the difference between jittery chaos and fluid perfection.

Difference Between Frames and Keyframes

Frames vs. keyframes—think of it like flipping through a photo album versus watching a movie.

Frames: every single snapshot. Hundreds, thousands, flashing by, forming the illusion of motion.

Keyframes: the stars of the show. They highlight the critical moments, the change-makers. Regular frames smooth out the path between them, filling in the gaps.

Keyframes are about intent and impact—they signal, “Look here, something important happens!”

Types of Keyframes

Linear interpolation keyframes.

Straight lines, predictable paths. That’s linear interpolation keyframes for you. They get you from point A to B in an unembellished, straightforward manner.

No fancy easing, no drama, just a steady, linear journey. Perfect for when you need consistency—a ball rolling, a character walking at a constant speed.

Bézier Interpolation Keyframes

Now, let’s crank up the elegance. Bézier interpolation keyframes are the maestros of motion. They use curves, not lines, adding a touch of finesse to transitions.

Control handles—oh, the power they offer! You can mold and shape your animation path, creating smooth, dynamic movements. Perfect for breathing life into sophisticated sequences.

Hold Interpolation Keyframes

Freeze! Hold interpolation keyframes command attention. They keep the frame steady until the next keyframe steps in. It’s like a dramatic pause in a speech, letting a moment resonate before moving on.

Great for abrupt changes—think sudden stops, intense gazes, sharp shifts in position. It’s all about making moments stand still, if only for a bit.

Keyframe Properties

Start here, end there. Position keyframes dictate where objects go. The cornerstone of motion paths, they move elements across the scene.

From a flying superhero to a bouncing ball—you plot positions, the animation follows suit. It’s the essence of kinetic movement.

Big to small, small to big—scale keyframes handle the growth and shrinkage.

Animating a character stretching? Zooming in on an object? Scale keyframes make it happen. They adjust size dynamically, adding depth and dimension.

Twist, turn, spin. Rotation keyframes bring objects into three-dimensional delight.

Whether it’s a subtle head tilt or a dramatic 360-degree spin, they manage angles and orientation. Keep it lively, keep it animated.

Fade in, fade out—opacity keyframes play with visibility. Perfect for transitions or ghostly appearances.

Control transparency, create mood shifts, bring in elements gradually. Opacity sets the tone, the drama, the punch. A master stroke for subtlety.

Application of Keyframes in Animation

Traditional animation techniques, hand-drawn animation.

Imagine an artist’s studio, paper and pencils scattered about. The rustle of flipping pages fills the air as each sketch morphs into the next. 

Hand-drawn animation —there’s something inherently raw, tactile about it. It’s where keyframes found their genesis.

Artists crafted each  keyframe  with painstaking detail, the cornerstone moments in motion. It was an act of patience, a dance between frame and vision. And then… the magic.

In-Between Frames

Enter the unsung heroes: in-between frames. These guys fill the gaps, flesh out the journey between keyframes. They breathe continuity, ensuring transitions are smooth.

In the traditional realm, junior animators or “in-betweeners” handled the task, carefully mapping out each incremental movement. They bridged keyframes with precision, making sure every flick of a character’s tail, every flutter of wings looked seamless.

Modern Animation Software

Keyframing in computer animation.

Now, let’s jump into the future, into the world of pixels and code.  Keyframes  never really changed, did they? Just evolved, adapted.

In computer animation, keyframing becomes a high-tech wizardry. Tools like  Blender  and  Maya  sweep in, offering endless possibilities. Create a keyframe at point A, another at point B, and voilà—algorithmic magic crafts the in-betweens.

The principles remain, but the process? Streamlined, efficient. No more paper cuts.

Keyframe Animation Tools

Modern animation requires robust tools. We’re talking  SVGator , where you can animate SVG graphics with precision, every keyframe meticulously plotted. Vectors come alive—responsive, scalable.

Or  PowerDirector , a dynamo for video editing. Its keyframe capabilities? Beyond superb. Position, scale, opacity, rotation—manipulate them all with finesse. Whether it’s tweaking a title animation or orchestrating an intense zoom-in, keyframes are your best friends.

From those quaint studios of hand-drawn art to today’s digital realms, keyframes remain irreplaceable. Perhaps now, you have a clearer picture of what is a keyframe in animation.

Practical Uses of Keyframes in Video Editing

Keyframes in video transitions.

Imagine a scene, slowly coming alive, drawing viewers in with deliberate intent. That’s the  slow zoom .

Keyframes, placed strategically, allow for this gradual approach. The camera glides inward, creating tension, anticipation—the kind you feel when something epic is about to unfold.

Now flip it. The  fast zoom . It’s dynamic, full of energy. In a split-second, you’re thrown into the heart of the action.

Keyframes make it happen in a blink. From establishing a scene to landing you into the core, it’s an adrenaline-packed ride. Think action films, that sudden jolt—yeah, that’s the keyframes doing their magic.

Motion and Effects

Tracking motion.

Objects on the move—they need tracking. With keyframes, you can lock onto a subject.

Be it a runner dashing through the streets or a car speeding down a highway,  tracking motion  keeps your focus glued. The subject stays centered, crisp, in every frame.

Rotation Effects

Spin it around!  Rotation effects  take ordinary scenes and twist them, quite literally.

Picture a title spiraling into view, or an object turning on its axis. Keyframes handle each degree of rotation, making it fluid, hypnotic. It’s like adding a choreographed dance move to your visual story.

Title Animations (Zoom, Pop, Reveal Mask)

Titles—don’t just make them appear, make them  arrive .

Zoom : Start small, grow larger, dominate the frame.

Pop : A burst into existence, catching eyes with exaggerated enthusiasm.

Reveal Mask : The elegant entry—letters revealed as if unveiled from the shadows.

Keyframes choreograph every aspect, converting static text into a dramatic entrance.

Lens Flare and Light Rays

Shooting stars aren’t just for dreams.  Lens flare  and  light rays  can add a celestial touch.

Subtle at times, blinding at others—keyframes dictate the flare’s journey across the screen. From sci-fi to sun-soaked vistas, they dial up the drama.

Masking Objects

Sometimes, elements need a disguise.  Masking objects  with keyframes offers a creative cloak.

Let something hide, reveal itself, or blend seamlessly into the background. It’s like animation wizardry. Keyframes define the boundaries, the entrances and exits.

Design visually attractive and high-performing websites without writing a line of code

WoW your clients by creating innovative and response-boosting websites fast with no coding experience. Slider Revolution makes it possible for you to have a rush of clients coming to you for trendy website designs.

Creating Smooth Transitions with Keyframes

Setting up keyframes.

Marking Initial and Final States

First things first—plot your journey. Pin down your starting point, the initial state. Then, decide where the magic ends, the final state. It’s like mapping a quest. This is where you sketch out the trajectory, the path your object will take.

No fluff. No guesswork. Crisp and clear.

Defining Animation Duration

Time to set the clock—how long does the journey take? Define your animation duration. Is it a quick dart across the screen or a slow, graceful glide?

Milliseconds matter . Keyframes don’t just guide action; they choreograph time itself. Feel the rhythm, set the pace.

Keyframe Interpolation

Understanding interpolation.

Now, let’s dive deeper—interpolation. It’s the secret sauce. The bridge between  keyframes . Think of it as filling in the gaps, crafting the flow.

Linear for straightforward, no-nonsense motion.

Bézier for those curvy, nuanced dance moves.

Applying Easing Effects

Time to add some flair—easing effects make transitions smoother. Imagine a car slowing down to a gentle stop or speeding up gradually.

Ease in ,  ease out . Create that natural ebb and flow, the subtle deceleration or acceleration that breathes life into animation. Sharp edges—gone. Everything becomes liquid.

Advanced Techniques

Reusing keyframes.

Efficiency is key—why reinvent the wheel? Reusing keyframes is a trick up any animator’s sleeve.

Got a particularly stunning rotation or a perfect scaling effect?  Duplicate .  Reapply . Achieve consistency and save time. It’s about working smart.

Adjusting Timing and Position

Now, refine, tweak, polish. Adjust the timing and position for that impeccable finish. Shift keyframes minutely, stretch or compress durations.

Nuance matters .

Millimeters, milliseconds—adjustments that transform good into brilliant.

Advantages of Keyframe Animation

Efficiency in animation, speeding up the process.

Speed. It’s the holy grail of animation. With keyframes, you’re not redrawing every frame. Place your keyframes, let the software handle the in-betweens. Bam. Done.

Efficiency isn’t just a buzzword—it’s life-saving. With keyframes, an epic scene that would’ve taken weeks is now manageable in days. It’s like having a turbo button for creativity.

Creating Smooth Transitions

Transitions that glide, not jerk. That’s the keyframe promise. Each keyframe meticulously placed, the journey from point A to B becomes buttery smooth.

Imagine the fluid turn of a spaceship, the elegant arc of a dancer’s leap. Keyframes string these moments together like pearls, each transition a seamless marvel. Smooth isn’t just nice—it’s necessary.

Flexibility and Precision

Ease of making changes.

Tweaks? No biggie. With keyframes, alterations aren’t a nightmare. Adjust a position here, a rotation there. It’s modular, flexible.

Need to change the path of a character’s jump? Shift a few keyframes. The rest? They fall in line. What’s magic is how undemanding it feels. Flexibility on tap.

Reusability of Keyframes

Build once, use forever. That’s the joy of reusable keyframes. Your perfect walk cycle? Save it. Your ideal zoom transition? Save it.

Animation isn’t born from scratch every time. Professional software like  Blender  and  Adobe Animate  offer libraries where keyframes can be reused endlessly. Efficient. Smart. Genius.

Limitations of Keyframe Animation

Time-consuming setup.

Ah, the  curse of the clock . Setting up keyframes can be a laborious task, a digital labyrinth. Each keyframe meticulously plotted, each movement calculated with a mathematical precision.

Imagine plotting a character’s arc through a bustling cityscape or choreographing a complex dance sequence. Time slips away.

Keyframing says, “Patience is a virtue,” but the ticking clock can test even the zen of animators. It’s a necessary evil, but a time-sink nonetheless.

Complexity of Realistic Movements

Realism —the holy grail that’s maddeningly elusive. To create movements that mimic the intricate dance of reality, you need more than just talent. You need a profound understanding of physics, biology, and motion.

Imagine replicating the delicate flutter of a bird’s wings or the flowing motion of human hair in the wind.

Keyframes can create lifelike fluid dynamics, but the complexity can often be overwhelming. Each nuanced movement requires dozens of keyframes, each coordinate choreographed to perfection.

Simple, it is not.

Managing Multiple Keyframes

Enter the tangled web—managing multiple keyframes. It’s like juggling chainsaws on a unicycle while blindfolded.

One wrong move, and your animation goes from flawless to fiasco.

Adjust one keyframe, and you could inadvertently affect twenty others. Keeping track of where everything is, ensuring each keyframe is perfectly aligned and timed, can be a digital nightmare. Complexity multiplies as layers build up: position, scale, rotation, opacity—each with its set of keyframes.

Answering the question of what is a keyframe in animation, while fascinating, also shines a light on these limitations the animator wrestles with.

The beauty and beast of keyframe animation lie in this balance.

FAQ On Keyframes In Animation

What exactly is a keyframe in animation.

A  keyframe  is a specific point in the  animation timeline  where a particular position, shape, or parameter of an object is defined. It sets the start and end points for any action or transformation, enabling smooth  frame sequencing  and creating the foundation for  tweening .

How do keyframes work within animation software?

In  animation software  like Adobe After Effects, Blender, or Toon Boom Harmony,  keyframes  are plotted on the  timeline  to dictate the movement or transformation of elements.

The software interpolates between these frames, generating intermediate  in-between frames  that make the transition appear smooth.

What is the difference between keyframes and tweening?

Keyframes  represent the primary points of significant change, such as the start or end of a motion.  Tweening  fills in the gaps between these points, generating the transitional frames.

While keyframes set the essential poses or states, tweening ensures seamless movement between them.

Why are keyframes crucial in character animation?

In  character animation , keyframes define pivotal poses in a character’s movement cycle, allowing animators to focus on critical moments.

This technique outlines the  character animation  process, setting up essential positions for actions like walking, jumping, or talking, and making it easier to animate fluidly.

How do I set up keyframes in digital animation?

To set up  keyframes  in digital animation, place your object in its initial position on the  animation timeline  and mark a keyframe.

Move to a later frame, change the object’s position or property, and set another keyframe. The software will interpolate the in-between states automatically.

What tools assist in managing keyframes?

Tools like the  graph editor  and  control points  assist in managing  keyframes . The graph editor visualizes animations, allowing for precise adjustments of properties. Control points help fine-tune movements, making your  frame sequencing  more accurate and versatile.

What is the role of keyframe interpolation in animation?

Keyframe interpolation  involves calculating the transitional frames between keyframes. This technique uses various algorithms to determine how an object moves between two points, ensuring smooth and realistic motion.

Proper interpolation is crucial for maintaining the natural flow of  motion graphics .

Can keyframes be used in motion graphics?

Absolutely!  Keyframes  are foundational in creating  motion graphics . By setting specific  key poses  or transformations at distinct points on the  timeline , you can control animations, allowing elements to move, scale, or rotate precisely as needed, significantly impacting the overall visual storytelling.

What are the principles of using keyframes in animation?

The principles involve setting distinct  key poses , understanding the timing and spacing (central  animation principles ), and using tools like  graph editors  for precise control.

Adopting these methods ensures that your  digital animation  remains smooth and natural, enhancing the viewer’s experience.

How does frame rate interact with keyframes?

Frame rate  determines how many frames are displayed per second, directly affecting the  animation timeline  and the placement of  keyframes .

A higher frame rate allows for more detailed transitions and smoother motion, whereas a lower frame rate reduces the number of possible in-between frames.

In grasping  what is a keyframe in animation , we’ve unlocked the power to bring static images to life. Keyframes act as the pivotal moments within your  animation timeline , setting the stage for seamless  frame sequencing  and dynamic  motion graphics .

By mastering tools like the  graph editor  and understanding  keyframe interpolation , you can control every nuance of your character’s movement and the flow of your scenes.

Remember,  keyframes  are not just technical placements but the heartbeats of your  digital animation .

They signify critical points, ensuring your vision translates into fluid, compelling motion.

From setting initial  key poses  to achieving complex  character animation , every frame and transition contributes to the overarching narrative and visual impact.

In sum, delving deeply into the  keyframing processes  equips you with the skills to make your animations not only technically sound but also visually enchanting.

Keep experimenting, and let each keyframe narrate your story with precision and artistry.

FREE: Your Go-To Guide For Creating Awe-Inspiring Websites

Get a complete grip on all aspects of web designing to build high-converting and creativity-oozing websites. Access our list of high-quality articles and elevate your skills.

Slider Revolution high priest on the mission to find the line between not enough coffee and just a little too much coffee. Same with beer.

For any inquiries or additional resources related to this blog post or else, please don't hesitate to comment below or email me at [email protected] .

Liked this Post? Please Share it!

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

Save my name, email, and website in this browser for the next time I comment.

From The Blog

Good color combinations for website design, innovative css masonry examples to inspire your layouts, the ultimate guide to css animations, popular resources, optimizing load speed and performance, quick setup – slider revolution, create a basic responsive slider, get productive fast.

Join over 35.000 others on the Slider Revolution email list to get access to the latest news and exclusive content.

Privacy Overview

• Editing Techniques

What are Keyframes in Animation — Origins and Modern Uses

• What is Animation
• Types of Animation
• What is Pre-Production in Animation
• What Does an Animator Do
• What is Cel Animation
• What is Pixilation Animation
• What is Tweening in Animation
• What is Anime
• What is Claymation
• What are Keyframe
• What is Persistence of Vision
• What is Stop Motion Animation
• What is Rotoscope Animation
• What is Character Design
• What is Vector Animation
• How to Direct an Animated Movie
• Animation Storyboard Tips
• Unreleased Aladdin Animatic
• Best Animated Music Videos
• Best Animation Software
• Best Animated Movies of All Time
• Best Hayao Miyazaki Movies
• Pixar Storytelling Formula
• Best Pixar Features
• Best Pixar Shorts
• Best Stop Motion Movies
• Best Rankin Bass Christmas Movies

H ave you ever wondered what exactly keyframes are? How about the difference between keyframes in animation vs. keyframes in video editing? The two are different but might be more interconnected than you realize. In this post, we’ll be covering everything you need to know about keyframes.

Watch:  The Ultimate Guide to Animation

Subscribe for more filmmaking videos like this.

What are Keyframes in Animation

First, let’s define keyframe.

If you encounter any other unfamiliar terms in the wild, our ultimate guide to filmmaking terminology is a great resource for looking up definitions.

KEYFRAME DEFINITION

What is a keyframe.

A keyframe , also written as “key frame,” is something that defines the starting and/or ending point of any smooth transition. That something can be a drawing in animation or a particular frame of a shot when dealing with film or video. Any shot, animated or live-action, is broken down into individual frames. You can think of keyframes as the most important frames of a shot that set the parameters for the other frames and indicate the changes that will occur throughout as transitions. For more information, refer to our filmmaker’s guide to frame rates.

Keyframe Characteristics:

• Important individual frames from within a shot
• Keyframes exist in animation and live-action
• Sets a start/stop point for a transition

What is a Key Frame in Animation

The origin of keyframes.

These days, the word keyframes is often associated with video editing, but they originated in animation long before digital video editing. However, if you are familiar with keyframes in the context of editing, then you might already have a pretty good idea as to what is a keyframe in animation.

How to layout and keyframe in animation  •  Keyframes definition

In traditional animation, each frame is drawn by hand. To learn more about different animation techniques, check out our articles about cel animation , rotoscope animation , and stop-motion animation .

Because of the heavy demands and time-consuming nature of animation, those films are typically made by a number of artists working together. One efficient way to both save time and ensure quality is to have the lead animators draw the most important frames and leave the transitional frames between them to the junior animators.

Animators react and discuss keyframes  •  Keyframes example

These important frames drawn by the lead animators became known as keyframes. The transitional frames that connected the various keyframes together become known as in-betweens.

Even if a single artist is drawing an entire scene, it is still common practice to begin by drawing the keyframes, then going back and adding the in-betweens.

What is the Difference Between Frame and Keyframe

Keyframes in the modern age.

Even though animated films are rarely drawn entirely by hand these days, keyframes are still used in computer animation. Keyframes remain just as important today as ever, it is only the method of creation that has changed.

Keyframing with modern technology  •  Keyframing animation

Keyframes no longer apply to just animation, however. Since the advent of non-linear editing software and video editing apps , keyframes have become an integral element of standard video editing. The literal definition is still the same — a frame that designates the start/end of a transition — but it now manifests differently and can apply to new contexts.

What are keyframes in editing?  •  Keyframes example

As used in the context of video editing, keyframes might set the parameters for motion, they might be used to fine-tune a video transition . Or they might control timed adjustments made to effects applied to a video.

The good thing about using modern technology when working with keyframes is that the editing software can fill in all of those pesky in-betweens automatically, saving you a great deal of time and effort.

What is Cel Animation?

So, those are the basics of keyframes in both animation and video editing. But, there is a lot more to learn about the world of animation. One of the most tried and true methods of traditional animation was known as cel animation, but what does that mean? Learn everything you need to know about cel animation.

Up Next: Cel Animation →

Showcase your vision with elegant shot lists and storyboards..

Create robust and customizable shot lists. Upload images to make storyboards and slideshows.

Learn More ➜

Leave a comment

Your email address will not be published. Required fields are marked *

• Pricing & Plans
• Product Updates
• Featured On
• StudioBinder Partners
• Ultimate Guide to Call Sheets
• How to Break Down a Script (with FREE Script Breakdown Sheet)
• The Only Shot List Template You Need — with Free Download
• Managing Your Film Budget Cashflow & PO Log (Free Template)
• A Better Film Crew List Template Booking Sheet
• Best Storyboard Softwares (with free Storyboard Templates)
• Movie Magic Scheduling
• Gorilla Software
• Storyboard That

A visual medium requires visual methods. Master the art of visual storytelling with our FREE video series on directing and filmmaking techniques.

We’re in a golden age of TV writing and development. More and more people are flocking to the small screen to find daily entertainment. So how can you break put from the pack and get your idea onto the small screen? We’re here to help.

• Making It: From Pre-Production to Screen
• What is a Femme Fatale — Definition, Characteristics, Examples
• What is Method Acting — 3 Different Types Explained
• How to Make a Mood Board — A Step-by-Step Guide
• What is a Mood Board — Definition, Examples & How They Work
• How to Make a Better Shooting Schedule with a Stripboard
• 1 Pinterest

An Interactive Guide to Keyframe Animations

Introduction.

CSS keyframe animations are awesome . They're one of the most powerful, versatile tools in CSS, and we can use them for all sorts of nifty things.

But they're also often misunderstood. They're a bit quirky, and if you don't understand those quirks, using them can be quite frustrating.

In this tutorial, we're diving deep into CSS keyframes. We'll figure out how they work, and see how to build some pretty swanky animations with them. ✨

Link to this heading Syntax

The main idea with a CSS keyframe animation is that it'll interpolate between different chunks of CSS.

For example, here we define a keyframe animation that will smoothly ramp an element's horizontal position from -100% to 0% :

Each @keyframes statement needs a name! In this case, we've chosen to name it slide-in . You can think of this like a global variable. *

Keyframe animations are meant to be general and reusable. We can apply them to specific selectors with the animation property:

As with the transition property, animation requires a duration. Here we've said that the animation should last 1 second (1000ms).

The browser will interpolate the declarations within our from and to blocks, over the duration specified. This happens immediately, as soon as the property is set.

We can animate multiple properties in the same animation declaration. Here's a fancier example that changes multiple properties:

Link to this heading Timing functions

In “ An Interactive Guide to CSS Transitions ”, we learned all about the different timing functions built into CSS.

We have access to the same library of timing functions for our keyframe animations. And, like with transition , the default value is ease .

We can override it with the animation-timing-function property:

Link to this heading Looped animations

By default, keyframe animations will only run once, but we can control this with the animation-iteration-count property:

It's somewhat rare to specify an integer like this, but there is one special value that comes in handy: infinite .

For example, we can use it to create a loading spinner:

Note that for spinners, we generally want to use a linear timing function so that the motion is constant (though this is somewhat subjective—try changing it and see what you think!).

Link to this heading Multi-step animations

In addition to the from and to keywords, we can also use percentages. This allows us to add more than 2 steps:

The percentages refer to the progress through the animation. from is really just syntactic sugar ? for 0% . And to is sugar for 100% .

Importantly, the timing function applies to each step . We don't get a single ease for the entire animation.

In this playground, both spinners complete 1 full rotation in 2 seconds. But multi-step-spin breaks it into 4 distinct steps, and each step has the timing function applied:

Unfortunately, we can't control this behaviour using CSS keyframe animations, though it is configurable using the Web Animations API. If you find yourself in a situation where the step-by-step easing is problematic, I'd suggest checking it out !

Link to this heading Alternating animations

Let's suppose that we want an element to "breathe", inflating and deflating.

We could set it up as a 3-step animation:

It spends the first half of the duration growing to be 1.5x its default size. Once it reaches that peak, it spends the second half shrinking back down to 1x.

This works, but there's a more-elegant way to accomplish the same effect. We can use the animation-direction property:

animation-direction controls the order of the sequence. The default value is normal , going from 0% to 100% over the course of the specified duration.

We can also set it to reverse . This will play the animation backwards, going from 100% to 0%.

The interesting part, though, is that we can set it to alternate , which ping-pongs between normal and reverse on subsequent iterations.

Instead of having 1 big animation that grows and shrinks, we set our animation to grow, and then reverse it on the next iteration, causing it to shrink.

Link to this heading Shorthand values

We've picked up a lot of animation properties in this lesson, and it's been a lot of typing!

Fortunately, as with transition , we can use the animation shorthand to combine all of these properties.

The above animation can be rewritten:

Here's a piece of good news, as well: the order doesn't matter. For the most part, you can toss these properties in any order you want. You don't need to memorize a specific sequence.

There is an exception: animation-delay , a property we'll talk more about shortly, needs to come after the duration, since both properties take the same value type (milliseconds/seconds).

For this reason, I prefer to exclude delay from the shorthand:

Link to this heading Fill Modes

Probably the most confusing aspect of keyframe animations is fill modes . They're the biggest obstacle on our path towards keyframe confidence.

Let's start with a problem.

We want our element to fade out. The animation itself works fine, but when it's over, the element pops back into existence:

If we were to graph the element's opacity over time, it would look something like this:

Why does the element jump back to full visibility? Well, the declarations in the from and to blocks only apply while the animation is running .

After 1000ms has elapsed, the animation packs itself up and hits the road. The declarations in the to block dissipate, leaving our element with whatever CSS declarations have been defined elsewhere. Since we haven't set opacity for this element anywhere else, it snaps back to its default value ( 1 ).

One way to solve this is to add an opacity declaration to the .box selector:

While the animation is running, the declarations in the @keyframes statement overrule the opacity declaration in the .box selector. Once the animation wraps up, though, that declaration kicks in and keeps the box hidden.

So, we can update our CSS so that the element's properties match the to block, but is that really the best way?

Link to this heading Filling forwards

Instead of relying on fallback declarations, let's consider another approach, using animation-fill-mode :

animation-fill-mode lets us persist the final value from the animation, forwards in time .

"forwards" is a very confusing name, but hopefully seeing it on this graph makes it a bit clearer!

When the animation ends, animation-fill-mode: forwards will copy/paste the declarations in the final block, persisting them forwards in time.

Link to this heading Filling backwards

We don't always want our animations to start immediately! As with transition , we can specify a delay, with the animation-delay property.

Unfortunately, we run into a similar issue:

For that first half-second, the element is fully visible!

The CSS in the from and to blocks is only applied while the animation is running. Frustratingly, the animation-delay period doesn't count. So for that first half-second, it's as if the CSS in the from block doesn't exist.

animation-fill-mode has another value that can help us here: backwards . This will apply the CSS from the first block backwards in time .

“Forwards” and “backwards” are confusing values, but here's an analogy that might help: imagine if we had recorded the user's session from the moment the page loaded. We could scrub forwards and backwards in the video. We can scrub backwards, before the animation has started, or forwards, after the animation has ended.

What if we want to persist the animation forwards and backwards? We can use a third value, both , which persists in both directions:

Personally, I wish that both was the default value. It's so much more intuitive! Though it can make it a bit harder to understand where a particular CSS value has been set.

Like all of the animation properties we're discussing, it can be tossed into the animation shorthand salad:

Link to this heading Dynamic animations with CSS variables

Keyframe animations are cool enough on their own, but when we mix them with CSS variables (AKA CSS custom properties), things get ⚡️ next-level ⚡️.

Let's suppose that we want to create a bouncing-ball animation, using everything we've learned in this lesson:

CSS animations are meant to be generic and reusable, but this animation will always cause an element to bounce by 20px. Wouldn't it be neat if different elements could supply different "bounce heights"?

With CSS variables, we can do exactly that:

Our @keyframes animation has been updated so that instead of bouncing to -20px , it accesses the value of the --bounce-offset property. And since that property has a different value in each box, they each bounce to different amounts.

This strategy allows us to create reusable, customizable keyframe animations. Think of it like props to a React component!

Link to this heading Just the beginning

As I was building the last couple demos, I realized just how much CSS has evolved in the past few years!

It's become an incredible language , expressive and flexible and powerful. I love writing CSS.

And yet, so many front-end developers have a very different relationship with the language. I've spoken to hundreds of JavaScript developers who find CSS frustrating and confusing. Sometimes, the exact same CSS will behave totally differently! It feels so inconsistent.

I have a theory about this: unlike with JS, so much of CSS is implicit and behind-the-scenes. It's not enough to know the properties ; you need to know the principles driving them.

I've spent the last year working full-time on a course that will help teach CSS at a deeper, more fundamental level. If you found this blog post helpful, you'll love the course.

It's called CSS for JavaScript Developers , and it's just been released to the public. You can learn more at css-for-js.dev .

Last Updated

September 28th, 2021

A front-end web development newsletter that sparks joy

My goal with this blog is to create helpful content for front-end web devs, and my newsletter is no different! I'll let you know when I publish new content, and I'll even share exclusive newsletter-only content now and then. No spam, unsubscribe at any time.

If you're a human, please ignore this field.

Animation basics: 5 pro tips for using keyframes

Learn how to get the most out of your keyframes with these useful tips from Rob Redman.

Animation is a task that lends itself to minimalism but by this I'm not talking about the action in your scene (although simple can be best there too) but in the keyframing of your action. It's very easy to add keys to everything and you will soon be faced with a timeline that looks intimidating at best and unusable at worst.

I'm going to show you a few ways you can work with your keyframes to minimise the pain and to allow you to work efficiently and effectively, letting you concentrate on the action in the scene.

01. Its all in the rig

Keyframe animation can be a time consuming and intricate task. Make life easier for yourself by ensuring you are working with clean, easily understood rigs.

My robot model has a few simple controls added to the viewport that are only visible when he is selected. This means I don't need to trawl through hierarchies to find what I need and can concentrate on only the parts I'm animating at any given time.

The little circle to the left of each label means I can add keyframes directly as well, so I don't need to lose focus on the actual motion.

02. Be choosy about what you keyframe

Most software will have a default set of parameters that will be keyframed when you hit the record button. These are usually position, rotation and scale, and quite often a point level or parameter option too. On top of that many apps will have auto record turned on. While this can help you get up and running it really isn't good practice.

Try to deactivate recording anything that you don't choose yourself. Putting yourself in the driver's seat will enable you to understand your timeline better and will reduce unwanted clutter. It will also make it a whole lot easier to go back and make adjustments should you need to (and you will).

03. Blocking out

The first step is to block out your animation. This is best thought of as as start, middle and end position, although it can be far more complex depending on your needs. Try to keep details to a minimum here and concentrate on getting the overall pacing right before you move on to other things.

Start with just the overall position and orientation. You can add in limb movement and hip sway later. You can see in the timeline that I have a full two second animation set up with keyframes at the start and end that are identical. This makes step 4 more logical.

04. Duplicating keyframes

Once you have the basics in place you can use them to create loops. In my example you can see my robot is doing a little dance where he moves in a circle with a little twirl at the end. All I've keyed here is some simple position and rotation for the body, head and arms. I've actually duplicated some of them, specially the keys on frame 49, so I don't get any odd jerkiness at each loop point.

I can now select all those keys and Ctrl-click and drag to generate a fresh loop. I can then click on the name of each item and add in a little variation for a more natural look. By grabbing the end handle of any selected set of frames you can change the speed of that section, with a smooth interpolation.

05. Keep a handle on the curves

While the keyframe view is great for organisation the true power is in the curve editor, so swap over to that (the space bar in C4D) and all your keys will be there with colour coded curves and each keyframe represented by a black dot with handles. These handles are where you fine tune things and add some personality to your animation. There are a few preset styles of interpolation from easy ease, to ease in or ease out but really you want to be taking manual control.

Curves work in a logical way, moving along the timeline so if you want to ramp up your motion you would take the handle to the left of a key and adjust that. You can pull them up down and left and right, making nailing that velocity and falloff easy. Generaly speaking the shorter the handle the quicker the change and the steeper the angle (compared to the one on the other side) the sharper the motion.

Another massive benefit to the curve editor is that you can actually change the vaule of a keyframe, not just it's interpolation. If you grab a handle you can move it up and down, which raises or lowers the value. That means no need to navigate your scene to find the right part, move to the correct frame, set what you are recording then make then change and record it. You just select one dot and move that. No re-recording needed.

Words : Rob Redman

Creative director and VFX supervisor at Pariah Studios , Rob is also often found speaking at industry events and teaching either on location or at his studio.

Like this? Read these!

• The Cinema 4D secrets of Iron Man's interface
• Hands-on review: Adobe After Effects CC
• How to land your dream job in VFX

Thank you for reading 5 articles this month* Join now for unlimited access

Enjoy your first month for just £1 / $1 / €1

*Read 5 free articles per month without a subscription

Join now for unlimited access

Try first month for just £1 / $1 / €1

Get the Creative Bloq Newsletter

Daily design news, reviews, how-tos and more, as picked by the editors.

The Creative Bloq team is made up of a group of design fans, and has changed and evolved since Creative Bloq began back in 2012. The current website team consists of eight full-time members of staff: Editor Georgia Coggan , Deputy Editor Rosie Hilder , Ecommerce Editor Beren Neale , Senior News Editor Daniel Piper , Editor, Digital Art and 3D Ian Dean , Tech Reviews Editor Erlingur Einarsson and Ecommerce Writer Beth Nicholls and Staff Writer Natalie Fear , as well as a roster of freelancers from around the world. The 3D World and ImagineFX magazine teams also pitch in, ensuring that content from 3D World and ImagineFX is represented on Creative Bloq. 

Related articles

• 2 Camden Town WFC’s new kit is actually designed for women
• 3 Serif vs. sans-serif: how to choose the right font for your project
• 4 Why creatives need a new alliance to face the existential threat of AI
• 5 Microsoft Surface Laptop 7 review: Copilot+ personified

The PowerPoint Blog

What is a “key frame”.

A movie is a series of images played rapidly to create movement. Remember webcasts are all about throwing away unneeded data. The Key Frame is a critical component in what the file size will be.

Key Frames Within video software is the option to change the Key Frame Rate. The fewer Key Frames you use the smaller the file size. Here are two examples: 1. For PowerPoint content, where little changes from frame-to-frame, use a high Key Frame Rate (10-80). 2. For live video, where the constant motion has lots of change from frame-to-frame, use a low Key Frame Rate (1-12).

Frames Per Second and Key Frame Rate Within video software the Frames Per Second (fps) can be adjusted. We need to consider the fps when setting the Key Frame Rate. We need to really consider how much changes from frame-to-frame how often. Here are to examples: 1. If encoding at 30 fps and the Key Frame Rate is 80, a new Key Frame will occur very quick – about every 2.5 seconds. 2. If encoding at 5 fps and the Key Frame Rate is still set at 80, the Key Frames are nearly 40 seconds apart.

– Troy @ TLC

All Subjects

Key frames are specific frames in animation or video editing that mark the start and end points of a transition or movement. They are crucial for defining the major changes in position, timing, and effect within a scene, allowing creators to control the flow of action and visual storytelling effectively.

5 Must Know Facts For Your Next Test

• Key frames are essential for defining critical moments in animation and video projects, as they mark where significant changes occur.
• By setting key frames at specific points in a timeline, creators can create complex animations with precise movements and effects.
• Key frames can be adjusted to modify the timing and pacing of transitions, giving flexibility to the overall flow of a project.
• The use of key frames allows animators to save time by automatically generating in-between frames through interpolation.
• In pitch decks and visual aids, key frames can serve as focal points to highlight critical shifts in narrative or design elements.

Review Questions

• Key frames play a vital role in storytelling by marking the significant points where actions or emotions change within a scene. By strategically placing key frames, creators can emphasize important narrative moments, enhance dramatic tension, or depict character development. This helps to guide the audience's understanding and emotional response to the unfolding story.
• The placement of key frames directly influences the pacing and timing of an animated sequence. If key frames are positioned closer together, it creates faster movements, while wider spacing results in slower actions. This manipulation allows animators to convey urgency or calmness as needed, effectively shaping how viewers perceive the rhythm and flow of the scene.
• In pitch decks, key frames serve as powerful tools for visual storytelling, helping to illustrate pivotal moments and transitions that reinforce the message being conveyed. By using key frames to highlight important data points or design shifts, presenters can maintain audience interest and focus. This dynamic presentation style can make complex information more digestible and memorable, leading to a stronger impact during pitches.

Related terms

Animation : A technique in which individual frames are manipulated to appear as moving images, creating the illusion of motion.

Timeline : A visual representation of a sequence of events in video editing or animation, showing the duration and timing of clips, effects, and key frames.

Interpolation : The process of calculating intermediate frames between two key frames to create smooth transitions in animation or video.

" Key Frames " appears in:

Study guides ( 1 ).

• Advanced Film Writing - 12.2 Creating Pitch Decks and Visual Aids

Subjects ( 4 )

• 2D Animation
• Art Direction
• Storytelling for Film and Television

© 2024 Fiveable Inc. All rights reserved.

Ap® and sat® are trademarks registered by the college board, which is not affiliated with, and does not endorse this website..

• Free Training
• Testimonials

Top 10 Keyframing Animation Tips for Smooth Animations

Make More Money as an Animator

Do You Want To Master The Business of Animation? FREE MASTERCLASS “Articulation is how much detail there is in the movement. How many drawings for a particular amount of time on screen. Quality animation is all about detail and movement.” Will Gadea, IdeaRocket Creative Director, and Founder 

GIF via Future Reality Lab

Creating high-quality, Hollywood-style animations is as easy as clicking a few buttons with modern video editing software. There are countless video editing effects that animators use to elevate their videos. One popular method is to create keyframing animations. 

Adding keyframing animations to your video is an easy way to enhance your footage without learning any special in-depth animation skills . This makes it ideal for new animators with limited industry experience who want to make their mark on the animation industry. 

This extremely versatile technique can be applied to many aspects of your videos. All you need is creativity and a basic understanding of keyframing animations to use it to your advantage. 

In this blog, we’ll teach you how to improve the quality of your animations through keyframing. By the end of this blog, you’ll know all the basics of keyframing and how to use it.

What is Keyframing?

GIF via School of Motion

Many factors go into professional animation, but keyframe frequency and articulation are at the top of the list. Animation is made of numerous frames, and when those frames are shown at a certain speed, we perceive the individual frames as moving images. 

Keyframes are the important frames that contain information on a start and end point of movement. A keyframe relays information about two things: first, it tells you what the action of your frame is at a certain point in time; second, it tells you what time that action occurs.

Keyframing is the process of indicating the beginning and end of a specific animation and allowing your editing program to fill in the transition between these two points. This filling-in creates the illusion of continuity between the still images of each keyframe marker.

Simply put, keyframes are the markers or anchor points that animators use to indicate how they want their animations to look at a specific time. Keyframes are used to create keyframe intervals that tell an editing program where one clop should transition from one state to another in a specified time. 

In other words, keyframes contain states of how you want things to be, whether it’s a zoom amount, a specific color, or any of hundreds of possible parameters and thousands of possible combinations of parameters. 

As the animation plays between the keyframes, your editing program calculates steps between the two states. It then interpolates intermediate frames between the keyframes and creates a natural flow between one state to the next.

A state is a combination of parameters. States can be very complex and involve many parameters, like size, shape, color, rotation, brightness, and many more, which can change between keyframes. With so many independent parameters under your control, you can create amazingly sophisticated animations and effects with keyframing.

Each quality that you change in a given keyframe is called a parameter. You can adjust as many parameters as you want with keyframes, and your editing program will automatically transition from your desired starting point to your ending point.

Keyframe interpolation is the movement by which your desired parameters reach their end state. While your editing program determines much of this transition itself, you still have some control over how your animation moves throughout your video and transforms into your desired state.

However, you have to remember that less is more when it comes to keyframes. It's very easy to add keyframes to everything, and you will soon be faced with a timeline that looks intimidating at best and unusable at worst. Keep it simple for the best results. 

Using keyframes may sound intimidating, but programs like PowerDirector, Adobe, and Blender make it easy to learn how to use these editing tools to your advantage. Once you know the basics of keyframes, keyframe interpolation, keyframe interval, and keyframe animation, you can use these animation techniques to create custom animations and transitions in just a few seconds. 

The Benefits of Keyframing Animations

GIF by MeikDraws via DeviantArt

Keyframing allows you to smoothly implement transitions on overlays, animations, and other parameters over a given period. Using keyframes creates clean, natural movement, removing the manual labor involved in creating animations from scratch. 

Keyframes also allow you to customize the effects' duration in your video. Instead of adding an overlay to the entire clip, you can fade the effect in and out over the exact period in which you would like it to appear. 

Animators typically use keyframes to implement a wide range of effects and transitions in their animations, including camera effects like panning from side to side, zooming in or out, and rotating the video. 

Other visual effects can also be added to your animations, including fading out to 0% opacity and adding overlays. You can also animate an object and change its scale or position with keyframes. 

Most editing programs offer a variety of interpolation methods , and the one you should use depends on your parameters and your desired effects. 

If you want your movement to look natural enough that viewers may not notice, utilizing Bezier interpolation will help you create this effect. If you want your change to be sudden, such as a surprising fade to black, hold interpolation will help you produce this movement. 

Without keyframes, your animation would jump from one position to the next like a flipbook. You would have to insert more positions to create a smoother finished product . Playing around with your keyframes and interpolation methods is the best way to achieve your desired movements.

Tips for Keyframing Animations

Design a Good Rig

GIF by Indigustiva via Blender 

Like most things in animation, keyframing can be a time-consuming and intricate task. You can make things easier for yourself by ensuring that you have a clean and easy-to-understand rig.  

This will make it easier for you to find what you need without searching through seemingly endless hierarchies. You will be able to concentrate only on the parts that you are animating at any given time. 

Some editing programs also allow you to add keyframes directly, meaning that you don’t need to lose focus on the actual motion you are working on. 

Don’t Keyframe Everything

Image by rico345100 via blenderartists.org

Keyframing animation is made easy with modern technology and editing programs that often have a  default set of parameters that will be keyframed when you hit the record button. Usually, these include position, rotation, and scale, and sometimes a point level or parameter option too. 

While this has the potential to make your life as an animator a lot simpler, you should try to avoid recording anything that you don't choose yourself. You need absolute control to make sure your animation turns out exactly the way you want it to. 

Putting yourself in control will enable you to understand your timeline better and will reduce unwanted clutter. It will also make it much easier to go back and make adjustments should you need to.

Blocking Out

Image via Animation Addicts

The first step of keyframing is blocking out your animations by creating a basic framework consisting of a start, middle, and end. Although you can make this framework more complicated down the line, you should focus on keeping things simple for now. 

It’s important to get the overall pacing of your animation right before you move on to more complicated things. If you start with just the overall position and orientation and add in things like limb movement and hip sway later.

A tip for making this process easier is to make your start and end keyframes identical. 

Duplicating Keyframes

GIF by Emir via Blender

Once you have the basic keyframes and frameworks in place, you can use them to create loops. Duplicating and recycling keyframe animations will make your job much easier and allow you to add just a few unique movements here and there to spice up your animation. 

Duplicating keyframes will also prevent any odd jerkiness at each point where your animation changes from one movement to another.

After creating your basic set of keyframes, you can just select them and drag them to generate a new loop and add in a little variation for a more natural look. By grabbing the end handle of any selected set of frames, you can change the speed of that section with a smooth interpolation.

Keep an Eye on Your Curves

GIF by 1ucasvb via Tumblr

While keeping track of your keyframes is great for organization, keeping an eye on your curves is also a good idea. Most editing programs have this feature, allowing you to fine-tune things and add some personality to your animation through interpolation. 

Curves work in a logical way, moving along the timeline, so if you want to ramp up your motion, you would take the handle to the left of a key and adjust that. You can pull them up and down or left and right, making nailing the velocity and falloff easy. 

Another massive benefit to the curve editor is that you can change the value of a keyframe, not just its interpolation. That means no need to navigate your scene to find the right part. Move to the correct frame, set what you are recording, then make your change and record it. 

Studying Other Animated Films

Image via Pixar Online Library

The key to good animation is studying the work of other animators and animation studios to learn from their techniques. This counts for keyframing animation as well.

If you want to do a good keyframe design, studying the film composition and the visual representation is a good idea. This gives you a better understanding of how to serve the overall story in the film and animation through keyframing animation.

Studying the work of other animators will also allow you to gain a better understanding of composition. To design keyframes and show them in the best possible way, we need to understand the composition elements of keyframe design. For more information on this, you can refer to the 4 Art Composition Design Tips .

Knowing the Aspect Ratio

Image via Wonderfox

While we’re on the subject of composition, you need to know the size, or aspect ratio, of each composition for good keyframing animation. 

Some of the more commonly used aspect ratios are:

• 1:85:1 ( Flat)
• 2:39:1 ( Scope)
• 16:9 ( Quad HD)

When you understand how aspect ratio works , how it can influence the look and feel of your animation, and how to use it to manipulate and improve the quality of your animations, you’ll be on track to creating breathtaking animations in no time. 

The Balance of the Keyframes

Image via Academy of Art University

Balancing keyframes is another important tip for keyframing animations and creating smooth animations. Balancing keyframes largely depends on how you place the light and how you resize the objects in the frame. 

The sense of balance in key frames affects the viewer’s perception of the subject. When objects in the frame are placed evenly, it makes the composition feel balanced. For example, an unbalanced composition may be one in which the focal point points to only one area of the frame, accompanied by confusion, chaos, and tension.

The Focal Point of the Frame

Image by Behance via Pinterest

The focal point of the frame or the center of interest in the composition should be the easiest for the viewer to notice. This means that you have to focus on the design to make the focal point clear to the viewer. 

Creating a clear focal point in each scene will also aid in your visual storytelling and help drive the narrative along. Remember that the focal point can include one or more subjects and this largely depends on your specific animation, the scene you are animating, and the keyframes at any given point in the animation.

Adding Depth to Keyframes

Image by Fenharell via Funcom

By adding depth-of-field and depth cues, you can increase interest in the two-dimensionality of the frame to produce a believable three-dimensional space. The relative size it cues is based on the idea that two objects are the same size, and by seeing the smaller object, this will be perceived as farther away, thus creating the illusion of depth.

Add More Keyframes to Articulate Your Animation

GIF via Tenor

We explained above that adding too many keyframes can make your animation too complicated to work on. But the truth is, the more keyframes you add, the smoother your animation will be. This is called articulation and it’s the key to creating smooth animations. More key frames per second mean better articulation for every action and smoother animation. 

While adding too many keyframes can be a waste of time, you should still be sure to add enough to make your animation appear seamless and as natural as possible (if that’s what you’re aiming for). 

However, keep in mind that every keyframe can’t be treated equally. If every second of animation is made up of 24 frames per second, all of those frames can’t be important moments. Your eye just doesn’t register images that quickly. 

Instead of spending time and energy drawing 24 perfect frames that will just fly by, you can create the basics and let your editing software fill in the surrounding frames with less-than-perfect animation or the bare minimum movement to get the point across. 

Why Keyframing Animations are Key to Creating Smooth Animations

Keyframing animations is one of the most important steps in the animation process as it offers the most benefits in terms of creating smooth and natural-looking animations. 

There are countless benefits to understanding and using keyframes to your advantage. In this blog, we shared some of our top tips for keyframing animations so you can create the best quality animations possible. 

If you implement these simple tips into your animation process, you’ll be sure to up your game and drastically improve your animation quality! If you’re interested in knowing more about animation and the business side of animation, check out our free masterclass , download a copy of our free marketing handbook , and check out our blog on “ How to Start an Animation Studio ”!

Lacking Business Skills as an Animator?

Feeling Stuck in Your Animation Career? Learn How to Break the $10,000 Per Month Barrier!

How to Create Key-frame Concept Art

Learn how to create your own concept story keyframes using the compositing techniques in movies with Concept Artist Nikola Angelkoski

MattePaint Team

Think of Keyframes as important moments in the story

A Keyframe concept is a concept illustration of a “Key Scene” from the movie script that is usually used to convey the feel, style or look of a specific scene. Think of Keyframes as important moments in the story that are usually drawn by the storyboard artists. Due to the importance of these, they are further developed into concept illustrations to show the directors or producers what the movie might look like.

Study from movies

To become better in Keyframe design first you have to study cinematic composition and how the visual representation serves the overall story in the movie.

One way to practice cinematic composition is to copy the look and feel of 'stills' from movies. Why study from a movie already made? Well even though they have amazing rendering skills, a lot of artists feel like the image they created is over cluttered with information or it feels as if the emotion the artist is trying to convey is not property captured in his/her piece. So, directly copying an existing frame from a movie helps you train your eye for what exactly you should be looking for when composing your next piece. Some sites to get movie stills are Filmgrab and Evan Richards .

Aspect Ratios

Every composition you create is first defined by the dimensions of your frame. The width and height of a frame is referred to as the aspect ratio. Some of the more commonly used aspect ratios are:

⦁ 1:85:1 (Flat) ⦁ 2:39:1 (Scope) ⦁ 16:9 (QuadHD)

If you want to explore more examples of other aspect ratios on different resolutions, I highly suggest you to visit this link: ChameleonDG

“The size of an object in the frame should be directly related to its importance in the story at that moment.” ~ Alfred Hitchcock

Compositional guides.

Think of composition as placement of your subjects in the scene, whether there is only one or several elements in the frame that can be used to visually show excitement, tensions or suspense for the audience.

Rule of Thirds

One of the oldest conventions used to create harmonious compositions is known as the Rule of Thirds. By dividing the frame into three pieces along the width and height you are creating spots where the lines of the thirds cross, providing a guide for the placement of the subject.

High Angles and Low Angles

Depending on the angle of your camera position, you can tell the audience how to perceive that subject. If you have a high angle camera above eye-level, it results in us looking down on the subject ( showing weakness, defeat, vulnerability). A low angle, placed below eye-level, shows us the subject in power (showing confidence, control, power etc.).

Balanced vs Unbalanced

Depending on how you place light or how you size an object in your frame can affect the perception of the audience of that subject, making the compositions feel balanced when objects in the frame are evenly placed. Unbalanced compositions for example could be when the focus is pointed only in one area of the frame while accompanied by disorder, chaos and tension.

Focal Points

Focal points, or the center of interest in the composition are the areas where the audience's view will focus on because of the arrangement of the elements in the frame. Focal points can include one or more subjects and are commonly created using the compositional guides as mentioned above.

Adding Depth

By adding depth I’m not just referring to depth of field, but depth cues. This way you’re adding interest to the two-dimensionality of the frame to produce a believable three dimensional space. The relative size of a depth cue is based on the idea that two objects are the same size, and by seeing a smaller one, this is perceived as being further away, thus creating the illusion of depth.

Different types of shots:

Establishing shot.

The Establishing shot is usually an exterior that showcases the location where the following action will take place.

Over the shoulder Shot

These shots are primarily used whenever there is a conversation between two or more characters or when a character is looking at something.

The Long Shot includes whole characters in the frame, along with a large portion of the area where the action will take place. There are also extreme long shots that can convey the smallness of a subject in the established area.

Medium Shot

Medium shots include one or more characters from the waist up and still including some of the surrounding area. There are also medium long shots that show the characters from the knees up in the frame.

Close up Shot

Close up shots are mainly used to show the character’s behavior and emotion.

In Practice

As an example, I’m going to do a breakdown of how I approached my project “Evolution 169”

Remember when I mentioned stealing compositions above? My main inspiration was the promotional material for Alien: Covenant, the making of Walter.

Alien: Covenant (2017) 20th Century Fox

Even though I follow the same idea of an android being constructed, I solely keep my focus to the focal points and camera angle ideas.

I wanted to draw the audience to the center of the frame

For this frame I went with a medium long shot. Why? Well because this is an interior and the overall action takes place in this room. I still wanted to show the emptiness and coldness of the environment while being accompanied by the sunset with the moon in the shot, representing the ending and beginning of something new.

Because this is a story about an android being constructed; I wanted to start off by showing the unseen. I wanted to draw the audience to the center of the frame (the covered body, a lifeless element).

The two characters in dark hazmat suits serve as both guides to the focal point and as a symbolic element of being the “creators” of the robot.

You will also notice that I’m at a low angle camera view. The reason why I went with that angle of view is the first main thing you will think is - “Oh its because the humans are it’s creators so normally they will be seen as the dominant one”. Well true, but also focus on the table. See the statues holding the bed of the android? That background detail alone helps you create an idea in your head that what is created might be worshiped or even seen as god.

In the background you will also notice a balanced split. On the left side are the humans and the statue of David, symbolizing the things created by nature itself. While on the right, the aspect of artificial; things created by the human hand. So naturally, the center where the robot lays, represents the merge of the two.

The depth of field and depth cues are added to guide the viewer to the center focal point.

For the second frame, again I went with a low angle, medium shot camera view. Here you will notice that the human is barely visible in this shot. I made his hazmat suit to be darker and pushed in the back so my main focal point (the android) is well lit and the center of attention.

The pipes and UI element serve as a guide to the head of the android, as this is the place where I want the viewers to focus.

The depth of field is not so shallow, because I wanted the viewers to see that the elements used to create this robot are a copy of the human anatomy. Symbolizing the look for perfection, and immortality.

The third frame is opposite from what we’ve seen so far. It has a high angle, medium shot. I’ve added the plastic cover to have small bounces of uneven lights covering the android as well as robotic arms covering half of her face. This way I have created an unbalanced composition that shows the facial expression on the android (fear). In this frame, the android has been awoken mid construction. The high angle represents her fear of her current state.

The fourth frame shows us the android being an exact copy of a human. She is sitting alongside her bed, calm and her eyes are pointed towards her creators off camera.

For this frame I intentionally went with an eye level camera to show her being an equal to mankind with a few elements of machinery in the background to show that she is different from the rest.

The final shot also includes a vast landscape viewed from the windows of the interior. Here I wanted to show her significance in this world while still having the Moon and Sun in the frame representing the unknown fate she will bring.

If you want to learn more about the cinematography techniques I highly recommend these books: - The Filmmaker's Eye - Filmmaker's Handbook: A Comprehensive Guide for the Digital Age

Learn. Create. Repeat.

I hope you have enjoyed my article and learnt some things! If you would like to see more of my work or want to connect, you can find me at:

Artstation | Linkedin

Improve your practice.

Enhance your soft skills with a range of award-winning courses.

How to Structure your Presentation, with Examples

August 3, 2018 - Dom Barnard

For many people the thought of delivering a presentation is a daunting task and brings about a  great deal of nerves . However, if you take some time to understand how effective presentations are structured and then apply this structure to your own presentation, you’ll appear much more confident and relaxed.

Here is our complete guide for structuring your presentation, with examples at the end of the article to demonstrate these points.

Why is structuring a presentation so important?

If you’ve ever sat through a great presentation, you’ll have left feeling either inspired or informed on a given topic. This isn’t because the speaker was the most knowledgeable or motivating person in the world. Instead, it’s because they know how to structure presentations – they have crafted their message in a logical and simple way that has allowed the audience can keep up with them and take away key messages.

Research has supported this, with studies showing that audiences retain structured information  40% more accurately  than unstructured information.

In fact, not only is structuring a presentation important for the benefit of the audience’s understanding, it’s also important for you as the speaker. A good structure helps you remain calm, stay on topic, and avoid any awkward silences.

What will affect your presentation structure?

Generally speaking, there is a natural flow that any decent presentation will follow which we will go into shortly. However, you should be aware that all presentation structures will be different in their own unique way and this will be due to a number of factors, including:

• Whether you need to deliver any demonstrations
• How  knowledgeable the audience  already is on the given subject
• How much interaction you want from the audience
• Any time constraints there are for your talk
• What setting you are in
• Your ability to use any kinds of visual assistance

Before choosing the presentation’s structure answer these questions first:

• What is your presentation’s aim?
• Who are the audience?
• What are the main points your audience should remember afterwards?

When reading the points below, think critically about what things may cause your presentation structure to be slightly different. You can add in certain elements and add more focus to certain moments if that works better for your speech.

What is the typical presentation structure?

This is the usual flow of a presentation, which covers all the vital sections and is a good starting point for yours. It allows your audience to easily follow along and sets out a solid structure you can add your content to.

1. Greet the audience and introduce yourself

Before you start delivering your talk, introduce yourself to the audience and clarify who you are and your relevant expertise. This does not need to be long or incredibly detailed, but will help build an immediate relationship between you and the audience. It gives you the chance to briefly clarify your expertise and why you are worth listening to. This will help establish your ethos so the audience will trust you more and think you’re credible.

Read our tips on  How to Start a Presentation Effectively

2. Introduction

In the introduction you need to explain the subject and purpose of your presentation whilst gaining the audience’s interest and confidence. It’s sometimes helpful to think of your introduction as funnel-shaped to help filter down your topic:

• Introduce your general topic
• Explain your topic area
• State the issues/challenges in this area you will be exploring
• State your presentation’s purpose – this is the basis of your presentation so ensure that you provide a statement explaining how the topic will be treated, for example, “I will argue that…” or maybe you will “compare”, “analyse”, “evaluate”, “describe” etc.
• Provide a statement of what you’re hoping the outcome of the presentation will be, for example, “I’m hoping this will be provide you with…”
• Show a preview of the organisation of your presentation

In this section also explain:

• The length of the talk.
• Signal whether you want audience interaction – some presenters prefer the audience to ask questions throughout whereas others allocate a specific section for this.
• If it applies, inform the audience whether to take notes or whether you will be providing handouts.

The way you structure your introduction can depend on the amount of time you have been given to present: a  sales pitch  may consist of a quick presentation so you may begin with your conclusion and then provide the evidence. Conversely, a speaker presenting their idea for change in the world would be better suited to start with the evidence and then conclude what this means for the audience.

Keep in mind that the main aim of the introduction is to grab the audience’s attention and connect with them.

3. The main body of your talk

The main body of your talk needs to meet the promises you made in the introduction. Depending on the nature of your presentation, clearly segment the different topics you will be discussing, and then work your way through them one at a time – it’s important for everything to be organised logically for the audience to fully understand. There are many different ways to organise your main points, such as, by priority, theme, chronologically etc.

• Main points should be addressed one by one with supporting evidence and examples.
• Before moving on to the next point you should provide a mini-summary.
• Links should be clearly stated between ideas and you must make it clear when you’re moving onto the next point.
• Allow time for people to take relevant notes and stick to the topics you have prepared beforehand rather than straying too far off topic.

When planning your presentation write a list of main points you want to make and ask yourself “What I am telling the audience? What should they understand from this?” refining your answers this way will help you produce clear messages.

4. Conclusion

In presentations the conclusion is frequently underdeveloped and lacks purpose which is a shame as it’s the best place to reinforce your messages. Typically, your presentation has a specific goal – that could be to convert a number of the audience members into customers, lead to a certain number of enquiries to make people knowledgeable on specific key points, or to motivate them towards a shared goal.

Regardless of what that goal is, be sure to summarise your main points and their implications. This clarifies the overall purpose of your talk and reinforces your reason for being there.

Follow these steps:

• Signal that it’s nearly the end of your presentation, for example, “As we wrap up/as we wind down the talk…”
• Restate the topic and purpose of your presentation – “In this speech I wanted to compare…”
• Summarise the main points, including their implications and conclusions
• Indicate what is next/a call to action/a thought-provoking takeaway
• Move on to the last section

5. Thank the audience and invite questions

Conclude your talk by thanking the audience for their time and invite them to  ask any questions  they may have. As mentioned earlier, personal circumstances will affect the structure of your presentation.

Many presenters prefer to make the Q&A session the key part of their talk and try to speed through the main body of the presentation. This is totally fine, but it is still best to focus on delivering some sort of initial presentation to set the tone and topics for discussion in the Q&A.

Other common presentation structures

The above was a description of a basic presentation, here are some more specific presentation layouts:

Demonstration

Use the demonstration structure when you have something useful to show. This is usually used when you want to show how a product works. Steve Jobs frequently used this technique in his presentations.

• Explain why the product is valuable.
• Describe why the product is necessary.
• Explain what problems it can solve for the audience.
• Demonstrate the product  to support what you’ve been saying.
• Make suggestions of other things it can do to make the audience curious.

Problem-solution

This structure is particularly useful in persuading the audience.

• Briefly frame the issue.
• Go into the issue in detail showing why it ‘s such a problem. Use logos and pathos for this – the logical and emotional appeals.
• Provide the solution and explain why this would also help the audience.
• Call to action – something you want the audience to do which is straightforward and pertinent to the solution.

Storytelling

As well as incorporating  stories in your presentation , you can organise your whole presentation as a story. There are lots of different type of story structures you can use – a popular choice is the monomyth – the hero’s journey. In a monomyth, a hero goes on a difficult journey or takes on a challenge – they move from the familiar into the unknown. After facing obstacles and ultimately succeeding the hero returns home, transformed and with newfound wisdom.

Storytelling for Business Success  webinar , where well-know storyteller Javier Bernad shares strategies for crafting compelling narratives.

Another popular choice for using a story to structure your presentation is in media ras (in the middle of thing). In this type of story you launch right into the action by providing a snippet/teaser of what’s happening and then you start explaining the events that led to that event. This is engaging because you’re starting your story at the most exciting part which will make the audience curious – they’ll want to know how you got there.

• Great storytelling: Examples from Alibaba Founder, Jack Ma

Remaining method

The remaining method structure is good for situations where you’re presenting your perspective on a controversial topic which has split people’s opinions.

• Go into the issue in detail showing why it’s such a problem – use logos and pathos.
• Rebut your opponents’ solutions  – explain why their solutions could be useful because the audience will see this as fair and will therefore think you’re trustworthy, and then explain why you think these solutions are not valid.
• After you’ve presented all the alternatives provide your solution, the remaining solution. This is very persuasive because it looks like the winning idea, especially with the audience believing that you’re fair and trustworthy.

Transitions

When delivering presentations it’s important for your words and ideas to flow so your audience can understand how everything links together and why it’s all relevant. This can be done  using speech transitions  which are words and phrases that allow you to smoothly move from one point to another so that your speech flows and your presentation is unified.

Transitions can be one word, a phrase or a full sentence – there are many different forms, here are some examples:

Moving from the introduction to the first point

Signify to the audience that you will now begin discussing the first main point:

• Now that you’re aware of the overview, let’s begin with…
• First, let’s begin with…
• I will first cover…
• My first point covers…
• To get started, let’s look at…

Shifting between similar points

Move from one point to a similar one:

• In the same way…
• Likewise…
• Equally…
• This is similar to…
• Similarly…

Internal summaries

Internal summarising consists of summarising before moving on to the next point. You must inform the audience:

• What part of the presentation you covered – “In the first part of this speech we’ve covered…”
• What the key points were – “Precisely how…”
• How this links in with the overall presentation – “So that’s the context…”
• What you’re moving on to – “Now I’d like to move on to the second part of presentation which looks at…”

Physical movement

You can move your body and your standing location when you transition to another point. The audience find it easier to follow your presentation and movement will increase their interest.

A common technique for incorporating movement into your presentation is to:

• Start your introduction by standing in the centre of the stage.
• For your first point you stand on the left side of the stage.
• You discuss your second point from the centre again.
• You stand on the right side of the stage for your third point.
• The conclusion occurs in the centre.

Key slides for your presentation

Slides are a useful tool for most presentations: they can greatly assist in the delivery of your message and help the audience follow along with what you are saying. Key slides include:

• An intro slide outlining your ideas
• A  summary slide  with core points to remember
• High quality image slides to supplement what you are saying

There are some presenters who choose not to use slides at all, though this is more of a rarity. Slides can be a powerful tool if used properly, but the problem is that many fail to do just that. Here are some golden rules to follow when using slides in a presentation:

• Don’t over fill them  – your slides are there to assist your speech, rather than be the focal point. They should have as little information as possible, to avoid distracting people from your talk.
• A picture says a thousand words  – instead of filling a slide with text, instead, focus on one or two images or diagrams to help support and explain the point you are discussing at that time.
• Make them readable  – depending on the size of your audience, some may not be able to see small text or images, so make everything large enough to fill the space.
• Don’t rush through slides  – give the audience enough time to digest each slide.

Guy Kawasaki, an entrepreneur and author, suggests that slideshows should follow a  10-20-30 rule :

• There should be a maximum of 10 slides – people rarely remember more than one concept afterwards so there’s no point overwhelming them with unnecessary information.
• The presentation should last no longer than 20 minutes as this will leave time for questions and discussion.
• The font size should be a minimum of 30pt because the audience reads faster than you talk so less information on the slides means that there is less chance of the audience being distracted.

Here are some additional resources for slide design:

• 7 design tips for effective, beautiful PowerPoint presentations
• 11 design tips for beautiful presentations
• 10 tips on how to make slides that communicate your idea

Group Presentations

Group presentations are structured in the same way as presentations with one speaker but usually require more rehearsal and practices.  Clean transitioning between speakers  is very important in producing a presentation that flows well. One way of doing this consists of:

• Briefly recap on what you covered in your section: “So that was a brief introduction on what health anxiety is and how it can affect somebody”
• Introduce the next speaker in the team and explain what they will discuss: “Now Elnaz will talk about the prevalence of health anxiety.”
• Then end by looking at the next speaker, gesturing towards them and saying their name: “Elnaz”.
• The next speaker should acknowledge this with a quick: “Thank you Joe.”

From this example you can see how the different sections of the presentations link which makes it easier for the audience to follow and remain engaged.

Example of great presentation structure and delivery

Having examples of great presentations will help inspire your own structures, here are a few such examples, each unique and inspiring in their own way.

How Google Works – by Eric Schmidt

This presentation by ex-Google CEO  Eric Schmidt  demonstrates some of the most important lessons he and his team have learnt with regards to working with some of the most talented individuals they hired. The simplistic yet cohesive style of all of the slides is something to be appreciated. They are relatively straightforward, yet add power and clarity to the narrative of the presentation.

Start with why – by Simon Sinek

Since being released in 2009, this presentation has been viewed almost four million times all around the world. The message itself is very powerful, however, it’s not an idea that hasn’t been heard before. What makes this presentation so powerful is the simple message he is getting across, and the straightforward and understandable manner in which he delivers it. Also note that he doesn’t use any slides, just a whiteboard where he creates a simple diagram of his opinion.

The Wisdom of a Third Grade Dropout – by Rick Rigsby

Here’s an example of a presentation given by a relatively unknown individual looking to inspire the next generation of graduates. Rick’s presentation is unique in many ways compared to the two above. Notably, he uses no visual prompts and includes a great deal of humour.

However, what is similar is the structure he uses. He first introduces his message that the wisest man he knew was a third-grade dropout. He then proceeds to deliver his main body of argument, and in the end, concludes with his message. This powerful speech keeps the viewer engaged throughout, through a mixture of heart-warming sentiment, powerful life advice and engaging humour.

As you can see from the examples above, and as it has been expressed throughout, a great presentation structure means analysing the core message of your presentation. Decide on a key message you want to impart the audience with, and then craft an engaging way of delivering it.

By preparing a solid structure, and  practising your talk  beforehand, you can walk into the presentation with confidence and deliver a meaningful message to an interested audience.

It’s important for a presentation to be well-structured so it can have the most impact on your audience. An unstructured presentation can be difficult to follow and even frustrating to listen to. The heart of your speech are your main points supported by evidence and your transitions should assist the movement between points and clarify how everything is linked.

Research suggests that the audience remember the first and last things you say so your introduction and conclusion are vital for reinforcing your points. Essentially, ensure you spend the time structuring your presentation and addressing all of the sections.

What the Hell Is a Keyframe and How to Use It Properly

Digital Animation is all about Keyframes. Inside After Effects, you’ll find a variety of those guys and knowing the best way to use them will help your animations to achieve the results you are aiming for.

We’ve talked before about  how to improve your skills and craft better animations , and presented some of the animation curves you can use to get different animations. Besides that, the  type of keyframe  you use for each moment,  tells a lot about the result you will get.

For those beginning in Animation and After Effects keyframes can sound very weird and understanding all of them can be difficult. In fact, there isn’t much information on the program about what each one of them does. Also, the way you apply Keyframes is very different from each other, making it harder to memorize in the beginning.

Even for those who already have some experience with After Effects, some Keyframes can be “new” since they are all somehow “hidden” at the same place.

That’s why we are here.  I’m going to  teach you about all the Keyframes  you will encounter inside After Effects and  how each one of them will affect your animations.

What The Hell is a Keyframe?

Keyframes are the most important moments of an animation. As the name says, they are the “Key” of all the frames of an animation. They are the ones that show a difference in the acting of a character, a crucial change in a movement or even a great pose.

In Motion Graphics, or Digital Animation, they act the same way — they determine what change from a previous pose or moment, and it’s where the actual animation is done.

Its difference from traditional animation is that animators make the interpolation between keyframes one by one, and they are responsible for creating the correct timing and pacing for the movement. In Digital Animation, the software is responsible for calculating and interpolating itself the animation. One of the things we can control on digital animation is how this interpolation will happen.

By knowing and applying the correct type of Keyframe, you can achieve the full power of automatic interpolation, making sure your animation will occur as you had in your head.

Keyframes types

I like to separate the Keyframes in  three categories  that make it easy to understand and visualize each one of them.

Linear Keyframes are those that create a linear velocity. It means that there is no change in speed during the movement.

Linear Keyframe

The Linear Keyframe is the default of AfterEffects. Whenever you begin to animate, the first Keyframe will always be Linear until you change them to another type. When you have an animation that starts and ends with a Linear Keyframe, it will act as we mentioned before — the animation will start and end with the same velocity.

Continuous Keyframe

The Continuous Keyframe or Continuous Velocity — when talking about interpolation — works between two Keyframes, which can be Linear or of any Keyframe. What they do is to add a third moment without changing the previous velocity of the two keyframes.

For example, if we have an object going from the point A to point B in the Scene, and we want to add a Point C, between them, the “normal” way is to have the animation starting at point A, moving and ending at point C, then starting again at Point C, moving, and ending at point B.  It can be very confusing, but the image below will make it better to understand.

When we change the Point C Keyframe to a Continuous Keyframe, what it does is that now, instead of having those starts and stops in the animation, we have the animation beginning at the Point A moving and ending at Point B but passing by the Point C in the middle of that. Instead of having Point C as a Start/End point, now it’s more like a “detour” of the movement.

The Eases Keyframes are the  most interesting kind of Keyframes.  They are the ones associated with the sensation of both acceleration and weight, which are essential in Motion Graphics.

There are  two basic Eases that comes from the concept of Traditional Animation — Ease In and Ease Out . Depending on your source of animation principles, you can also find them as Slow In and Slow Out.

However, most of the traditional resources about those Eases can explain it as the opposite of what they mean in After Effects. Since we are relating the Keyframes here to Digital Animation,  we will use the concept from Ease In and Ease Out presented in After Effects.

Ease In means that  the movement will “come in” slowly to this Keyframe . For example, if we set a Linear Keyframe as the first position and them an Ease In Keyframe as the second one, the animation will start in a linear velocity and gradually slow down until reaching the second position — Just like our guy above.

The  Ease Out is the opposite of the Ease In . Instead of having the animation slowing down into the second position (how it was with the Ease In Keyframe), the animation will “come out” slowly from the first position and end up linear, or whatever way you set the second Keyframe to be.

You can visualize the Ease Out as acceleration and the Ease In as a slowdown movement. No matter if you use an Ease In or Ease Out Keyframe, the next or previous Keyframe can be of any type. You can even have an Ease Out Keyframe to an Ease In, what will mean an object to start moving slowly, getting faster in the middle of the movement, and then slow down until reaching the final position.

The Easy Ease is the  combination of Ease In and Ease Out in one single Keyframe . Sometimes you have entire animations made only with that kind of Keyframes.

The Easy Ease means that the animation will  come in  slowly to this Keyframe and also  go out  slowly from this frame to the next one. The animation reaches some Start/Stop points like in the Linear Keyframe, but with moments of acceleration and deceleration.

The Hold Keyframes are  the ones that in fact “holds” the movement . When you need something to stay static or a particular pose to maintain as it is without moving, you should use a Hold Keyframe.

Hold Keyframe

The Hold Keyframe symbol says a lot about it. It makes no animation to happen.  It “blocks” the animation in this element or character.  It’s also very used when you want to “freeze” a particular frame of a movie you’ve imported, and also when “blocking” a character animation.

One method of character animation is to start by positioning the character poses over time to analyze the timing and spacing of the whole movement. When done in After Effects, we always begin working with Hold Keyframes instead of the Linear or Ease ones.

Hold In and Hold Out

The Hold In and Hold Out are two Keyframes that relates to the Ease In and Ease Out Keyframes I’ve mentioned above.  You can have an animation happening in an Ease In Keyframe and turn it into a Hold Keyframe or the opposite.

It’s most of the time used to prevent some bugs After Effects can make when interpolating Keyframes.

A common way to “hold” a position when you don’t need an animation by that time is to duplicate the last Keyframe and place it further on the timeline.

However, After Effects can misunderstand that it should hold the animation, and instead, he creates some weird interpolations that will generate animation between these two identical poses. The Hold Keyframes are there to help you prevent this from happening and avoiding undesired movements.

Keyframes names are related to its interpolation and most of them don’t have a specific name besides their  interpolation  mode. Even though I like to name them the way I did, It’s not a formal naming convention but one that I find easier to understand.

The shapes and names of the different Keyframes tell a lot about themselves. Play with them and see for yourself the different results you can achieve by mixing different Keyframes.

Combine it with the  Animation Curves.  It will expand your possibilities and also allow you to make whatever movement you want.

Related Posts

Motion graphics explained. Learn the definition of motion graphics, how it relates to animation, and…

In today’s world, it’s not that hard to face many ads during the day —…

What does the new decade bring? Animated content. Not only because of the emotional connections…

Felippe Silveira

• Posted by Felippe Silveira

3 Ways to Apply the Power of Animation in Health Tech Communications

Health tech marketing must communicate technical concepts in engaging ways. So, avoid the conventional. Learn 3 ways animation boosts health tech marketing.

• Posted by Raff Marques

Animation’s Evergreen Value for Health Tech Companies

Animation is a game-changer for health tech brands. Learn the true ROI of animation and how this medium will boost your health tech brand’s digital impact.

Keyframe Animation Syntax

Basic Declaration & Usage

You can use any number of “stops” in the @keyframe animation, and it’s one of the main strengths of keyframe animations. While CSS transition is only from one state to another, a keyframe animation can interpolate between many different states during its timeline.

If an animation has the same starting and ending properties, one way to do that is to comma-separate the 0% and 100% values:

Or, you could always tell the animation to run twice (or any even number of times) and tell the direction to alternate .

Calling keyframe animation with separate properties

Animation shorthand.

Just space-separate all the individual values. The order doesn’t matter except when using both duration and delay, they need to be in that order. In the example below 1s = duration, 2s = delay, 3 = iterations.

Combine transform and animation

Multiple animations.

You can comma-separate the values to declare multiple animations on a selector.

The steps() function controls exactly how many keyframes will render in the animation timeframe. Let’s say you declare:

If you use steps(10) in your animation , it will make sure only 10 keyframes happen in the allotted time.

The math works out nicely there. Every one second, the element will move 10px to the left and 10px down, until the end of the animation, and then again in reverse forever.

This can be great for spritesheet animation like this demo by samurai .

Browser Support

This browser support data is from Caniuse , which has more detail. A number indicates that browser supports the feature at that version and up.

Mobile / Tablet

More Resources

• MDN: Using CSS Animation
• Can I Use – Browser Support
• Video: Intro to CSS Animations

Old Vendor Syntaxes

is from & to equal to using 0% and 100%?

Yep! It is.

great information i really enjoy it thankx so much

If you use 0%, 50%, and 100% you can have in-betweens. (ex. 0% nothing 25% grow 50% skew 75% change color 100% shrink)

Hello Chris Coyier, The tips and tricks are nice, but if you include the demo link then it would be much helpful. For audience to view the effect live and you too to increase potential returning traffic.

Guys – this is the snippets section! It’s literally the only section of the site that doesn’t have demos/tutorials, it’s purely the grab-n-go snippets… As described!

Font Bulger: http://jsfiddle.net/keif/p8ytP/

Infinite Spinning: http://jsfiddle.net/keif/LTvLF/

Trust me, there is no better way to learn this stuff than to try and build your own demo.

This is even more true with JavaScript frameworks, and JSFiddle is an invaluable tool to throw down quick and dirty code to figure pieces out.

hi Adam, doesn’t the inability to do what Spence is looking for sort of defeats the purpose of animating in this way? I’m encountering exactly the same problem. Any further thoughts on how to resolve it without hover approach?

I agree with Hiren. A demo attached to the snippets (whenever possible) would be very nice.

Agreed on the demo attached to snippets. Just something super-simple.

Hi Chris, you can use -webkit-animation-delay to delay the effects.

hey, Chris, the delay can be found here: http://css-infos.net/property/-webkit-animation-delay

you can also use shorthand to set multiple declarations in one step: http://css-infos.net/property/-webkit-animation

cheers! Atg

Just experimenting, I found the webkit delay syntax is:

“-webkit-animation-delay: 5s;”

Pretty simple!

Hi and many thanks for the code snippet. Can anyone please help me with the following opacity animation:

The div layer starts invisible it then animates to fully visible (after a 2second delay) and remains in that state.

Currently with the code above (including the delay code) I can only get the following: The div layer starts visible it then animates (flashes invisible then animates) to fully visible (after a 2second delay) and remains in that state.

If I put an opacity: 0; on the div then the following occurs: The div layer starts invisible it then animates to fully visible (after a 2second delay) and then returns to invisible.

Is what I am attempting even possible or am I just being a Muppet?

many thanks.

Hey Spence –

I think what you’re looking for is only possible on hover/state change. This is how you do it:

1. Put your animate attributes on the element(s) you want to animate, ie:

.box { -webkit-transition: all 0.2s ease-in-out; }

2. On the hover:

.box:hover { background: red; }

That’s it! Change as necessary.

Adam, doesn’t the inability to do what Spence is looking for sort of defeats the purpose of animating in this way? I’m encountering exactly the same problem. Any further thoughts on how to resolve it without hover approach?

I was just scratching my head about the same question and came across this page:

http://css3animator.com/2010/12/how-to-control-your-css3-animations/

I think this is what you’re looking for, you want to use forwards to make the last keyframe of your animation persist so you can make it look as if something has appeared

I also discovered if you use from and to, this behaviour is the default but it you want to use several keyframed states then it will default back to its original state unless you state the fill-mode

Just use -webkit-animation-iteration-count: 1;

thanks Mairead

cool tut thank you

Hey Chris! Really informative snippet! You’re blog just keeps getting better and better =)

Anyway, you mistakenly repeated ‘has the same’ in “has the same starting…”

Thanks, fixed!

Hi there all! I’m new to CSS and a bit stuck! I have “2” images 70px By 70px I want “1” to stay still and “2” to rotate. image “2” has just a little dot and image “1” has an inner circle and an outer circle.. So I want the little dot to rotate between the inner and outer circle.. So far I’ve done it but the dot makes a big rotation off the screen!! I don’t no if it’s the stage of my rotation or not using margin’s the right way or the perspective origin not being there or non of the above. Thanks for any help in advance and hope someone knows what I’m going on about

Hi Lee, It looks like you have to define a different center for the rotation. When you use -webkit-transform: rotate(), you can define the center point for both 2D and 3D rotation with -webkit-transform-origin ( http://css-infos.net/property/-webkit-transform-origin )

This is really nice! Thanks guys for sharing this. I’m going to try it now :)

I’ve lost count how many times I’ve come back to this page. A great, simple writeup.

Hello everybody! What about overriding a @keyframe animation set? I can’t figure out if it’s possible (using chrome). I explain myself: I made a webapp with several CSS3 features and @keyframes named sets for UI to interact beautifully. My app is customizable by different connected users. The customization is made by importing a css file that override some colors and layouts. I want people how knows CSS3 awesomness to be able to override my standard animations (triggered via javascript). But it looks like if you write a second @keyframes with the same name than first’s, the second definition is ignored. Any help with this?

nice content but boring site colors, Older site is much better in terms of colors

thank u chris. love it.

I can not load this page in ie8. The loading time take more…

Hi Jack I realise in web design terms it’s a lifetime since you made your comment here but I was wondering whether you found a solution. I’ve had this problem for over a week and cannot find an answer. Through much trial and error it appears IE8 simply will not ignore keyframes and as a result just gives up and freezes. I’ve tried hiding the animations in a conditional comment and only serving them to IE9 + and all other browsers. Which works as expected, only as you probably know, IE10 does not support conditional comments! Any help would be greatly appreciated…

Hi guys I want to be able to control the “frames” so I have 5 frames inside of one CSS sprite. They represent each state. I want them to play out over a specified amount of time with out the even motion in between I just want it to go from one frame to another over a specified amount of time. It’s hard for me to explain but flash has a way to do this rather easily.

Hey Chris, I think you want to use the step-start property. I used it on this animation ( http://codepen.io/ScottJH/full/IAyEp ). I realize this comment is over a year old, but hey, thought I’d contribute if I could.

Hello Chris,

I found your website very helpful. I learnt & implemented lot of new ideas from your site.

I was thrilled when I learnt this specific animation trick – usage of keyframes. I have implemented this to highlight the Archives on my site ABAP Help when people hover on it.

Thanks again so much for all these tutorials.

Regards, Naimesh Patel

I have a question about combining animations – when I try the below I get either ‘leaffall’ OR ‘drop’ randomly on refresh. I was expecting both to happen simultaneously.

Anyone seen this problem?

Thank you so much for this snippet, Chris.

Regards from Brazil!

This is great! I was staying away from CSS3 animations because I thought the syntax was too difficult, but you’ve made it pretty simple to understand. Thanks.

For the infinite loop you could change the “to” value to 359deg to prevent it stopping

Its awesome, very easy to understand and a great thank for giving such a wonderful tutorial. I want to know if i want an animation (eg: text animation -fading) on sentance after another how can I achieve it. We do similar kind of text animation when we create presentations.

“The order doesn’t matter except when using both duration and delay, they need to be in that order. In the example below 1s = duration, 2s = delay, 3 = iterations.”

This doesn’t work, if you set the number of iterations to infinite. Instead infinite has to be at the end, hasn’t it?

Thanks for this useful snippets! :)

It should be noted that the keyframe sequence is effectively “reset” if the element is altered via jQuery or some other thing. Keyframes for cycling through border colors for a play button on a player gets reset everytime the button switches from “play” to “pause”. For instance:

@keyframes change_border { 0% {border-color: #202424;} 25% {border-color: #a09b8c;} 75% {border-color: #69707d;} 100% {border-color: #737e88;} } div#player_control { animation: change_border 66s infinite; animation-direction:alternate; -moz-animation: change_border 66s infinite; -moz-animation-direction:alternate; -webkit-animation: change_border 66s infinite; -webkit-animation-direction:alternate; }

This “infinite” sequence of color changing would be interrupted every time the #player_control layer was altered. This is undesirable when other things are also tied to this animation sequence, but aren’t subject to the interruption. Could also simply be undesirable.

I’ve had this same problem. I am animating some list elements on page load so they kind of fly in from the side one at a time and then they are the navigation. On hover, the text in the list element twitches to the side a bit, but when it returns to the default state after the hover effect, the animation is queued again and the list element comes back in from the side again. Very obnoxious.

Have you, or has anyone found a solution to this?

Thank you for this script.

Works great :) Thanks.

It seems that Android 2.2 doesn’t perform well if you specify more than two states in percentages:

a.) this doesn’t work

b.) this works

You can’t spot the problem if you change user agent in webkit browser (etc. safari) or using Android SDK Tools. Bug seems to appear just on particular Android 2.2 devices.

HTC Wildfire Mozilla/5.0 (Linux; U; Android 2.2.1; sv-se; HTC Wildfire Build/FRG83D) AppleWebKit/533.1 (KHTML, like Gecko) Version/4.0 Mobile Safari/533.1

HTC Evo Mozilla/5.0 (Linux; U; Android 2.2; en-us; Sprint APA9292KT Build/FRF91) AppleWebKit/533.1 (KHTML, like Gecko) Version/4.0 Mobile Safari/533.1

thanks “pirohy”

You have some of the best CSS tips. Your directions are often better (more detailed and advanced) than the W3 Schools I appreciate your quality posts Chris. Thank you.

Is there a way to point the animation to another div? Say you make one div like a button to make the other div do the action?

In this example im trying to make it when I hover on the black box the white box will move out of the way. I haven’t seen anything on triggering an animation of one div by hovering on another before, at least not with pure CSS.

Nevermind, I figured it out.

by putting .blackbox:hover and .whitebox in the same selector it acts just the way I want it too.

Awesome snippet by the way. Very helpful.

You could do it with nesting and sibling selectors. You could put both elements in a container div and when you want the white box to move, set the selection to something like:

You could do it with transitions, or you could queue off an animation the same way if it’s not as simple as a transition.

How do you set the animation parameters to be the parameters of the element from the moment the page loads.

So what happens here is that the element will appear on screen, then after 2 seconds it will dissapear and run through the animation. The desired effect is that it appears at opacity 0 until the animation starts, 2 seconds after the page is loaded.

Is there a property I’m missing?

Typical!! Of course I search for ages before I post, then 2 minutes after posting I carry on looking and find the answer!

animation-fill-mode: backwards;

was the solution.

I created this site http://www.css3builder.com it automates the css3 gallery creation by doing the math for the keyframes etc. etc. automatically. You can create one in less than 1 minute for all browsers.

Just a note that the upcoming IE10 will support keyframe syntax without prefixes. So, since IE9 doesn’t support keyframe animations, and since nobody will be using ‘IE10pp’ or whatever, then all examples should now omit the ‘-ms-‘ part.

Also, all the examples on this page should be updated to include the standard syntax, for IE10, upcoming FF16 and other browsers that will eventually support keyframe animations unprefixed.

Page updated! Thanks Louis.

Android browser 2.2 supports it?

I have a animation in an infinite loop. Have any way to set an interval between the loops?

The property “animation-delay” just add a delay before the animation start, then loop without the delay.

I know that I can resolve with a simple “setInterval” script, but I’m trying to figure if have any way to reach to the expected result without that.

When an iOS device is in the process of completing animations #1 and a user scrolls down before animation #2 started, it will simply not load animations set to load later… Any workaround ?

Hey, it would be great if you include the steps() timing function. You can see more details here, for example:

https://hacks.mozilla.org/2011/09/taking-steps-with-css-animations/

Hello Chris for the basic animation you can use the simple code:

.box{ background: white; -webkit-transition: background 5s; }

.box:hover{ background: olive; -webkit-transition: background 1s; }

I don’t think anyone has asked this before:

how do you use the @-webkit-keyframes syntax in SASS (or LESS, for that matter), where the @ is a special character?

bro, where can i test this code with implementing in my site.. i.e sandbox / testdrive ?? Thanks for the content.

This is a very helpful article. Just wondering how you could include audio in the js so it syncs with the play/reset button?

Here is an example of FELIX the cat episode of “The magic bag” done completely using css3 key frame animations. Implemented for webkit browsers only. No JS just pure CSS3 http://pikcle.com/felix/

I couldn’t find any obvious way to delay at either ‘end’ of my looped animation (left to right, in this case). Ended up using:

pre> @-webkit-keyframes pan { 0%, 10% { -webkit-transform: translate3d( 0%, 0px, 0px); } 90%, 100% { -webkit-transform: translate3d(-50%, 0px, 0px); } }

Just a bit annoying having to have a fake animation duration to account for delays.

Is it possible to combine 2 animations for one target.. For example rotating a menu and in the menu time changing its font-size.. Please let me know.. :)

yes its possible

I have used a lot of animations here http://pikcle.com/felix/ works only on webkit browsers, have combined many types of animations for single targets :)

Thanks Melwyn.. Will Try it Out… :)

I think it’s worth pointing out that there are additional properties of animation-direction other than normal and alternate ; These are reverse and alternate-reverse .

The MDN docs has the full set of possible values , but their browser compatibility list looks a little stale. If anyone can confirm browser support for these properties I’m sure Mozilla would appreciate it :)

Is it actually possible to make an animation with the property content ? Something like…

I ran into that just recently. No browser supports it but according to spec writers, they should, and will catch up eventually.

Ya, I think “the solution” for the moment would be to apply a mask over the “…” and animating it to cover the last char, the last two chars, and so on..

Does anyone know of a way to prevent keyframe animations from firing when the browser is resized and it crosses breakpoints? How about on page load?

We’ve a animation in a infinite loop. Have in any manner to set an interval between the loops?

The property “animation-delay” just convey a delay before this animation start, then loop devoid of the delay.

I understand that I can resolve which has a simple “setInterval” piece of software, but I’m looking to figure if have in any manner to reach on the expected result without having that.

Hi, I want to have multiple elements (divs) with different animations. So I have .item1 as you can see below. But I want to make an animation for another item (For example: .item2 with a different slide animation/position), but how can I assign different animations to different elements? Hope someone can help! Thanks!

.item1 { position: absolute; background:url(../img/item.png) 0 0 no-repeat; top: -900px; width: 400px; height: 887px; -webkit-animation: slide 0.5s forwards; -webkit-animation-delay: 0.5s; animation: slide 0.5s forwards; animation-delay: 2s; animation-name: slide; }

@-webkit-keyframes slide { 100% { top: -300px; } }

@keyframes slide { 100% { top: -300px; } }

Hello! I have something like this now: but I don’t know if this is the right way to do it ( name it slide 1 and slide 2)…

Anybody know that either Compass/Sass support keyframe animation inbuild?

I’m sure they have support, but I use a couple of simple mixins I wrote:

Thanks for sharing Josh!

Can I get the webkitAnimationDuration (javascript) equivalent for mozilla and safari browser support same in javascript.

The order doesn’t matter except when using both duration and delay, they need to be in that order. In the example below 1s = duration, 2s = delay, 3 = iterations.”

This doesn’t work, if you set the number of iterations to infinite. Instead infinite has to be at the end, hasn’t it?

I have created one demo. Check it out

http://jsbin.com/alOPUko/11/edit

THANKS Admin . This is a very helpful article. Just wondering how you could include audio in the js so it syncs with the play/reset button?

good luck ….Keep it up

Nice resource dear but is there a way to prevent keyframe animations from firing when the browser is re-sized and it crosses breakpoints? How about on page load?

I am trying to mix this code with another code of css but the problem I am facing is that at run time only one code is showing effect where as another one is not showing it.

Creating a reduced test case is the way to go with issues like this.

BTW thanks for such a nice tutorial.

This is exactly what i was looking for..so simple to implement dude..nice work and looks cure too :)

“-webkit-animation-delay: 5s;”

Pretty simple! http://www.results-2014.com/

Anyone who wanted to put Time Interval between infinite/loop of the animation, please follow this Stackoverflow solution: http://stackoverflow.com/questions/18812055/css-animation-with-time-interval

Happy CSS3 animation!

Nice Job, This is exactly what i was looking for, CSS3 animations is difficult, because for me the syntax was very difficult, but you have done a awesome work, you just understand others in a simple way. Thanks For Sharing.

facing issues here on:

http://www.resultsnews.com/

Is there any way to stop keyframe animations from firing at time of browser re sizing ? I am facing issue on my website http://bit.ly/1j84UFh

made an animation here http://jsfiddle.net/ipsjolly/DDT9b/ But it is not that smooth. it bumps a little when it enters from one frame to other. how can i make it more jelly type like buttons in candy crush .

I am creating an animation of a character who must move his head to the right and left while moving his hands opening and closing them.

It happens that in the CSS I think an animation with two input and both have the fill-mode parameter value with forwards. This will run the first animated properly, but the second is never executed. If I remove the first animation forwards to the second runs properly, but the first animation to end abruptly returns to its starting point. Can anyone tell me how to fix this.

This is my css code:

[Admin note]: Deleting big code dump – feel free to post a link to a Pen on CodePen.

the site very nice and very helpfull fro beginers

Once again, I am landed to this page,

Hey Chris, do you have any idea how to create some magic mixin which create CSS3 Keyframe animation including CSS3 Browser Vender Prefix?

Once I tried to use some Mixin available over the internet, it product something like following where I want to ignore all other browser related code while it was under @-webkit.

@-webkit-keyframes hover-icon { 0% { -webkit-transform: scale(1, 1); -moz-transform: scale(1, 1); -ms-transform: scale(1, 1); -o-transform: scale(1, 1); transform: scale(1, 1); } 100% { -webkit-transform: scale(1.2, 1.2); -moz-transform: scale(1.2, 1.2); -ms-transform: scale(1.2, 1.2); -o-transform: scale(1.2, 1.2); transform: scale(1.2, 1.2); } }

Here’s what you’re looking for: http://leaverou.github.io/prefixfree/

Thanks Chris! Any time I have a question – you always there. Thanks again!!

Thank you very much for the demo. This solved my animation problems I had in IE.

This is a massively important detail:

Internet explorer does not get to claim the w3 standards' spec. That needs to be updated. No seriously- Internet Explorer has been the bane of every web developers' existence for entirely too long for them to get to claim that comment. It should not be confused, even for a second, that Internet Explorer is the reason that any modern css standard exists. That comment should be

[1]: Firefox 16+, IE 10+.

Hi Chris, you can use -webkit-animation-delay to delay the effects. It will much better if you can add demo as well.

Hey I need to do ripple effect with 4 circles. Could you guys please help me on it using css3. Thanks in advance.

I don’t think it works that way

Thanks so much for this. Works great with HTML5 video, just got to remember that -webkit prefix

Is there a reason why you don’t condense the repetitive declarations into a comma-separated selector? i.e:

It’d be good if we could bundle the various @keyframes like CSS selectors (to save on repetition):

@keyframes A, @-webkit-keyframes A, @-o-keyframes A { … opacity: 1; … /*no [-*-]transform…*/ } But apparently it does not work, it isn’t recognized, I tried… (Just in case someone else has the same idea)

It works fine on all browsers with zoom in and out. But on safari (all versions) when you zoom in and out and it breaks. Anybody suggest any fix PLEEEAAASSSSSEEEEE. Thanx many and many thanx.

forgot to put the demo link here which you have posted up there: http://jsfiddle.net/simurai/CGmCe/light/ Any FIX for the SAFARI ZOOM IN or OUT

This keyframe slide is an intresting tool but I can only find examples of it in use with images. Is it possible to use this framework for sliding color change across text only? I mean text like in a p or span class sentence where it only affects the text with no background effects.

Seems like Safari 9 has issue with “0% {}” when declaring keyframes. So if we have simple animations that have only 2 keyframes we can use ‘from’ and ‘to’. or we can simply use ‘0’ instead of ‘0%’

Hi Chris. Is it fine to include a number in the keyframe naming declaration? e.g. banner_circle_10, banner_circle_11 or should it be banner_circle_ten, banner_circle_eleven, etc. ?

This doesn’t really explain anything… it just gives some examples without explaining how or why they work, or what the rules are. What is keyframes and when/why would you use it? How?? What??

Anything specific you need to know? There is place to ask questions…

Thanks for that! Still a very valid and useful article!

Allow me to point out a little error in the properties table: “duration-count” should be “iteration-count”

I m using text resizing and it doesn’t animate correctly in Safari, text blinks to a large size before animating. Are there some glitches with Safari regarding animating font-size? No issue with Firefox or Chrome.

Does it matter at all if the @keyframes are declared before or after where they called from?

I’m 99% sure it does not.

Hey! This is just what I was looking for. It works great in Chrome, but I’m getting nothing in IE11. I put some text in the div–the text shows, but there is NO background color or frame around it.

Is there something I’m missing in the HTML file to make it work?

Are there ways to declare frames without from/to or percentages? Really annoying to have to calculate the percentages and change them all when you add a frame.

Hi guys I’m using the animation that is from 0deg to 360deg and so the object rotates 360 degrees but I want it rotate as an animation while it stays still not moving around in a circle, any tips?

I’m using transform keyframes animation. However, the section above and this section are parallax, which when the zoom happens is creating a weird space at the bottom. Even with heights set, overflow:hidden, etc. Does this just not work well with multiple parallax sections?

Is there or was at some point any reason to use the prefixes on the “keyframes” key? I can’t find any prefixer that will add the “keyframes” prefixes, including sass.

Can you use this to animate the percentage steps on a graduated background?

I’m trying to make a “gleem” effect that has a thin graduated white highlight run from one side of a LI item to the other like someone’s headlights lights up your room when they pull into the driveway at night.

Hello, may I ask how to stop animation at certain keyframe and keep on next one? if I want to connect several animations one after another – i need to stop the one and keep on the next

Thank very much, amazing article, at last i’ve found best CSS animation tutorial.

If a keyframe includes transform, how should it be prefixed? According to ShouldIPrefix.com, we only need -webkit with animation and we need -webkit and -ms with transform.

I have never witnessed such kind of a tutorial ever before. Have to say the kind of visuals it presents over here is enough to make a guy understand animation coding.

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

Save my name, email, and website in this browser for the next time I comment.

Copy and paste this code: micuno *

Leave this field empty

Tutorial Playlist

Css tutorial, the best guide to understand css selectors, the ultimate guide to css background image, the best guide to understand css colors, your one-stop guide to master the display property in css, css box model, css grid layout: the best guide to understand grid layout, css flexbox: the best guide to understand flex model, css grid vs. flexbox: a tutorial to understand the key differences, a beginner's guide on how to create a navbar in css, css keyframes: a brief introduction.

CSS Hover Effect - An Introduction

Everything You Need to Know About CSS Animation

A Tutorial to Learn Some Useful CSS Effects for Your Webpage

Learn how to add css transitions to your webpage, position elements on a web page using css positioning, what is css responsive web design and how to implement it.

CSS Tricks: Five Tricks To Enhance Your Web Page

CSS Advanced Tutorial to Understand the A-Z Of CSS

Solana Crypto: The Rising Star of the Crypto-Market

What is anova understanding the fundamentals of anova, all you need to know about c++ memory management.

Everything You Need to Know About CSS

Lesson 10 of 22 By Aryan Gupta

Table of Contents

CSS stands for Cascading Style Sheets . It is a style sheet language that is used to describe the look and formatting of a document written in a markup language. It provides an additional feature to HTML, thus making it very useful. It is generally used with HTML to change the style of web pages and user interfaces. In this article, we will try to understand an important part of CSS animation that is Keyframes.

What Are CSS Keyframes?

In CSS, keyframes are used for animations. It gives you more control over the animation you want to perform. The animation is created by gradually changing from one style to another. You can change the CSS styles as many times as you want.

@keyframes [name]{

      from{

        [styles];

• The [name] value defines the name of the animation. You can name it anything you want.
• The [styles] value defines the CSS style properties.

Keep in mind that these do not have any selectors on them. They are at the root of the CSS page.

Basics to Advanced - Learn It All!

Keyframes Timing Functions

The timing function is used to control the rate of animation. We can use the following values for better understanding:

• Linear: It means the transition will be constant from start to end.
• Ease:  It means that the animation will start slowly, and then after a time period, the speed will increase, and before the end speed will slow down again.
• Ease-in: It is very similar to ease, but the animation will end quickly.
• Ease-out: It is very similar to ease. The only difference is that it starts fast.

CSS Keyframes Example 1

Let’s look at a basic example to understand how keyframes work. 

Here, we have created a box with a height and width of 200px. The color of the box is red. A keyframe named ‘myframes’ is created which will increase the height of the box all the way to 500px.

Animation name specifies the name of @keyframes defined animations that should be applied to the selected element.

Animation duration specifies how many seconds or milliseconds that animation takes to complete one cycle of the animation.

The timing function specifies how the animation will progress over the duration of each cycle i.e. the easing functions.

Iteration count specifies the number of times an animation cycle should be played before stopping.

CSS Keyframes Example 2

In this example. we will see another simple way to write keyframes with multiple style declarations.

CSS Keyframes Example 3

In this example, we will see multiple keyframes with multiple style declarations.

Here's How to Land a Top Software Developer Job

Combine Transform and Keyframes

In this example, we have combined the keyframes with the transform property to create a rotating box.

With this, we come to the end of this CSS keyframes article. We discussed several examples and animation properties to understand the working of keyframes. By working with different keyframes and animation properties, you will be able to understand the concepts better. Get in touch with experts by enrolling in Simplilearn's Post Graduate Program in Full Stack Web Development course.

If you have any questions or feedback regarding our CSS keyframes article, let us know in the comments section. Our experts will get back to you as soon as possible.

Find our Full Stack Java Developer Online Bootcamp in top cities:

About the Author

Aryan is a tech enthusiast who likes to stay updated about trending technologies of today. He is passionate about all things technology, a keen researcher, and writes to inspire. Aside from technology, he is an active football player and a keen enthusiast of the game.

Recommended Resources

CSS Preprocessors

• PMP, PMI, PMBOK, CAPM, PgMP, PfMP, ACP, PBA, RMP, SP, and OPM3 are registered marks of the Project Management Institute, Inc.

CSS Reference

Css properties, css @keyframes rule.

Make an element move gradually 200px down:

More "Try it Yourself" examples below.

Definition and Usage

The @keyframes rule specifies the animation code.

The animation is created by gradually changing from one set of CSS styles to another.

During the animation, you can change the set of CSS styles many times.

Specify when the style change will happen in percent, or with the keywords "from" and "to", which is the same as 0% and 100%. 0% is the beginning of the animation, 100% is when the animation is complete.

Tip: For best browser support, you should always define both the 0% and the 100% selectors.

Note: Use the animation properties to control the appearance of the animation, and also to bind the animation to selectors.

Note: The !important rule is ignored in a keyframe (See last example on this page).

Browser Support

The numbers in the table specifies the first browser version that fully supports the rule.

Numbers followed by -webkit-, -moz- or -o- specify the first version that worked with a prefix.

Advertisement

Property Values

More Examples

Add many keyframe selectors in one animation:

Change many CSS styles in one animation:

Many keyframe selectors with many CSS styles:

Note: The !important rule is ignored in a keyframe:

Related Pages

CSS tutorial: CSS Animations

COLOR PICKER

Contact Sales

If you want to use W3Schools services as an educational institution, team or enterprise, send us an e-mail: [email protected]

Report Error

If you want to report an error, or if you want to make a suggestion, send us an e-mail: [email protected]

Top Tutorials

Top references, top examples, get certified.

• Open access
• Published: 29 September 2016

Towards key-frame extraction methods for 3D video: a review

• Lino Ferreira   ORCID: orcid.org/0000-0003-0648-6067 1 , 2 , 4 ,
• Luis A. da Silva Cruz 2 , 3 &
• Pedro Assuncao 1 , 4  

EURASIP Journal on Image and Video Processing volume  2016 , Article number:  28 ( 2016 ) Cite this article

3552 Accesses

6 Citations

Metrics details

The increasing rate of creation and use of 3D video content leads to a pressing need for methods capable of lowering the cost of 3D video searching, browsing and indexing operations, with improved content selection performance. Video summarisation methods specifically tailored for 3D video content fulfil these requirements. This paper presents a review of the state-of-the-art of a crucial component of 3D video summarisation algorithms: the key-frame extraction methods. The methods reviewed cover 3D video key-frame extraction as well as shot boundary detection methods specific for use in 3D video. The performance metrics used to evaluate the key-frame extraction methods and the summaries derived from those key-frames are presented and discussed. The applications of these methods are also presented and discussed, followed by an exposition about current research challenges on 3D video summarisation methods.

In the last years, new features have been implemented in video applications and terminal equipments due to users demand, who are always seeking for new viewing experiences more interactive and immersive, such as those provided by 3D video. This new visual experience is created by depth information that is part of 3D video and absent in classic 2D video. The inclusion of depth information in video signals is not a recent innovation, but the interest in this type of content and aspects related to it, such as acquisition, analysis, coding, transmission and visualisation, have been increasing recently [ 1 , 2 ]. Lately, 3D video has been attracting attention from industry, namely content producers, equipment providers, distributors and from the research community mostly on account of the improvements in Quality of Experience that it provides to viewers [ 3 ], as well as due to the new business opportunities presented by this emerging multimedia format.

In the past, video repositories were relatively small so that indexing and retrieval operations were easy to perform. More recently, the massification of 3D video and its applications have resulted in the generation of huge amounts of data, increasing the need for methods that can efficiently index, search, browse and summarise the relevant information with minimum human intervention. Furthermore, 3D video description and management is also required to enable quick presentation of the most important information in a user-friendly manner [ 4 , 5 ]. Video summarisation is a video-content representation method that can fulfil these requirements. In contrast to summarisation of 2D video, which has been the subject of a significant amount of research, 3D video summarisation is still a relatively unexplored research problem which deserves more attention.

A video summary is a short version of a full-length video that preserves the essential visual and semantic information of the original unabridged content. In the video summarisation process, a subset of key-frames or a set of shorter video sub-sequences (with or without audio) are chosen to represent the most important segments of the original video content according to predefined criteria [ 4 ]. This video content representation can be used in the promotion of movies, TV channels or other entertainment services. Video summarisation can also be used for content adaptation operations in constrained communication environments where bandwidth, storage capacity, decoding power or visualisation time is limited [ 6 ].

The literature defines two types of video summaries, namely those based on key-frames and those comprised of video skims [ 7 ]. A video summary based on key-frames is made up of a set of relevant frames selected from the video shots obtained from the original video. This type of summary is static, since the key-frames, being temporally distant and non-uniformly distributed, do not enable adequate rendering/reproduction of the original temporal evolution of the video content. Here, the video content is displayed in a quick and compact way for browsing and navigation purposes, without complying with timing or synchronisation requirements. Video skims are usually built by extracting the most relevant temporal segments (with or without audio) from the source sequence. After the extraction, all temporal segments are concatenated into sequential video with much shorter length than the source sequence. The methods used for computation of key-frames and video skims summaries are quite distinct, but these two types of representations for video content can be transformed from one to the other. A video skim can be generated from a key-frame summary by adding frames or segments that include the key-frames, while a key-frame summary can be created from a video skim by uniform sampling or selecting the most representative frame from each video skim segment [ 4 ].

In regard to 3D video content, a detailed study of the existing scientific literature reveals that comprehensive comparative studies of 3D video summarisation methods are missing. To help filling this gap, this paper presents a review of 3D video summarisation methods based on key-frames. This overview of the current state-of-the-art is mainly focused on the methods and features that are used to generate and evaluate 3D video summaries and not so much on the limitations or performance of specific methods. Since experimental set-ups, 3D video formats and features used for summarisation are considerably different from one computational method to another, a fair comparative analysis of the results, advantages and shortcomings of all methods is almost impossible. This paper also identifies open issues to be investigated in the area of 3D key-frame extraction for summarisation.

The remainder of the paper is organised as follows. Section 1.1 presents the existing 3D video representation formats and relevant features for the purpose of summarisation; then, in Section 1.3 , the generic framework normally used in 3D key-frame extraction methods is presented, after which Section 1.4 reviews the most important shot boundary detection (SBD) methods for 3D video. Then, Section 1.5 characterises the relevant methods used in 3D key-frame extraction for summarisation while Section 1.6 addresses common methods used for presentation of key-frames. Section 1.7 describes performance evaluation methods suitable for 3D video summaries based on key-frames, and Section 1.8 describes some applications of this kind of summaries. Section 1.9 discusses the prospects and challenges of the 3D key-frame extraction methods, and finally Section 2 concludes the paper.

1.1 3D video representation formats

In this review article, ‘3D video’, is defined as a representation format which differs from 2D video by the inclusion of information that allows viewers to perceive depth. This depth information can be conveyed either indirectly via two or more views of the scene (e.g. left and right views) or explicitly through either depth maps of geometric representation of connected 3D points and surfaces.

The most common formats used used to represent 3D visual scenes include natural video and/or geometric representations.

Stereoscopic video is composed of two slightly shifted video views of the same scene, where one corresponds to what would be observed by a left eye and the other by the right eye of a human observer. Since these are two views of the same scene, the corresponding images are related by the binocular disparity, which refers to the difference in the image plane coordinates of similar features captured in two stereo images. The scene depth is perceived from the disparity when using stereoscopic displays and can also be computed for different types of computer vision applications (e.g. measuring distances in 3D navigation).

Multiview video (MVV) is composed of more than two video views shifted in the vertical and/or horizontal position. Typically, MVV acquisition is done using an array of synchronised cameras with some spatial arrangement, which capture the visual scene from different viewpoints. The MVV format is useful for applications supported by autostereoscopic displays with or without head tracking, which render a denser set of 3D views that are displayed through lenticular and parallax barriers. With this type of display, viewers are able to see the portrayed scene from different angles by moving the head along a horizontal plane. A typical application of this video format is freeview navigation where users are given the option of freely choosing the preferred viewpoint of the scene.

Video-plus-depth (V+D) is composed of a video signal (texture) and respective depth map. Each value of the depth map represents the distance of the object to the camera for the corresponding pixel position. Typically, the depth information is quantised with 8 bits, where the closest point is represented with value 255 and the most distant point is represented with 0. Additional virtual views (i.e. not captured) of the same scene imaged can be synthesised from the V+D original information by using 3D warping transformations. Several different applications and services can benefit from the V+D format, due to its inherent backward compatibility with 2D video systems and higher compression efficiency achievable when compared to stereoscopic video. For instance, 3DTV services can be seamlessly deployed while maintaining compatibility with legacy 2D video services.

Multiview video-plus-depth (MVD) is composed of video and depth maps for more than two views of the specific scene. The depth information can be computed from different views or captured directly using time-of-flight (ToF) sensors. MVD can be used to support dense multiview autotereoscopic display in a relatively efficient manner. From a relatively small set of different views and corresponding depth maps, a much larger set of views can be synthesised at the display side, avoiding coding and transmission of a great deal of data while enabling smooth transitions between viewpoints. Several emerging applications such as free viewpoint video and free viewpoint TV will use the MVD format due to its compact representation of 3D visual information. Mixed-reality applications and gaming are also important application fields for MVD.

3D computer graphics use a geometry-based representation, where the scene is described by a set of connected 3D points (or vertices), with associated texture/colour mapped onto them. The data content of this format can be organised into geometry, appearance and scene information [ 8 ]. The geometry includes the 3D position of vertices and polygons (e.g. triangles) that are constructed by joining these vertices. The appearance is an optional attribute which associates some properties (e.g. colour, texture coordinates) to the geometry data. Finally, the scene information includes the layout of a 3D scene with reference to the camera (or view), the light source and description of other 3D models if they are present in the scene. 3D computer graphics can provide better immersive and interactive experience than conventional 2D video, since the user is provided more freedom to interact with the content and get a realistic feeling of ‘being there’. Relevant applications can be found in quite different fields, such as medicine, structural engineering, automobile industry, architecture and entertainment.

Plenoptic video is composed of a very large of the number of views (e.g. hundreds or thousands) captured simultaneously. This multiple view acquisition process can be interpreted as a partial sampling of the plenoptic function [ 9 ], which represents not only spatial or temporal information but also angular information of about the captured light rays, i.e. captures a segment of the whole observable scene represented by a light field. In practice a 3D plenoptic image is captured by a normal image sensor placed behind an array of uniformly spaced semi-spherical micro-lenses. Each micro-lens works as an individual low resolution camera that captures the scene from an angle (viewpoint) slightly different from that of its neighbours. Plenoptic video, also known as light field video, is an emerging visual data representation with known applications in computational photography, microscopy, visual inspection and medical imaging among others.

1.2 3D video features for summarisation

The scene depth is the additional information that is either implicitly or explicitly conveyed by 3D video formats. Therefore, depth is also the signal component that mostly contributes to distinguish 3D video summarisation methods from those used for 2D video. One of the first works combining depth with features of 2D video to summarise 3D video was done by Doulamis et al. [ 10 ]. The authors proposed an algorithm jointly operating on both the depth map and the left channel image to obtain a feature vector for use in video segmentation and key-frame extraction methods. The feature vector including segment size, location, colour and depth. Another important feature is the depth variance associated with the temporal activity, which was used by Ferreira et al. in [ 11 ] for temporal segmentation of 3D video. The average stereo disparity per frame and temporal features such as image difference and histogram difference are computed and combined in feature vectors which in turn are used by a clustering algorithm to partitioning 3D video in temporal segments. Another work using frame intensity histogram distributions as features and the Jensen-Shannon difference to measure frame difference in feature space is presented in [ 12 ]. This is used to segment a video clip into shots, and then to choose key-frames in each one. More recently, Papachristou et al. in [ 13 ] also segmented 3D video using low-level features obtained from disparity, colour and texture descriptors computed from histograms and wavelet moments. An improved three-dimensional local sparse motion scale invariant feature transform descriptor is used in [ 14 ], for RGB-D videos, based on grey pyramid, depth pyramid and optical flow pyramids that are built for both colour frames and depth maps. Point features are determined with a SIFT descriptor are combined with depth information of the point as well as optical-flow derived motion information. Although this work is focused on gesture recognition, the features and similarity measures may also be used in key-frame extraction methods. This work presented in [ 15 ], uses vectors containing moments (mean, standard deviation, skewness and kurtosis) of signature profiles of blocks with variable size for the luminance and disparity frames. A descriptor of frame moments was developed for summarising stereoscopic video through key-frame extraction and also to produce stereoscopic video skims. Yanwei et al. in [ 16 ] proposed a multiview video summarisation method which uses low-level and high-level features. The low-level features are based on visual attribute of video, as colour histogram, edge histogram and wavelet while for high-level features the authors use the Viola-Jones as face detector in each frame.

Geometric features are also relevant for temporal segmentation and extraction of key-frames in 3D visual information. For instance, Assa et al. in [ 17 ] proposed a method to produce an action synopsis of skeletal animation sequences for presenting motion in still images. The method selects key-frames based on skeleton joints and their associated attributes (joint positions, joint angles, joint velocities, and joint angular velocities). In [ 18 ], the authors also use geometric features, such as the number and location of vertices of surface, to produce video summaries of animation sequences. Other geometric features used by Jianfeng et al. in [ 19 ] to summarise 3D video are features vectors formed by the histograms of the vertices in the spherical coordinate system. A different type of features rely on 3D shape descriptors. For instance, Yamasaki et al. in [ 20 ] used the shape of 3D models to split the video in different motion/pose temporal segments. Relevant shape features, such as shape histograms, shape distribution, spin image and spherical harmonics were studied in [ 21 ], where the performance of shape similarity metrics is evaluated for applications in 3D video sequences of people. Since similarity measures are of utmost importance in summarisation this is a relevant work in the context of this paper. Another type of features used in [ 22 ] for key-frame extraction is based on deformation analysis of animating meshs while the vertices positions in mesh models and motion intensity were used by Xu et al. in [ 23 ] in temporal segmentation of 3D video.

1.3 3D key-frame extraction framework

Summarisation of 3D video follows a generic processing chain that is extended from 2D video by considering the inherent depth and geometric information as relevant feature contributors for selecting the dominating content in 3D moving scenes. A possible approach is based on clustering by grouping similar frames according to some similarity measure [ 24 ], without any prior processing or feature extraction. However, a more generic and systematic approach that better suits the problem of 3D video summarisation follows the three-step framework of Fig. 1 , where the entire video sequence is first divided into video shots based on scene transitions using an SBD method, followed by a key-frame extraction method applied to each video shot to extract the most representative frames, based on the specific properties of the video content and similarity measures. Finally, the extracted key-frames are either presented to the viewers or stored, following some predefined presentation structure.

A conceptual framework for key-frame summarisation

Following the conceptual framework shown in Fig. 1 , the input video is segmented into video shots, mostly based on spatio-temporal criteria, but other criteria can be used such as based on motion [ 20 , 25 ] or the combination of the temporal and depth features [ 11 ]. More details can be found in Section 1.4 . After this segmentation, one or more key-frames are extracted from each video shot according to user-defined parameters, or based on specific requirements (in Fig. 1 , only one key-frame is extracted). The most relevant key-frame extraction methods are presented in Section 1.5 . Once the key-frames are extracted, they need to be presented in an organised manner for easy viewing during video browsing or navigation. In this framework, three key-frame presentation methods are described, static storyboard, dynamic slideshow and single image based on stroboscopic effect, but other methods can be found in the literature (see Section 1.6 ). The key-frame presentation methods are independent of the key-frame extraction operation and thus the same key-frame summary can be presented to viewers in different ways.

1.4 Shot boundary detection

In the recent past, development of SBD methods for 2D video received a lot of the attention from the video processing research community. However, very few works have investigated the SBD problem in the context of 3D video, especially taking into account depth information. Relevant surveys of video SBD methods with specific application in 2D video can be found in the literature [ 26 – 28 ]. In this section, we briefly introduce the main concepts behind these methods for 2D video. Then, the most promising and better performing SBD methods used for 3D key-frame extraction are explained in detail.

A video segment can be decomposed into a hierarchical structure of scenes, video shots and frames, with the linear video first divided into video scenes, which may comprise one or more video shots (set of correlated frames). A video scene is defined as a set of frames which is continuous and temporally and spatially cohesive [ 29 ], while a video shot may also be defined by camera operations, such as zoom and pan. Thus, the video shot is the fundamental unit in the content structure of a video sequence. Since its size is variable, the identification of start and end of the video shots is done using specific SBD methods.

Figure 2 presents a generic framework of a SBD method. While this framework is similar for both 2D and 3D video, the actual algorithms used for each type of content are not the same due to the difference in their relevant features. Firstly, the relevant visual features are computed, in general forming feature vectors for each video frame as described in Section 1.2 . In the second step, the visual features of consecutive frames are compared using specific similarity measures some decision criteria are used to identify shot boundaries. The decision methods used to find shot boundaries can be based on static thresholds (as in Fig. 2 ), adaptive thresholds (thresholds depend on the statistics of the visual features used), B-splines fittings [ 30 ], support vector machines (SVM) [ 31 ] and K-means clustering [ 11 ]. The detection accuracy of SBD methods is improved by combining several visual features [ 32 ].

A generic diagram of SBD framework

Video shot boundaries can be classified into two types: abrupt shot boundary (ASB) (as in Fig. 2 ) and gradual shot boundary (GSB), according to a certain classification of scene transition, which in general is related to content variation over time. This is common in 2D and 3D video, despite the fact that scene transition in 3D video may include depth changes besides the visual content itself. In ASB, the scene transition occurs over very few frames, usually a single frame defines the boundary. In the case of GSB, the transition takes place gradually over a short span of frames. The most common gradual transitions are fade-ins, fade-outs, dissolves and wipes [ 26 – 28 ]. A common problem in SBD is the correct discrimination between camera operations and object motion that originate the gradual transitions, since the temporal variation of the frame content can be of the same order of magnitude and take place over the same number of frames. This similarity of visual effects caused by camera operations and object motion can induce false detections of gradual shot boundaries. This problem is aggravated for video sequences with intense motion.

1.4.1 SBD methods

Doulamis et al. in [ 10 ] proposed a key-frame extraction method for stereo video which includes a SBD method. Here, the entire video sequence is divided into video shots using an algorithm based on the analysis of DC coefficients of compressed videos, following the solution proposed in [ 33 ]. More recently, Papachristou et al. in [ 13 ] presented a framework for stereoscopic video shot classification, that uses a well-known method designed for 2D video to segment the original stereoscopic video into shots [ 34 ]. However, this method was applied only to the colour channels of the videos to be summarised. Ferreira et al. [ 11 ] proposed an algorithm to detect 3D shot boundaries (3DSB) based on a joint depth-temporal criterion. The absolute frame difference and sum of absolute luminance histogram difference are used as the relevant measures in the temporal dimension, while in the depth dimension, the variance of depth in each frame is used. A K-means clustering algorithm that does not require training and does not use thresholds is applied to choose the 3DSB transition frames. Ferreira’s method is independent of the video content and can be applied to 2D or 3D video shot boundary identification. In the case of the 2D video, absolute frame differences and sum of absolute luminance histogram difference are used.

Some methods target segmentation of 3D mesh sequences using properties of 3D objects as the shape and motion/action (e.g. human body motion, raise hands) to detect the shot boundaries. Yamasaki et al. [ 20 ] proposed a temporal segmentation method for 3D video recordings of dances, which is based on motion speed, i.e. when a dancer/person changes motion type or direction, the motion becomes small during some short period and in some cases it is even paused for some instants, according to the type of dance. To seek the points where motion speed becomes small the authors used an iterative close point algorithm proposed in [ 35 ] which is employed in the 3D space (spherical coordinates). In contrast to conventional approaches based on thresholds, the authors devised a video segmentation scheme appropriate for different types of dance. In this scheme, each local minimum is compared with local maxima occurring before ( l m a x bef ) and after ( l m a x aft ) the local minima. When l m a x bef and l m a x aft are 1.1 times larger than the local minimum, a temporal segmentation point is declared to occur at the minimum location. Since the decision rule is not based on absolute values and thresholds, rather on relative values of extrema, it is more robust to data variation (like type of dance) and no empirically derived decision thresholds are used.

Another method which uses the motion speed of the 3D objects was presented by Xu et al. [ 23 ]. To reduce computation time of motion information the authors used the point distance (DP) instead of vertices position in Cartesian coordinates. DP is defined as Euclidean distance between one fixed point and all 3D objects’ vertices coordinates of each frame. Figure 3 shows the point distance for 2 frames of Batter’s sequence. Before determination of scene transitions, the histogram of point distance of each frame is calculated.

Point distance of the frames #38 and #39 of Batter sequence. Grey values means the point distance from (0,0,0) [ 23 ]

To detect abrupt and gradual transitions of 3D video, the Euclidean distance between the histograms of point distance and three thresholds are used, where the threshold values were derived empirically.

Ionescu et al. [ 36 ] used a histogram-based algorithm specially tuned for animated films to detect ASB. From GSB only fades and dissolves are detected, since they are the most common gradual transition. The GSB detection is done using a pixel-level statistical approach proposed by [ 37 ]. The authors proposed the Short Colour Change (SCC) detection algorithm to reduce the cut detection false positives. The SCC is the effect that accompanies short term frame colour changes, caused by explosion, lightning and flash-like visual effects. More recently, Slama et al. [ 38 ] proposed a method based on the motion speed to split a 3D video sequence into segments characterised by homogeneous human body movements (e.g. walk, run, and sprint). However, the author only considers as significant video shot transition indicators changes of type of movement. Here, video shots with small differences from previous shots and small number of frames are avoided. The motion segmentation used in this work is based on the fact that when humans modify the motion type or direction, the motion magnitude decreases significantly. Thus, finding the local minimum of motion speed can be used to detect the break point where human body movements changes and consequently to segment the entire video into shots.

1.4.2 Evaluation metrics

Three well-known performance indicators are used in the evaluation of the SBD methods for 2D video: recall rate (R), precision rate (P) [ 39 ] and accuracy measure F1 [ 40 ]. The computation of these values is based on the comparison of manual segmentation (ground-truth) and computed segmentation. If a ground-truth is available these metrics can be applied to 3D video SBD methods.

Recall rate is defined as the ratio between the number of shot boundaries detected by an algorithm and the total number of boundaries in the ground-truth dataset (see Eq. ( 1 )). Precision rate, computed according to Eq. ( 2 ), is defined as the ratio between the number of shot boundaries detected by an algorithm and the sum of this value with the number of false positives. F1 is a measure that combines P and R, see Eq. ( 3 ).

where D is the number of shot boundaries correctly detected by the algorithm, D M is the number of missed boundaries and D F is the number of false detections. For good performance, the recall and precision rates should have values close to 1. The best performance is reached when F1 is equal to 1, while the worst occurs at 0.

The recall rate, precision rate and measure F1 were used to evaluate the performance of temporal segmentation methods for 3D video in [ 11 , 23 , 38 ], while Yamasaki et al. [ 20 ] only used recall and precision rates in the evaluation process. Although, these 3D SBD methods used the same evaluation metrics, the comparison of the results and performance obtained from such SBD methods is not possible because different datasets were used.

1.4.3 Discussion

Since the major difference between 2D and 3D video is the implicit or explicit availability of depth information, the visual features used in the SBD methods for 3D video must take depth into account, i.e. the temporal segmentation must also consider depth information in order to use depth discontinuities in shot detection. Until now, most research works on SDB for 3D video, did not use the depth information in the detection process. For example, Doulamis et al. in [ 10 ] proposed a key-frame extraction method for stereo video which includes a SBD method. However, this algorithm does not take into account the depth information of the stereo video and it is only applied to one view of the stereo sequence, for instance the left view. Another drawback of Doulamis’ work is the lack of performance evaluation of the proposed temporal segmentation method. A method to segment stereo video was proposed in [ 13 ], but the proposed procedure does not take depth into account either.

In [ 20 , 23 , 38 ], the authors proposed SDB methods for 3D video, which are only applicable to 3D mesh models and require modifications to be used with most common pixel-based 3D video formats, like stereo or video-plus-depth. Finally, Ferreira et al. [ 11 ] proposed a method which uses the depth and temporal information for automatic detection of 3D video shots from the 3D video sequence that uses a K-means clustering algorithm to locate the boundaries. This algorithm has the advantage of not using any explicit thresholds or training procedure.

A common problem with the 2D video SBD methods described in the literature is the lack of common comparison grounds, as few works use the same dataset to test the methods proposed and evaluate their performance. This is a serious problem as it limits the number of comparisons that can be made to compare the different SBD methods. For the 3D case, the lack of comparative analyses is even more severe, due to the reduced number of SBD methods developed so far specifically for this type of visual information. The few works that have been proposed for SBD in 3D video usually use the Recall and Precision rate to evaluate performance, but the lack of benchmark 3D video sequences with ground-truth shot segmentations severely limit the number and types of performance evaluations that can be made. As mentioned above, the evaluation metrics presented in Section 1.4.2 are based on comparison between manual and computed segmentation. Therefore, besides being very important to have common test datasets, the development of universal and objective measures, which are specific for SBD and can be applied in different content domains and 3D video formats is highly recommended and desired.

1.5 3D key-frame extraction

In this section, we briefly introduce the main concepts behind key-frame extraction methods for 2D video and describe key-frame extraction methods for 3D video. The key-frame extraction methods under review are grouped into seven categories: non-optimised, clustering, minimum correlation, minimum reconstruction error (MRE), curve simplification, matrix factorisation and other methods.

1.5.1 Non-optimised methods

The simplest method for 3D key-frame summarisation is uniform sampling (UnS). This method selects key-frames at regular time-intervals (see Fig. 4 a ), e.g. selecting one video frame every minute to be a key-frame. This will result in a set of key-frames evenly distributed throughout the video. However, the selected key-frames might not contain meaningful or pertinent visual content or there may be two or more similar key-frames. For instance, the selected key-frame might show a bad image (e.g. unfocused) or no key-frame exists for some video shots, thus failing to effectively represent the video content.

a UnS method: uniform sampling at equal intervals. b PoS method: selecting the first frame of each video shot

Another simple and computationally efficient frame selection method is position sampling (PoS). In PoS, once the boundaries of a video shot are detected, the method selects frames according to their position in the video shot, and e.g. the first, or the last or the middle frame of the video shot (see Fig. 4 b ) can be chosen as key-frames. Thus, the size of key-frame summary corresponds to the number of video shots of the entire video. In some summarisation applications, one key-frame per video shot is not enough, and the PoS method can be adapted to this need by allowing the selection of multiple frames at fixed positions within the video shot. For 3D video, UnS and PoS are used mostly as references for comparisons with other methods, as in [ 18 , 41 , 42 ]. Ionescu et al. [ 36 ] selected as key-frames the frames in the middle of the video shot to reduce temporal redundancy and computation cost of his animation movies summarisation method. Yanwei et al. [ 16 ] used the middle frame of each video skim segment to represent this summary in a storyboard. In general the above non-optimised methods may be used in both 2D and 3D video with minor adaptations.

1.5.2 Clustering

Clustering can be used to partition a large set of data into groups minimising intra-group variability and maximising inter-group separation. After partitioning, the data in each cluster have similar features. The partitioning can be based on the similarities or distances between the data points where each datum represents a vector of features of a frame. These points are grouped into clusters based on feature similarity and one or more points from each cluster are selected to represent the cluster, usually the points closest to the cluster centre. The representative points of the clusters can be used as key-frames for the entire video sequence. A significant number of clustering methods reported in the literature use colour histograms as the descriptive features, and the clustering is performed using distance functions such as Euclidean distances or histogram intersection measures. These methods are very popular due to its good clustering robustness and the simplicity inherent to computing colour histograms [ 43 , 44 ]. Other features can also be used in clustering-based methods. For example, Ferreira et al. in [ 11 ] used temporal and depth features with a clustering algorithm to segment 3D video sequences into 3D video shots.

K-means is one of the simplest algorithms used to solve the clustering problem. This clustering algorithm can be applied to extract key-frames from short video sequences or shots, but its application to longer video sequences must be done with care taking into account the large processing time and memory requirements of the algorithm. To reduce the number of frames used by the clustering algorithm some authors pre-sample the original video, as proposed in [ 24 ]. The quality of the summaries may not be affected by this operation but the sampling rate must be chosen carefully. Although K-means is a popular and well-known clustering algorithm it has some limitations such as the need to pre-establish the number of clusters desired priori and the fact that the sequential order of the key-frames may not be preserved. Huang et al. [ 18 ] used the K-means clustering algorithm for extracting a set of 3D key-frames to be compared with the output of their key-frame extraction method.

1.5.3 Curve simplification

In the curve simplification method, each frame of the video sequence can be treated as a point in multidimensional feature space. The points are then connected in sequential order through an interpolating trajectory curve. The method then searches for a set of points which best represent the curve shape. Binary curve splitting algorithm [ 45 ] and discrete contour evolution [ 46 , 47 ] are two curve simplification algorithms used in the key-frame extraction methods. Curve simplification-based algorithms preserve the sequential information of the video sequence during key-frame extraction; however, the search for the best representation curve has high computational complexity. The curve simplification method proposed in [ 48 ] was used by Huang et al. [ 18 ] in the evaluation process of the 3D key-frame extraction method they proposed.

1.5.4 Minimum correlation

Minimum correlation based methods extract a set key-frames such that the inter-key-frame correlation is minimal, i.e. it extracts the key-frames that are more dissimilar from each other. Formally, the optimal key-frame extraction based on minimum correlation can be defined as

where Corr(.) is a correlation measure, n is the frame number of original sequence F , l i is the frame index in F and K is the set of resulting key-frames with m frames, \(K=\{f_{l_{0}},f_{l_{1}}, \ldots,f_{l_{m-1}}\}\) . Different algorithms can be used to find the optimal solution, such as logarithmic and stochastic search or a genetic algorithm [ 4 ]. The key-frame extraction method for stereoscopic video based on minimum correlation has been presented first by Doulamis et al. in [ 10 ]. Here, colour and depth information are combined to summarise stereo video sequences. After the segmentation of the entire video sequence, a shot feature vector is constructed based on size, location, colour and depth of each shot. To limit the number of shot candidates, a shot selection method based on similarly between shots is applied. Finally, the stereo key-frames are extracted from each of the most representative shots. The extraction is achieved by minimising a cross correlation criterion and uses a genetic algorithm [ 49 ]. Since this approach selects frames by minimisation of cross correlation, they are not similar to each other in terms of colour and depth.

1.5.5 Minimum reconstruction error

In MRE based methods, the extraction of the key-frames is based on the minimisation of the difference between the original video sequence/shot and the sequence reconstructed from the key-frames. A frame interpolation function \(\mathcal {I}(t,K)\) is used to compute the frame at time t , of the reconstructed sequence, from a set of key-frames K . The frame-copy method can be used to reconstruct the video sequence/shot (i.e. performing zero-order interpolation), but more sophisticated methods like motion-compensated interpolation might be used as proposed in [ 50 ]. The reconstruction error \(\mathcal {E}(\textbf {F},K)\) is defined as

where d (.) is the difference between two frames, F is video sequence/shot with n frames, F ={ f 0 , f 1 ,…, f n −1 }, where f i is the i th frame.

The key-frame ratio R ( K ) defines the ratio between the number of frames in the set K , m and the total number of frames in video sequence/shot F , n , i.e. R ( K )= m / n . Given a key-frame ratio constraint R m , the optimum set of key-frames K ∗ is the one that minimises the reconstruction error, i.e.

Thus, the MRE is defined by

For example, given a shot F with n =10 frames and a key-frame ratio R ( K )=0.2, this algorithm extracts at most 2 frames as key-frames, i.e. m =2.

Xu et al. in [ 19 ] presented a key-frame extraction method to summarise sequences of 3D mesh models, wherein the number and location of key-frames are found through a rate-distortion optimisation process. As in all shot-based methods, in Xu’s method shot detection is performed before key-frame extraction. Here, the SBD is based on the motion activity of a human body in dancing and sports videos. The motion activity is measured by the Euclidean distance between feature vectors of neighbouring 3D frames. The feature vectors are derived from three histograms (one for each spherical coordinate r , θ and ϕ ) of all vertices of the 3D frames. Before the computation of spherical histograms, the Cartesian coordinates of vertices are transformed to the spherical coordinates. One of the three histograms is computed by splitting the range of the data in equal size bins. Then, the number of points from the data set that fall into each bin is counted. After shot detection, the key-frames are extracted in each shot. The key-frame extraction method is based on a rate-distortion trade-off expressed by a Lagrangian cost function, cost(Shot k ) =Distortion(Shot k ) + λ Rate(Shot k ) where Rate is the number of key-frames in a shot and Distortion is the Euclidean distance between feature vectors.

Huang et al. [ 51 ] also presented a key-frame extraction method for 3D video based on rate-distortion optimisation, where Rate and Distortion definitions are similar to those used in [ 19 ]. However this method is not based on shot identification, since it produces 3D key-frame summaries without requiring prior video shot detection. The key-frame summary sought should minimise a Conciseness cost function, which is a weighted sum of the Rate and Distortion functions defined in the work. A graph-based method for extracting the key-frames is used, such that the key-frames selection is based on the shortest path in the graph that is constructed from a self-similarity map. The spherical histogram of the 3D frames is used to compute the self-similarity map.

More recently, Ferreira et al. [ 42 ] proposed a shot-based key-frame extraction method based on rate-distortion optimisation for 2D and 3D video. For each video-shot, a corresponding set of key-frames is chosen via dynamic programming by minimising the distortion between the original video shot and the one reconstructed from the set of key-frames. The distortion metric comprises not only information about frame difference, but also the visual relevance of different image regions as estimated by and aggregated saliency map, which combines three saliency feature maps computed from spatial, temporal and depth information.

1.5.6 Matrix factorisation

Another class of methods use matrix factorisation techniques to extract frames from a video sequence. Matrix factorisation (MF) techniques are based on approximating a high dimension matrix A (original data) by a product of two or more lower dimension matrices. The A matrix can be composed of different features of the video or image, e.g. Gong and Liu [ 52 ] used the colour histograms to represent video frames while, Cooper et al. [ 53 ] computed the MF of the similarity matrix into essential structural components (lower dimension matrices). In addition to dimension reduction, the MF techniques allow reducing significantly the processing time and memory used during the operation. The MF techniques found in these key-frame extraction methods include singular value decomposition (SVD) and non-negative matrix factorisation.

Gong and Liu [ 52 ] proposed a key-frame extraction method based on SVD. To reduce the number of frames to be computed before the SVD, only a subset is taken from the input video at a pre-defined sample rate. Then, colour histograms (RGB) are used to create a frame-feature matrix A of the pre-selected frames. Next, the SVD is performed on matrix A to obtain an orthonormal matrix V in which each column vector represents one frame in the defined feature space. Then a set of key-frames are identified by clustering the projected coefficients. According to user’s request, the output can be a set of key-frames (one of each cluster) or a video skim with a user specified time duration. To construct the set of key-frames, the frames that are closest to the centres of the clusters are selected as key-frames. Non-negative similarity matrix factorisation based on low-order discrete cosine transforms [ 53 ] and sliding-window SVD [ 54 ] are other approaches for key-frame extraction based on matrix factorisation.

In [ 18 ], Huang et al. proposed a method to be used with 3D video to represent an animation sequence with a set of key-frames. Given an animation sequence with n frames and m vertices of a surface in each frame, an n × m matrix A is built with the vertices coordinates. This matrix A is then approximately factorised into a weight n × k matrix W and a key-frame k × m matrix H , where k is the predefined number of key-frames. As k is selected to be smaller than n and m , this decomposition results in compact version of the original data A ≈ W H . An iterative least square minimisation procedure is used to compute the weights and extract the key-frames. This procedure is driven by user-defined parameters such as a number of key-frames and an error threshold. Lee et al. [ 22 ] introduced a deformation-driven genetic algorithm to search good representative animation key-frames. Once the key-frames are extracted, similar to [ 18 ], the animation is reconstructed by a linear combination of the extracted key-frames for better approximation. To evaluate the performance of the proposed method, the authors compare it with Huang’s method proposed in [ 18 ].

1.5.7 Other methods

The methods described in this section could not be easily classified into the preceding categories, mostly on account of the diversity of approaches followed in solving the key-frame extraction problem. As such, and given their importance, they are described all together in this section.

Assa et al. proposed a method to create an action synopsis image composed of key poses (human body motion) based on the analysis through motion curve. The method integrates several key-frames into a single still image or a small number of images to illustrate the action. Currently, it is applied in 3D animation sequences and 2D video as documented in [ 17 ].

Lee et al. [ 25 ] proposed a method to select key-frames from 3D animation video using the depth information of the animation. The extracted key-frames are used to compose a single image summary. The entire video sequence is divided into temporal segments based on the motion of the slowest moving objects, and then a summarisation method is applied to the segments. The depth information and the respective gradient (computed with depth values of each frame) is used to compute the importance of each frame. A single image summary composed of several foreground visual objects is built based on the importance of each frame. The authors proposed a threshold based approach to control the visual complexity (number of foreground objects) of the single image summary (one for each video sequence), as it is showed in Fig. 5 . By using this approach, the number of video frames to be analysed is reduced, but in some cases the method can miss important information contained in the temporal segments.

Single image key-frame presentation method [ 25 ]

Jin et al. [ 41 ] proposed a key-frame extraction method for animation sequences (skeletal and mesh animations). The method uses animation saliency computed from the original data to aid the selection of the key-frames that can be used to reconstruct the original animation with smaller error. Usually, an animation sequence is characterised by a large amount of information. For computational efficiency, the animation sequence is projected to a lower-dimensional space where all frames of the sequence are represented as points of curves defined in the new lower-dimensional space. Then, the curves in the lower-dimensional space are sampled and these sampled points are used to compute the Gaussian curvature values. Next, the points with the largest curvature value are selected as candidate key-frames. Finally, a key-frame refinement method is employed to minimise an error function which incorporates visual saliency information. The aim of a visual saliency is to identify the regions of an image which attract higher human visual attention. Lee Huang et al. [ 55 ] expanded this idea to 3D video and computed mesh saliency for use in a mesh simplification algorithm that preserves much information of the original input. More recently, visual saliency has also been used in 3D key-frame extraction, in the method proposed by Ferreira et al. in [ 42 ].

Yanwei et al. [ 16 ] proposed a multiview summarisation method for non-synchronised views, including 4 of them covering 360°, which results in small inter-view correlation, thus more difficult to compute similarity measures. In this method, each view is segmented into video shots and general solution combines features of different shots and uses a graph model for the correlations between shots. Due to the correlation among multi-view shots, the graph has complicated connectivity, which makes summarisation very challenging. For that purpose, random walks are used to do shot clustering and then the final summary is generated by a multi-objective optimisation process based on different user requirements, such as the number of shots, summary length and information coverage. The output of Yanwei’s method is a multiview storyboard, condensing spatial and temporal information.

1.5.8 Discussion

The problem of key-frame extraction for 3D video has been presented first by Doulamis et al. in [ 10 ] who proposed a method combining colour and depth information to summarise stereo video sequences. Papachristou et al. in [ 13 ] developed a video shot classification framework for stereoscopic video, in which the key-frame extraction method used is based on mutual information. Even though the framework was proposed for stereoscopic video, the key-frame extraction method only uses one view of the stereoscopic video. Until now, only some specific 3D video formats were considered by the existing key-frame extraction methods. Stereoscopic video was used in [ 10 , 42 ], V+D is used by Ferreira et al. in [ 42 ] and 3D computer graphics format in [ 17 – 19 , 22 , 25 , 51 ]. Thus, further room exists for research on efficient key-frame extraction methods that can be applied to other 3D video formats, such as MVV, MVD and holoscopic video.

Most 3D key-frame extraction methods cited in this paper were developed for specific content and 3D format and only four of them include comparisons with similar methods [ 18 , 22 , 41 , 51 ]. In [ 18 ], curve simplification, UnS and clustering methods were utilised as reference methods for performance evaluation and comparison of the proposed matrix factorisation methods. The authors showed that the method based on matrix factorisation extracts more representative key-frames in comparison with the other three competing methods [ 22 , 41 , 51 ]. However, the algorithm is very slow with quadratic running time complexity. In [ 22 ], the proposed method based on genetic algorithm is compared with Huang’s method [ 18 ] in terms of the PSNR and computational complexity. The former is very efficient in terms of computation time when compared to the latter but qualitywise (average PSNR) it is slightly worse. However, Huang’s method [ 18 ] is slightly better when comparing maximum and minimum PSNR.

Peng Huang et al. in [ 51 ] confront their key-frame extraction method with the method used in [ 19 ] and the results show improved performance for all 3D video sequences used. Jin et al. in [ 41 ] compare the proposed method with the UnS and Principal Component Analysis methods [ 56 ]. The results show that the proposed method achieves much better reconstruction of skeletal and mesh animation than the other methods under analysis.

As mentioned before, most of the key-frame extraction methods for 3D video, rely on a previous SBD step. However, the methods just described, from [ 18 , 22 , 41 , 51 ], do not perform any pre-analysis of the video signal to identify shots and their boundaries. The quality of key-frame summaries obtained using these approaches can be negatively affected when accurate shot segmentation is not available. Another important issue is the definition of the number of key-frames need to represent the original sequence. This number depends on user requirements and on the content of the video to be summarised and its choice frequently involves a trade-off between the quality and efficiency of the key-frame summary.

1.6 Key-frame presentation

Once the key-frames are extracted, they need to be presented in an organised manner to facilitate video browsing and navigation operations by the user. The video presentation methods aim to show the key-frames in some meaningful way allowing the user to grasp the content of a video without watching it from beginning to end [ 4 ]. The most common methods for key-frame presentation are the static storyboard, dynamic slideshow and single image, see Fig. 1 .

Static storyboard presents a set of miniaturised key-frames spatially tiled in chronological order, allowing a quick browsing and viewing of the original video sequence. This presentation method was used with 3D video in [ 10 , 18 , 19 , 22 , 41 , 51 ]. The second method is the dynamic slideshow that presents the key-frames one by one on the screen, which allows browsing over the whole video sequence. Other presentation method is the single image, which morphs parts of different key-frames in chronological to produce a single image. Normally, in this presentation type the background and foreground objects (time shifted) are aggregated in single image, as exemplified in Fig. 6 . In this figure, the foreground is the children who plays in bars of a playground. Here, we can see 3 positions of the children in the bars which corresponds to 3 key-frames of video sequence.

Video synopsis proposed [ 43 ]

Qing et al. [ 12 ] proposed a generic method for extracting key-frames in which the Jensen-Shannon divergence is used to measure the difference between video frames to segment the video into shots and to choose key-frames in each shot. The authors also proposed a 3D visualisation tool, used to display key-frames and the useful information related to the process of key-frame selection. More recently, Nguyen et al. [ 57 ] proposed the Video Summagator. This method provides a 3D visualisation of a video cube of static and dynamic video summaries. Assa et al. proposed a method to create an action synopsis image from a 3D animation sequence or 2D video [ 17 ]. Lee et al. also proposed a method to summarise a 3D animation into a single image based on depth information [ 25 ].

In [ 58 ] a 3D interface (3D-Ring and 3D-Globe) was proposed as an alternative to the 2D grid presentation for interactive item-search in visual content databases, see Fig. 7 . Even though this system was designed to be used with a large database it can also be applied to visualise key-frames summaries of 2D and 3D video.

a 3D-Ring interface, b 3D-Globe interface and c 2D grid presentation (figures based on [ 58 ])

1.6.1 Discussion

Most of the 3D key-frame extraction methods proposed in the literature until now are focused on the extraction rather than in the presentation of key-frame sets to the viewers. So far only Assa et al. and Lee et al. proposed in [ 17 ] and [ 25 ] two presentation solutions distinct from the static storyboard used in association with most of 3D key-frame extraction methods [ 10 , 18 , 19 , 22 , 41 , 51 ]. In this scenario, with only two presentation solutions, it is foreseeable that the development of new 3D video and image display devices will lead to the creation of new methods to display 3D video summaries or key-frame collages providing the user with more immersive and more meaningful ways to observe these types of time-condensed video representations.

1.7 Quality evaluation of 3D key-frame summaries

One of the most important topics in video summarisation algorithmic development is the evaluation of the key-frame extraction methods. In this section, we present current key-frame summary evaluation methods and some related issues. These methods are aggregated into three groups: result description, subjective and objective methods, as it was proposed in [ 4 ].

1.7.1 Result description

This is the most common and simple form of evaluation key-frame extraction methods since it does not require a reference, either for objective or subjective comparison with other methods. Usually, it is used to explain and describe the advantages of some method compared with others based on presentation or/and description of the key-frames extracted (visual comparison), as in [ 18 , 19 , 22 , 25 , 41 , 51 ]. This type of evaluation can also be used to discuss the influence of specific parameters or features of the method and also the influence of the content in the key-frame set, as in [ 10 , 19 ]. In some works, this type of evaluation method is complemented with objective and/or subjective methods as in [ 19 , 25 ]. However, the result description method has some limitations, such as the reduced number of methods which can be compared at same time, i.e. it is inadequate to compare key-frame summaries of a large number of video sequences or methods. Another drawback is the subjectivity inherent to this type of evaluation, since the underling comparisons results are usually user-dependent and so prone to inter and intra observer fluctuations.

1.7.2 Subjective methods

Subjective methods rely on the independent opinion of a panel of users judging the quality of the generated key-frame video summaries according to a known methodology. In this type evaluation, a panel of viewers are asked to observe both the summaries and the original sequence and then respond to questions related to some evaluation criteria, (e.g. ‘Was the key-frame summary useful?’, ‘Was the key-frame summary coherent?’) or if each key-frame is ‘good’, ‘fair’, or ‘poor’ according to the original video sequence.

The experiments can include a set of absolute evaluations and/or a set of relative evaluations, in which two key-frame summaries are presented and compared. Usually, the summary visualisation and rating steps are repeated for each video in the evaluation set by each viewer. During the evaluation of the key-frame summaries, it is also required taking into account the external factors which can influence the ratings of the summaries, such as the attention and fatigue specially when there are long evaluation sessions with many video summaries. In addition to these factors, the experiments must follow standard recommended protocols prepared specifically for subjective assessment of video quality [ 59 ].

Subjective evaluation methods were used in [ 16 , 60 – 63 ]. In [ 60 ], subjective assessment was used to grade the single key-frame representations as ‘good’, ‘bad’ or ‘neutral’ for each video shot and also give appreciations on the number of key-frames with possible grades being ‘good’, ‘too many’ and ‘too few’ in the case of multiple key-frames per shot. In [ 61 , 63 ], the quality of the key-frame summary is evaluated by asking users to give a mark between 0 to 100 for three criteria, ‘informativeness’, ‘enjoyability’ and ‘rank’ after watching the original sequences and the respective key-frames summaries. Ejaz et al. [ 62 ] used subjective evaluations to compare the proposed method with four prominent key-frame extraction methods: open video project (OV) [ 45 ], Delaunay triangulation (DT) [ 64 ], STIll and MOving Video Storyboard (STIMO) [ 65 ] and Video SUMMarisation (VSUMM) [ 24 ]. In this case, the evaluation is based on mean opinion scores (MOS) and viewers are asked to rate the quality of the key-frame summary on scale of 0 (minimum value) to 5 (maximum value) after watching the original sequences and the respective summaries generated by all the methods.

In [ 16 ] subjective assessments were also used to evaluate multiview video summaries. The aim is to grade the ‘enjoyability’, ‘informativeness’ and ‘usefulness’ of the video summary. Here, three questions were put to the viewer to evaluate the method: Q 1 : ‘How about the enjoyability of the video summary?’ Q 2 : ‘Do you think the information encoded in the summary is reliable compared to the original multiview videos’ and Q 3 : ‘Will you prefer the summary to original multiview videos if stored in your computer?’. In reply to the questions Q 1 and Q 2 , the viewers assigned a score between 0 (minimum value) to 5 (maximum value) and for Q 3 the viewers only need to respond with ‘yes’ or ‘no’. From all 3D key-frame extraction methods reviewed, only [ 16 , 17 , 25 ] used subjective evaluations.

1.7.3 Objective methods

Although subjective evaluation provides a better representation of the human perception than objective evaluation, it is not suitable for practical implementations due to the time required to conduct the opinion collection campaigns. Objective evaluation methods are reproducible and can be specified analytically, and since they are automatable can be used to rate the proposed method on large number of videos of variable genres and formats. These methods can be applied to all types of video formats without requiring the services of video experts and can be performed rapidly and automatically if suitable quality measures are available. Besides being faster, simpler and easily replicable, this type of method is more economical than the subjective evaluation.

The works reviewed in this article, which use objective quality evaluation, employ several quality measures originally developed for 2D video, but can be also applied to 3D video, after being modified to take into account the specific features of 3D visual information. The shot reconstruction degree (SRD) distortion measure [ 66 ] and the fidelity measure (Fm) defined in [ 67 ] follow two different approaches. Fidelity measure employs a global strategy, while SRD uses a local evaluation of the key-frames. To judge the conciseness of a key-frame summary a measure of the Compression Ratio (CR) is used [ 68 ]. If a ground-truth summary is available the Comparison of User Summaries (CUS) [ 24 ], Recall rate, Precision rate and accuracy measure (F1) measures can be used. These measures compare the computed summaries with those manually built by users. More details on these measures are presented in the next sub-sections.

1.7.3.1 Shot reconstruction degree:

SRD measures the capability of a set of key-frames to represent the original video sequence/shot. Assuming a video shot F ={ f 0 , f 1 ,…, f n −1 } of n frames and \(K=\{f_{l_{0}},f_{l_{1}}, \ldots,f_{l_{m-1}}\}\) a set of m key-frames selected from F , the reconstructed scene shot F ′ ={ f 0′, f 1′,…, f n −1′} is obtained from the K set by using some type of frame interpolation. The SRD measure is defined as

where n is the size of the original video sequence/shot F and \(\mathcal {S}im(.)\) is the similarity between two video frames. In Liu et al. [ 66 ], the similarity measure chosen was peak signal-to-noise ratio (PSNR), but other similarity metrics that include 3D features can also be used in the evaluation of 3D key-frame summaries. A K key-frame summary is a good representation of the original F when the magnitude of its SRD is high.

1.7.3.2 Fidelity

Fm is computed as the maximum of the minimal distances between the set of key-frames K and each frame of the original F , i.e. a Semi-Hausdorff distance d sh . Let F be a video sequence/shot containing n frames, and the set \(K=\{f_{l_{0}},f_{l_{1}}, \ldots,f_{l_{m-1}}\}\) of m frames, selected from F . The distance between the set K and a generic frame f k s.t. 0≤ k ≤ n −1 belonging to F can be calculated as follows.

Then the semi-Hausdorff distance d sh between K and F is defined as

The Fidelity measure is defined as

where MaxDiff is the largest possible value that the frame difference measure can assume. The function d ( f a , f b ) measures the difference between two video frames a and b . The majority of the existing dissimilarity measures can be used for d (,), such as the L 1 -norm (city block distance), L 2 -norm (Euclidean distance) and L n -norm [ 67 ]. As it was mentioned before, the Fm measure can be used for 3D video with the necessary changes in the d (,) distance. Whenever Fm is high, this means that the selected key-frames provide an accurate representation of the whole F .

1.7.3.3 Compression ratio:

A video summary should not contain too many key-frames since the aim of the summarisation process is to allow viewers to quickly grasp the content of a video sequence. For this reason it is important to quantify the conciseness of the key-frame summary. The conciseness is the length of the key-frame video summary in relation to the original video segment length and can be measured a compression ratio, defined as the relative amount of ‘savings’ provided by the summary representation:

where m and n are the number of frames in the key-frame set K and the original video sequence F respectively. Generally, high compression ratio is desirable for a compact video summary [ 68 ].

1.7.3.4 Comparison of user summaries (CUS):

CUS is a quantitative measure based on the comparison of summaries built manually by users and computed summaries. It was proposed by Avila et al. in [ 24 ]. The user summaries are taken as reference, i.e. the ground-truth, and the comparison between the summaries is based on specific metrics. The colour histogram is used for comparing key-frames from different video summaries, while the distance between them is measured using the Manhattan distance. Two key-frames are similar if the Manhattan distance of their colour histograms is below a predetermined threshold δ . In [ 24 ], this threshold value was set to 0.5. Two evaluation metrics, accuracy rate CUS A and error rate CUS E , are used to measure the quality of the computed summaries. They are defined as follows:

where n match and n no-match are, respectively, the number of matching and non-matching key-frames between the computed and the user generated summary and n US is the total number of key-frames in the summary. CUS A varies between 0 and 1, where CUS A =0 is the worst value indicating that none of the key-frames from the computed summary matches those of the user summary. A value of CUS A =1 is the best case and indicates that all key-frames from both summaries perfectly match each other. A null value for CUS E indicates a perfect match between both summaries.

1.7.3.5 Computational complexity:

Another relevant performance metric taken into account in the evaluation of key-frame extraction methods is the computational complexity, which is usually equated with the time spent to construct a key-frame summary. This metric was used in [ 24 , 62 , 63 , 68 , 69 ] for 2D video summaries. In 3D key-frame extraction methods, the computational complexity metric is only used by Lee et al. in [ 22 ], where the computational complexity of Lee’s and Huang’s et al. [ 18 ] methods are compared.

1.7.3.6 Other methods:

Other methods and measures were used for objective evaluation of 3D key-frames summaries. In [ 19 , 51 ] a rate-distortion curve is used, modelling a monotonic relationship between rate and distortion, with increases of the former leading to decreases of the latter. In the case of [ 18 ], the root mean square error (RMSE) distance between the original and reconstructed animation was used as the objective quality measure (with an inverse relationship in this case). This measurement is the same as in [ 70 ] and [ 71 ]. Lee et al. [ 22 ] used PSNR to measure the reconstruction distortion. Jin et al. in [ 41 ] measure reconstruction error of the animation from the extracted key-frames, using average of degrees of freedom (DOF) of reconstruction error magnitude.

1.7.4 Discussion

Conciseness, coverage, context and coherence are desirable attributes in any key-frame summary. Some of these attributes are mostly subjective as is the case of the context and coherence. Conciseness is related to the length of the key-summary, while the coverage evaluation is based on comparison between computed key-frames summary and ground-truth summary, expressed by the recall rate, precision rate, CUS A and CUS E .

Most evaluation metrics reviewed above were developed for 2D video. However, some of them, such as Fm and SRD, have also been extended to evaluate 3D video summaries after some adaptation. This is the casse of the 3D key-frame extraction method presented by Ferreira et al. in [ 42 ], where the Fm and SRD metrics were used. To measure the Recall rate, Precision rate, CUS A , CUS E , computational complexity and compression ratio in 3D video summarisation, no adaptation is needed.

The key-frame extraction methods are often application-dependent (e.g. summarisation of sports videos, news, home movies, entertainment videos and more recently for 3D animation) and the evaluation metrics must be adapted to the intended use. A good summary quality evaluation framework must be based on a hybrid evaluation scheme which includes the strengths of subjective and objective methods and also the advantages of result description evaluations.

1.8 Applications

In this section, some applications of 3D key-frame extraction methods and some aspects related to these applications are presented. These applications are grouped into five categories: video browsing, video retrieval, content description, animation synthesis and others.

1.8.1 Video browsing

The video browsing and associated problem has been investigated by the research community for decades, [ 72 ]. However, the growing use of 3D video and the specific characteristics of this type of visual information make 3D video browsing a more interesting and challenging problem. The access to databases or other collection of videos could be eased by the use the key-frame extraction methods to abstract/resume long video sequences in the repository of interest. With this kind of abridged video representation, a viewer can quickly find the desired video in a large database. For example, once an interesting topic has been identified through display of the key-frames, a simple operation as a click on the respective key-frame can initiate video playback of the original content at that particular instant. Many video browsing methods have been proposed for 2D video [ 72 ]. However, to the best of the authors’ knowledge, in the case of 3D video there are no works reported in the literature.

1.8.2 Video retrieval

In contrast to video browsing, where viewers often just browse interactively through video summaries in order to explore their content, in video retrieval the viewers search for certain visual objects (e.g. objects, people and scenes) in a video database. In this type of retrieval processes, viewers are typically expected to know exactly what they are looking for. Therefore, it is crucial to implement appropriate search mechanisms for different types of queries provided by distinct viewers and with particular interests. The matching between the viewers’ interests (queries) and the database content can be made with recourse to textual or image based descriptions or combinations of both. Some 2D video search and retrieval applications have combined video browsing and retrieval in the same platform [ 72 ]. In the case of 3D video this problem is still open for research. Finally, it is worth to point out that work done on 3D object recognition techniques which can also be used in retrieval, as published in [ 73 – 75 ].

1.8.3 Content description

Vretos et al. [ 76 ] presented a way of using the audio-visual description profile (AVDP) of the MPEG-7 standard for 3D video content description. The description of key-frames is contemplated in the AVDP profile through the MediaSourceDecompositionDS (i.e. MediaSourceDecompositionDS is used in the AVDP context to decompose an audiovisual segment into the constituent audio and video channels). Thus this content description scheme, allows that 3D key-frames can be used for fast browsing and condensed representation of query results of 3D video search tasks. Other application of key-frames to content description was proposed by Sano et al. [ 77 ]. Here, the authors proposed and discussed how the AVDP profile of the MPEG-7 can be applied to multiview 3D video content [ 56 ].

1.8.4 Animation synthesis

Blanz et al. [ 78 ] proposed a morphable 3D face model by transforming the shape and texture of example into a new 3D model representation. According to this modelling approach, new or similar faces and expressions can be created by forming linear combinations of the 3D face models. A similar concept to the proposed in [ 78 ] can be applied to generate 3D models [ 79 ] or to synthesise new motion from captured motion data [ 80 ]. Animation synthesis based on key-frames [ 81 ] using the same concept has been presented in [ 78 – 80 ], to interpolate frames between two key-frames. However, the quality of the interpolated frames is dependent on the inter-key-frame distance and on the interpolation method used.

1.8.5 Others

Assa et al. [ 17 ] proposed the use of action synopsis images as icons (personal computer desktop and folders) and thumbnails of the 3D animation. Assa et al. also proposed an automatic or semi-automatic generation method to create comic strips and storyboards for 3D animation. Lee et al. [ 25 ] presented a method to create a single image summary of a 2D or 3D animation, which can be used in the same application as Assa’s work. Halit et al. [ 56 ] proposed a tool for thumbnail generation from motion animation sequences. Several authors, as [ 82 – 85 ] have used key-frame extraction methods in the 2D-to-3D video conversion.

1.9 Prospects and challenges

Although some significant work has been done in the 3D video summarisation domain, many issues are still open and deserve further research, especially in the following areas.

1.9.1 SBD and key-frame extraction methods

The selection of the features used by shot boundary and key-frame extraction methods is still an open research problem, because these features depend on the application, video content and representation format. For instance, in fast-motion scenes edge information is not the best choice to detect shot boundaries due to motion-induced blur. Thus, it may be better to automatically find the useful features based on some assumptions about the video-content.

The majority of key-frame extraction methods published in the literature use low-level features and content sampling approaches to identify the relevant frames that should be included in the key-frame summary. Recently, the inclusion of perceptual metrics in the SBD and key-frame methods are gaining some space. Recently and in the context of 2D video, some key-frame extraction methods based on visual attention models have emerged as, [ 60 – 63 , 86 ]. However, for 3D video only two solutions are available [ 41 , 42 ]. Hence, key-frame extraction in 3D video still poses relevant research problems to be investigated and efficiently solved.

Another open challenge is the combination of the visual features with additional information, such as audio features, text captions and content description, for use in shot boundary detection and selection of the optimal frames in 3D video. In the current literature, there is also a lack of summarisation methods based on key-frames or video skims, for the most recent 3D video formats such as MVD and plenoptic video. Another topic open to further research is the application of scalable summarisation to 3D formats [ 87 ]. Despite the fact that several previous works addressed scalable summarisation for 2D video, e.g. [ 88 , 89 ], such methods were not extended to 3D and multiview, which leads to open research questions.

1.9.2 Evaluation

In the past evaluation frameworks for 2D key-frame summarisation methods were proposed in [ 90 , 91 ]. More recently, Avila et al. [ 24 ] also proposed another evaluation setup, wherein the original video and the key-frame summaries of several methods are available for downloading, together with the results of several key-frame extraction methods for 2D video. Unfortunately for the case of 3D video, there is not as yet any similar framework, where key-frame summaries and the respective original sequences are available for research use.

The number and diversity of evaluation metrics (objective, result description and subjective) used to compare state-of-the-art key-frame extraction methods make their comparative assessment a difficult task. Therefore, the development of metrics which can be used in the evaluation of key-frames summaries in different domains and 3D video formats is a very important area of video-summarisation related research. Furthermore the focus of the evaluation process must be application-dependent. For instance, in browsing applications, the time spent by the user to search or browse for a particular video is the most important factor, but on the other hand, in detection events, the evaluation metric must focus on the successful detection of these events.

One other problem that arises in the evaluation process is the replication of results of previous works, as some works are not described with enough details to allow independent implementation or the input data is unavailable or else it is not easy to use due to data format incompatibilities or lack of information about their representation format. Thus, the best way to test and compare key-frame extraction methods for 2D and 3D is to build publicly accessible repositories containing test kits, made up of executable or web-executable versions of the methods and the test sequences.

1.9.3 Presentation

Another challenging topic in the research of 3D key-frame summarisation is the design of an efficient and intuitive visualisation interface that allows easy navigation and visualisation of the key-frame summaries. These applications should be independent of the terminal capabilities (display dimension, processing and battery power), i.e. should be usable on small screen devices such as smartphones as well as on ultra-high-definition displays. In addition, the visualisation interface should be independent from the key-frame summarisation method, to allow the visualisation of different formats of 3D key-frames video summaries, such as stereoscopic video or video-plus-depth and also 2D video in the same framework. The interface should be capable of dealing with the most common key-frame visualisation methods such as, static storyboard, dynamic slideshow and hierarchically arranged viewing. In particular, the most recent 3D interface for searching and viewing images or video in large databases, 3D-Ring and 3D-Globe, are interesting solutions which must be taken into account in the definition of new key-frame visualisation methods [ 58 ].

1.9.4 Video summary coding

In the past, the problem of scalable coding of video summaries was addressed in [ 88 , 92 – 94 ]. In [ 92 ] the authors propose a hierarchical frame selection scheme which considers semantic relevance in video sequences at different levels computed from compressed wavelet-based scalable video. In [ 93 ], a method to generate video summaries from scalable video streams based on motion information is presented; while in [ 94 ], the authors propose to partition a video summary into summarisation units related by the prediction structure and independently decodable. Ferreira et al. in [ 88 ] proposed a method to encode an arbitrary video summary using dynamic GOP structures in scalable streams. The scalable stream obtained was fully compatible with the scalable extension of the H.264/AVC standard. However, all approaches were proposed for 2D video and used older generation video coding methods. The application of video summary coding to the 3D video format and the use of the most recent video coding, such as HEVC, should also be explored to find efficient coding tools for such purpose.

2 Conclusions

In this paper, we have presented a review of 3D key-frame extraction methods covering the major results published in recent journal issues and conference proceedings. Different state-of-the-art methods for key-frame extraction and evaluation metrics were presented and examined. The most important presentation methods for key-frame summaries were also discussed.

Various suggestions for the development of future 3D video summarisation methods are made, particularly oriented for future research on 3D key-frame extraction methods and potential benefits that may be attained from further research based on visual attention models. So far, 3D video key-frame extraction methods based on visual attention have not been deeply researched, so this is an interesting point to be explored. More research effort should also be put on methods for performance evaluation of key-frame extraction algorithms. The current plethora of different objective and subjective evaluation methods, most of them not easily comparable between each other, motivates a research goal towards unified and comparable methods for performance evaluation and benchmarking of 3D video summaries.

One other important and interesting research topic is the design and implementation of methods and tools to present 3D key-frame summaries. It is clear that the way a key-frame set is presented to viewers influence the time and effort they have to devote to interpret the summarised visual data. Finally, efficient coding of video summaries also leads to research problems which are still open for further research, since no specific solutions for 3D video are currently available.

A Smolic, K Mueller, N Stefanoski, J Ostermann, A Gotchev, GB Akar, G Triantafyllidis, A Koz, Coding algorithms for 3DTV—a survey. Circ Syst Video Technol IEEE Trans. 17 (11), 1606–1621 (2007).

Article   Google Scholar  

P Merkle, K Müller, T Wiegand, 3D video: acquisition, coding, and display. Consum Electron Transac. 56 (2), 946–950 (2010).

S Chikkerur, V Sundaram, M Reisslein, LJ Karam, Objective video quality assessment methods: A classification, review, and performance comparison. Broadcasting IEEE Transac. 57 (2), 165–182 (2011).

BT Truong, S Venkatesh, Video abstraction: a systematic review and classification. ACM Trans Multimedia Comput. Commun. Appl. (TOMCCAP). 3 (1), 3 (2007).

W Hu, N Xie, L Li, X Zeng, S Maybank, A survey on visual content-based video indexing and retrieval. Syst Man Cybern. Part C: Appl Reviews, IEEE Trans. 41 (6), 797–819 (2011).

AG Money, H Agius, Video summarisation: a conceptual framework and survey of the state of the art. J Vis Commun Image Represent. 19 (2), 121–143 (2008).

Y Li, S-H Lee, C-H Yeh, C-CJ Kuo, Techniques for movie content analysis and skimming: tutorial and overview on video abstraction techniques. IEEE Signal Process Mag. 23 (2), 79–89 (2006).

K McHenry, P Bajcsy, An Overview of 3D Data Content, File Formats and Viewers. Technical Report NCSA-ISDA08-002 (2008). http://207.245.165.87/applied-research/papers/overview-3d-data.pdf .

EH Adelson, JR Bergen, in The plenoptic function and the elements of early vision , ed. by M Landy, JA Movshon (MIT PressCambridge, 1991), pp. 3–20.

ND Doulamis, AD Doulamis, YS Avrithis, KS Ntalianis, SD Kollias, Efficient summarization of stereoscopic video sequences. Circ Syst Video Technol IEEE Transac. 10 (4), 501–517 (2000).

L Ferreira, P Assuncao, LA da Silva Cruz, in Content-Based Multimedia Indexing (CBMI), 2013 11th International Workshop On . 3D video shot boundary detection based on clustering of depth-temporal features (IEEEVeszprem, 2013), pp. 1–6.

Chapter   Google Scholar  

Q Xu, P Wang, B Long, M Sbert, M Feixas, R Scopigno, in Systems Man and Cybernetics (SMC), 2010 IEEE International Conference On . Selection and 3D visualization of video key frames (Istanbul, 2010), pp. 52–59.

K Papachristou, A Tefas, N Nikolaidis, I Pitas, in Machine Learning for Signal Processing (MLSP), 2014 IEEE International Workshop On . Stereoscopic video shot classification based on weighted linear discriminant analysis (Reims, 2014), pp. 1–6.

J Lin, X Ruan, N Yu, R Wei, in The 27th Chinese Control and Decision Conference (2015 CCDC) . One-shot learning gesture recognition based on improved 3D SMoSIFT feature descriptor from RGB-D videos (Qingdao, 2015), pp. 4911–4916.

I Mademlis, N Nikolaidis, I Pitas, in Signal Processing Conference (EUSIPCO), 2015 23rd European . Stereoscopic video description for key-frame extraction in movie summarization (Nice, 2015), pp. 819–823.

Y Fu, Y Guo, Y Zhu, F Liu, C Song, Z-H Zhou, Multi-view video summarization. Multimedia IEEE Transac. 12 (7), 717–729 (2010).

J Assa, Y Caspi, D Cohen-Or, Action synopsis: pose selection and illustration. ACM Trans Graphics (TOG). 24 (3), 667–676 (2005).

K-S Huang, C-F Chang, Y-Y Hsu, S-N Yang, Key probe: a technique for animation keyframe extraction. The Visual Computer. 21 (8–10), 532–541 (2005).

J Xu, T Yamasaki, K Aizawa, Summarization of 3D video by rate-distortion trade-off. IEICE Transactions on Information and Systems. 90 (9), 1430–1438 (2007).

T Yamasaki, K Aizawa, Motion segmentation and retrieval for 3D video based on modified shape distribution. EURASIP J. Appl. Signal Process. 2007 (1), 211–211 (2007).

Article   MATH   Google Scholar  

P Huang, A Hilton, J Starck, Shape similarity for 3D video sequences of people. Int J Comput Vision. 89 (2), 362–381 (2010).

T-Y Lee, C-H Lin, Y-S Wang, T-G Chen, Animation key-frame extraction and simplification using deformation analysis. Circ Syst Video Technol IEEE Trans. 18 (4), 478–486 (2008).

J Xu, T Yamasaki, K Aizawa, Temporal segmentation of 3-D video by histogram-based feature vectors. Circ Syst Video Technol IEEE Trans. 19 (6), 870–881 (2009).

SEF de Avila, APB ao Lopes, A da Luz Jr., A de Albuquerque Araúo, VSUMM: a mechanism designed to produce static video summaries and a novel evaluation method. Pattern Recogn Lett. 32 (1), 56–68 (2011). Image Processing, Computer Vision and Pattern Recognition in Latin America.

H-J Lee, HJ Shin, J-J Choi, Single image summarization of 3D animation using depth images. Comput Animat Virtual Worlds. 23 (3–4), 417–424 (2012).

Y-J Zhang, Advances in image and video segmentation (IRM Press, USA, 2006).

Book   Google Scholar  

J Yuan, H Wang, L Xiao, W Zheng, J Li, F Lin, B Zhang, A formal study of shot boundary detection. Circ Syst Video Technol IEEE Trans. 17 (2), 168–186 (2007).

AF Smeaton, P Over, AR Doherty, Video shot boundary detection: seven years of TRECVid activity. Comput Vision Image Underst. 114 (4), 411–418 (2010). Special issue on Image and Video Retrieval Evaluation.

C Cotsaces, N Nikolaidis, I Pitas, Video shot detection and condensed representation. A review. Signal Process Mag IEEE. 23 (2), 28–37 (2006).

J Nam, AH Tewfik, Detection of gradual transitions in video sequences using B-spline interpolation. Multimed IEEE Trans. 7 (4), 667–679 (2005).

S Lian, Automatic video temporal segmentation based on multiple features. Soft Computing. 15: , 469–482 (2011).

P Sidiropoulos, V Mezaris, I Kompatsiaris, H Meinedo, M Bugalho, I Trancoso, Temporal video segmentation to scenes using high-level audiovisual features. Circ Syst Video Technol IEEE Trans. 21 (8), 1163–1177 (2011).

B-L Yeo, B Liu, Rapid scene analysis on compressed video. Circ Syst Video Technol IEEE Transactions on. 5 (6), 533–544 (1995).

Z Cernekova, I Pitas, C Nikou, Information theory-based shot cut/fade detection and video summarization. Circ Syst Video Technol IEEE Transactions on. 16 (1), 82–91 (2006).

PJ Besl, ND McKay, A method for registration of 3-D shapes. Pattern Anal Mach Intel IEEE Trans. 14 (2), 239–256 (1992).

B Ionescu, D Coquin, P Lambert, V Buzuloiu, A fuzzy color-based approach for understanding animated movies content in the indexing task. Eurasip J Image Video Process. 10 (2008), 1–17 (2008).

WAC Fernando, CN Canagarajah, DR Bull, in Image Processing, 1999. ICIP 99. Proceedings. 1999 International Conference On , 3. Fade and dissolve detection in uncompressed and compressed video sequences (Kobe, 1999), pp. 299–303.

R Slama, H Wannous, M Daoudi, 3D human motion analysis framework for shape similarity and retrieval. Image Vision Comput. 32 (2), 131–154 (2014).

U Gargi, R Kasturi, SH Strayer, Performance characterization of video-shot-change detection methods. Circ Syst Video Technol IEEE Transactions on. 10 (1), 1–13 (2000).

Y Yang, X Liu, A re-examination of text categorization methods, 42–49 (1999).

C Jin, T Fevens, S Mudur, Optimized keyframe extraction for 3D character animations. Comput Animat Virtual Worlds. 23 (6), 559–568 (2012).

L Ferreira, LA da Silva Cruz, P Assuncao, A generic framework for optimal 2D/3D key-frame extraction driven by aggregated saliency maps. Signal Processing: Image Commun. 39 Part A: , 98–110 (2015).

Google Scholar  

A Rav-Acha, Y Pritch, S Peleg, in Computer Vision and Pattern Recognition, 2006 IEEE Computer Society Conference On , 1. Making a long video short: dynamic video synopsis, (2006), pp. 435–441.

R Xu, ID Wunsch, Survey of clustering algorithms. Neural Netw IEEE Trans. 16 (3), 645–678 (2005).

D DeMenthon, V Kobla, D Doermann, in Proceedings of the Sixth ACM International Conference on Multimedia. MULTIMEDIA ’98 . Video summarization by curve simplification (ACMNew York, 1998), pp. 211–218.

LJ Latecki, D de Wildt, J Hu, in Multimedia Signal Processing, 2001 IEEE Fourth Workshop On . Extraction of key frames from videos by optimal color composition matching and polygon simplification (Cannes, 2001), pp. 245–250.

J Calic, E Izuierdo, in Information Technology: Coding and Computing, 2002. Proceedings. International Conference On . Efficient key-frame extraction and video analysis (Las Vegas, 2002), pp. 28–33.

IS Lim, D Thalmann, in Engineering in Medicine and Biology Society, 2001. Proceedings of the 23rd Annual International Conference of the IEEE , 2. Key-posture extraction out of human motion data, (2001), pp. 1167–1169.

ND Doulamis, AD Doulamis, Y Avrithis, SD Kollias, in Multimedia Signal Processing, 1999 IEEE 3rd Workshop On . A stochastic framework for optimal key frame extraction from MPEG video databases (Copenhagen, 1999), pp. 141–146.

B-D Choi, J-W Han, C-S Kim, S-J Ko, Motion-compensated frame interpolation using bilateral motion estimation and adaptive overlapped block motion compensation. Circ Syst Video Technol IEEE Trans. 17 (4), 407–416 (2007).

P Huang, A Hilton, J Starck, in 3DPVT ’08: Proceedings of the Fourth International Symposium on 3D Data Processing, Visualization and Transmission . Automatic 3D video summarization: key frame extraction from self-similarity (IEEE Computer SocietyWashington, DC, 2008).

Y Gong, X Liu, in Computer Vision and Pattern Recognition, 2000. Proceedings. IEEE Conference On , 2. Video summarization using singular value decomposition (Hilton Head Island, 2000), pp. 174–180.

M Cooper, J Foote, in Multimedia Signal Processing, 2002 IEEE Workshop On . Summarizing video using non-negative similarity matrix factorization (St. Thomas, 2002), pp. 25–28.

W Abd-Almageed, in Image Processing, 2008. ICIP 2008. 15th IEEE International Conference On . Online, simultaneous shot boundary detection and key frame extraction for sports videos using rank tracing (San Diego, 2008), pp. 3200–3203.

CH Lee, A Varshney, DW Jacobs, Mesh saliency. ACM Trans. Graph. 24 (3), 659–666 (2005).

C Halit, T Capin, Multiscale motion saliency for keyframe extraction from motion capture sequences. Comput Animat Virtual Worlds. 22 (1), 3–14 (2011).

C Nguyen, Y Niu, F Liu, in Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. CHI ’12 . Video summagator: an interface for video summarization and navigation (ACMNew York, 2012), pp. 647–650.

K Schoeffmann, D Ahlstrom, MA Hudelist, 3D interfaces to improve the performance of visual known-item search. Multimed IEEE Trans. 16 (7), 1942–1951 (2014).

ITU-R Recommendations 500-13. Methodology for the Subjective Assessment of the Quality of Television Pictures (International Telecommunication UnionGeneva, 2012).

Y-F Ma, X-S Hua, L Lu, H-J Zhang, A generic framework of user attention model and its application in video summarization. Multimedia IEEE Trans. 7 (5), 907–919 (2005).

J Peng, Q Xiao-Lin, Keyframe-based video summary using visual attention clues. IEEE Multimedia. 17 (2), 64–73 (2010).

N Ejaz, I Mehmood, S Wook Baik, Efficient visual attention based framework for extracting key frames from videos. Signal Process Image Commun. 28 (1), 34–44 (2013).

N Ejaz, I Mehmood, SW Baik, Feature aggregation based visual attention model for video summarization. Comput Electr Eng. 40 (3), 993–1005 (2014).

P Mundur, Y Rao, Y Yesha, Keyframe-based video summarization using Delaunay clustering. Int. J. Digit. Libr. 6 (2), 219–232 (2006).

M Furini, F Geraci, M Montangero, M Pellegrini, Stimo: still and moving video storyboard for the web scenario. Multimed Tools Appl. 46 (1), 47–69 (2010).

T-Y Liu, X-D Zhang, J Feng, K-T Lo, Shot reconstruction degree: a novel criterion for key frame selection. Pattern Recogn Lett. 25: , 1451–1457 (2004).

HS Chang, S Sull, SU Lee, Efficient video indexing scheme for content-based retrieval. Circ Syst Video Technol IEEE Trans. 9 (8), 1269–1279 (1999).

G Ciocca, R Schettini, An innovative algorithm for key frame extraction in video summarization. J Real-Time Image Process. 1 (1), 69–88 (2006).

Q-G Ji, Z-D Fang, Z-H Xie, Z-M Lu, Video abstraction based on the visual attention model and online clustering. Signal Process Image Commun. 28 (3), 241–253 (2013).

A Khodakovsky, P Schröder, W Sweldens, in Proceedings of the 27th Annual Conference on Computer Graphics and Interactive Techniques. SIGGRAPH ’00 . Progressive geometry compression (ACM Press/Addison-Wesley Publishing Co.New York, 2000), pp. 271–278.

HM Briceño, PV Sander, L McMillan, S Gortler, H Hoppe, in Proceedings of the 2003 ACM SIGGRAPH/Eurographics Symposium on Computer Animation. SCA ’03 . Geometry videos: a new representation for 3D animations (Eurographics AssociationAire-la-Ville, Switzerland, Switzerland, 2003), pp. 136–146.

K Schoeffmann, F Hopfgartner, O Marques, L Boeszoermenyi, JM Jose, Video browsing interfaces and applications: a review. J Photonics Energy. 1: , 018004–01800435 (2010).

B Bustos, DA Keim, D Saupe, T Schreck, Vranic, DV́, Feature-based similarity search in 3D object databases. ACM Comput Surv. 37 (4), 345–387 (2005).

T Napoléon, H Sahbi, From 2D silhouettes to 3D object retrieval: contributions and benchmarking. J Image Video Process, 1 (2010).

M Savelonas, I Pratikakis, K Sfikas, An overview of partial 3D object retrieval methodologies. Multimed Tools Appl,. 74: , 1–26 (2014).

N Vretos, N Nikolaidis, I Pitas, in 3DTV-Conference: The True Vision—Capture, Transmission and Display of 3D Video (3DTV-CON), 2012 . The use of audio-visual description profile in 3D video content description (Zurich, 2012), pp. 1–4.

M Sano, W Bailer, A Messina, J-P Evain, M Matton, in IVMSP Workshop, 2013 IEEE 11th . The MPEG-7 audiovisual description profile (AVDP) and its application to multi-view video (Seoul, 2013), pp. 1–4.

V Blanz, T Vetter, in Proceedings of the 26th Annual Conference on Computer Graphics and Interactive Techniques. SIGGRAPH ’99 . A morphable model for the synthesis of 3D faces (ACM Press/Addison-Wesley Publishing Co.New York, 1999), pp. 187–194.

C Shelton, Morphable surface models. Int J Comput Vis. 38 (1), 75–91 (2000).

Article   MathSciNet   MATH   Google Scholar  

J Xu, T Yamasaki, K Aizawa, in 3D Data Processing, Visualization, and Transmission, Third International Symposium on . Motion editing in 3D video database, (2006), pp. 472–479.

R Parent, Computer animation: algorithms and techniques (Morgan-Kaufmann, USA, 2012).

X Cao, Z Li, Q Dai, Semi-automatic 2D-to-3D conversion using disparity propagation. Broadcasting, IEEE Transactions on. 57 (2), 491–499 (2011).

W-N Lie, C-Y Chen, W-C Chen, 2D to 3D video conversion with key-frame depth propagation and trilateral filtering. Electronics Letters. 47 (5), 319–321 (2011).

Article   MathSciNet   Google Scholar  

D Wang, J Liu, J Sun, W Liu, Y Li, in Broadband Multimedia Systems and Broadcasting (BMSB), 2012 IEEE International Symposium On . A novel key-frame extraction method for semi-automatic 2D-to-3D video conversion (Seoul, 2012), pp. 1–5.

K Ju, H Xiong, in Proc. SPIE , 9273. A semi-automatic 2D-to-3D video conversion with adaptive key-frame selection (SPIEBeijing, 2014), pp. 92730M1–92730M8.

J-L Lai, Y Yi, Key frame extraction based on visual attention model. J Vis Commun Image Represent. 23 (1), 114–125 (2012).

H Schwarz, D Marpe, T Wiegand, Overview of the scalable video coding extension of the H.264/AVC standard. Circ Syst Video Technol IEEE Trans. 17 (9), 1103–1120 (2007).

L Ferreira, L Cruz, P Assuncao, in EUROCON 2011 IEEE . Efficient scalable coding of video summaries using dynamic GOP structures (Lisbon, 2011), pp. 1–4.

L Herranz, S Jiang, Scalable storyboards in handheld devices: applications and evaluation metrics. Multimed Tools Appl. 75: , 1–29 (2015).

P Over, AF Smeaton, G Awad, in Proceedings of the 2Nd ACM TRECVid Video Summarization Workshop. TVS ’08 . The trecvid 2008 BBC rushes summarization evaluation (ACMNew York, 2008), pp. 1–20.

P Over, AF Smeaton, P Kelly, in Proceedings of the International Workshop on TRECVID Video Summarization. TVS ’07 . The TRECVID 2007 BBC rushes summarization evaluation pilot (ACMNew York, 2007), pp. 1–15.

J Bescos, JM Martinez, L Herranz, F Tiburzi, Content-driven adaptation of on-line video. Signal Process Image Commun. 22 (7–8), 651–668 (2007).

M Mrak, J Calic, A Kondoz, Fast analysis of scalable video for adaptive browsing interfaces. Comp Vision Image Underst. 113 (3), 425–434 (2009).

L Herranz, J Martínez, An integrated approach to summarization and adaptation using H.264/MPEG-4 SVC. Signal Processing: Image Comm. 24 (6), 499–509 (2009).

Download references

Acknowledgements

This work was supported by the R&D Unit UID/EEA/ 50008/2013, Project 3DVQM and PhD Grant SFRH/ BD/37510/2007, co-funded by FEDER-PT2020, FCT/ MEC, Portugal.

Authors’ contributions

LF read and summarised some of the scientific articles reviewed and drafted the manuscript. LC and PA read and summarised some of the scientific articles reviewed, wrote some sections and revised the manuscript. All authors read and approved the final manuscript.

Competing interests

The authors declare that they have no competing interests.

Author information

Authors and affiliations.

Instituto de Telecomunicações (Leiria), Campus 2 Morro do Lena—Alto do Vieiro, Leiria, 2411-901, Portugal

Lino Ferreira & Pedro Assuncao

Dep. de Engenharia Electrotécnica e de Computadores, Universidade de Coimbra, Pólo II da UC, Coimbra, 3030-290, Portugal

Lino Ferreira & Luis A. da Silva Cruz

Instituto de Telecomunicações (Coimbra), Dep. de Engenharia Electrotécnica e de Computadores, Pólo II da UC, Coimbra, 3030-290, Portugal

Luis A. da Silva Cruz

Instituto Politécnico de Leiria/ESTG, Campus 2 Morro do Lena—Alto do Vieiro, Leiria, 2411-901, Portugal

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Lino Ferreira .

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License ( http://creativecommons.org/licenses/by/4.0/ ), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Reprints and permissions

About this article

Cite this article.

Ferreira, L., da Silva Cruz, L.A. & Assuncao, P. Towards key-frame extraction methods for 3D video: a review. J Image Video Proc. 2016 , 28 (2016). https://doi.org/10.1186/s13640-016-0131-8

Download citation

Received : 25 January 2016

Accepted : 12 September 2016

Published : 29 September 2016

DOI : https://doi.org/10.1186/s13640-016-0131-8

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• 3D key-frames extraction
• 3D video summarisation
• Shot boundary detection