Natural Keyframe Animation Using an Artist Doll

Prabath Gunawardane, Eddy Chandra, and Tien-Chieng Jack Feng

University of California, Santa Cruz

{prabath, echandra, txfeng}

A sequence of animated frames


Our project attempts to create high quality animation that can be easily used by anyone. We use an artists doll to generate the keyframe poses combined with a motion capture database to generate the in-between frames for smoother, life-like motion. The doll has joints painted in different colors and a stereo camera rig is used to determine its joint positions, which in turn is used to calculate the joint angles for that keyframe. Using a motion capture database, we match the keyframes to the existing data and synthesize the in-between frames. Our contribution is to bring these two techniques together to create a complete, simple-to-use animation tool which generates natural looking results.

Keywords: keyframe animation, motion capture, motion synthesis

Documentations and Data

Siggraph Sketch (1 page)  and  Video

Class Paper  and  Presentation  


Submitted by Prabath Gunawardane:

Arikan O., Forsyth D.A., 2002. Interactive motion generation from examples. SIGGRAPH 2002

Presents a technique to generate a complete motion sequence by combining cut portions from previously recorded motion data, subjective to a number of constraints. Relates to our project since we are interested in ways of combining motion too.

Gleicher M., 2000. Animation from observation: Motion capture and motion editing. SIGGRAPH-Graphics 2000

This is a general overview of the issues faced and the techniques used in creating animations by data obtained through motion capturing. Presents a good high-level overview of the whole process.

Kovar L., Gleicher M., Pighin F. , 2002. Motion Graphs. SIGGRAPH 2002

Given a motion database, this technique creates a motion graph, i.e. a directed graph containing actual motions and also synthesized transitions between them. In our project, transition from one motion sequence to another may also be implemented in a similar manner.

Li, Y., Wang, T., Shum, H.Y., 2002. Motion texture: a two-level statistical model for character motion synthesis. SIGGRAPH 2002

This paper presents a technique for statistical matching of human motion segments and motion texturing based on that. Each particular motion segment is modelled as a 'texton' (which is a linear dynamical system) and the transition between textons is defined via a matrix. Although we won't be using any of the actual systems in our implementation, the general statistical matching idea is something we might try.

Mukai, T., Kuriyama, S., and Kaneko, T. 2006. Keyframe animation of virtual humans via motion data learning. Syst. Comput. Japan 37, 5 (May. 2006), 25-33.

The technique presented in this paper tries to optimize the interpolated motion between the keyframes by using motion capture data. Imposes a hierachical structure on the posture for matching and interpolation. Related to our project since it presents another method for synthesizing inbetween frames using mocap data.

Pullen, K., Bregler, C. 2000. Motion capture assisted animation: Texturing and synthesis. In Proceedings of SIGGRAPH 2002.

Presents a method which allows the animator to specify the motion using a number of keyframes (which is similar to our project) or specifying several degrees of freedom and uses stored motion capture data to synthesize the other frames.

Submitted by Eddy Chandra:

Arikan, O., and Forsyth, D. A. 2002. Interactive motion generation from examples. Proc. SIGGRAPH 2002.

The paper is about automatic generation of new human motion from motion capture data at interactive rates. It relates to our project as we will be using a motion capture database to create the in-between frames.

Esposito, C., Paley, W. B., and Ong, J. 1995. Of mice and monkeys: a specialized input device for virtual body animation. In Proceedings of the 1995 Symposium on interactive 3D Graphics (Monterey, California, United States, April 09 - 12, 1995). SI3D '95. ACM Press, New York, NY, 109-ff. DOI=

The paper is about a user interface called the Monkey which is used to specify keyframes in an animation. It relates to our project as we have an artist doll as our user interface for specifying the keyframes.

Gleicher, M. 1999. Animation from observation: Motion capture and motion editing. SIGGRAPH Comput. Graph. 33, 4 (Nov. 1999), 51-54. DOI=

The paper is about an overview of computer animation, computer vision, and motion capture. It relates to our project as it gives the background of creating an animation from real motions data.

Lamouret, A. and van de Panne, M. 1996. Motion synthesis by example. In Proceedings of the Eurographics Workshop on Computer Animation and Simulation '96 (Poitiers, France). R. Boulic and G. H�gron, Eds. Springer-Verlag New York, New York, NY, 199-212.

The paper is about synthesizing new motions from motions in a motion capture database given the state of the character, the state of the environment, and the style of motion. It relates to our project as we need to synthesis motions between the starting (or ending) keyframe and the starting (or ending) motion from our motion capture database. This is one of the ways we can use to smooth a resulting animation.

Lasseter, J. 1987. Principles of traditional animation applied to 3D computer animation. In Proceedings of the 14th Annual Conference on Computer Graphics and interactive Techniques M. C. Stone, Ed. SIGGRAPH '87. ACM Press, New York, NY, 35-44. DOI=

The paper is about eleven principles of traditional 2D animation that can be applied to 3D animation. Our system reduces the amount of work an animator does to do corretly some of the principles such as timing, keyframing, and inbetweening through the use of an artist doll and a motion capture database.

Pullen, K., and Bregler, C. 2000. Animating by multi-level sampling. IEEE Computer Animation Conference, 36-42.

The paper is about creating a variety of motions from a motion in a motion capture database using frequency bands. This randomness in life-like motions brings life-like animations. So, for instance, when a walking motion is used multiple times, the motion is different each time. Furthermore, it maintains the correlations between the joints, and it is close to real different walking motions as if they are in the database. It relates to our project as we may introduce randomness in our motions too. Thus, it enriches the set of motions we have in our database.

Submitted by Tien-Chieng Jack Feng:

1. Brand, M., and Hertzmann, A. 2000. Style machines. proc. SIGGRAPH 2000, 183-192

The paper presents a method for generating new animated motion style using motion capture sequences. It is related to our project since we are also synthesizing new motion from existing mocap data.

2. Bruderlin, A., and Williams, L. 1995. Motion signal processing. In Proceedings of ACM SIGGRAPH 95, Annual Conference Series, ACM SIGGRAPH, 97-104.

The paper presents methods to apply signal processing techniques on manipulating animated motion. It is related to our project since it describes how to modify and combine interactively the existing motion data.

3. Davis, J., Agrawala, M., Chuang, E., Popovic', Z., and Salesin, D. 2003. A sketching interface for articulated figure animation. In Proceedings of the 2003 ACM Siggraph/Eurographics Symposium on Computer Animation (San Diego, California, July 26 - 27, 2003). Symposium on Computer Animation. Eurographics Association, Aire-la-Ville, Switzerland, 320-328.

This paper presents an intuitive interface for generating 3D animation from 2D keyframe sketches drawn by the animator. This paper relates to our project in the way that we are also creating a new interface which gives animators the ability to easily specify keyframes for 3D animation.

4. Lee, J., and Shin, S. Y. 2001. A coordinate-invariant approach to multiresolution motion analysis Graphical Models 63,2,87-105

The paper presented a new multiresolution method for analyzing and synthesizing motion. This paper is related to our project as we are also using frequency bands and applying them to motion data analysis.

5. Molina-Tanco, L., and Hilton, A. 2000. Realistic synthesis of novel human movements from a database of motion capture examples. In Proceedings of the Workshop on Human Motion, IEEE Computer Society, 137-142.

This paper presents a system that synthesizes human motion through the use of a statistical model on motion captured data. This paper relates to our project as it describes a method for synthesizing new motions from original captured data by specifying start and end keyframes.

6. Rose, C., Cohen, M., and Bodenheimer, B. 1998. Verbs and adverbs: Multidimensional motion interpolation. IEEE Computer Graphics and Application 18, 5, 32-40.

The paper describes a system that can be used to interpolate between motion data. For our project we will need to interpolate between motion data and the user specified keyframes.

7. Q. L. Li and W. D. Geng and T. Yu and X. J. Shen and N. Lau and G. Yu 2006. MotionMaster: authoring and choreographing Kung-fu motions by sketch drawings SCA '06: Proceedings of the 2006 ACM SIGGRAPH/Eurographics symposium on Computer animation

The paper describes a method using artist's sketches to generate 3D animation, specifically Kungfu motions. For our project we try to solve a similar problem, providing a simple to use animation interface

Return to the top

Updated on 03/14/07