SIGGRAPH 2002 Session Review
Animation from Motion Capture
Reviewer: Christian Graf
Ratings: Difficulty: 5/10
(Moderate)
Interest: 7/10 (High)
This session would give a good overview of approaches to automatically create and control motion of artificial characters. The results are excellent and auspicious. The algorithms are very descriptive and do not require heavy math background.
Best Image of Session
Animation from Motion Capture
From the paper “Motion Texture: A Two-Level Statistical Model for Character Motion Synthesis”
Paper Evaluation
Motion Graphs
Kovar, Gleicher, Pighin
Ratings: Difficulty: 3/10
(Easy)
Interest: 5/10 (Moderate)
The paper explains how to create realistic, controllable motion of synthetic characters automatically. This is done by the combination of given motion and automatically generated transitions along a path under predefined user constraints. It mimics the realism of motion capture while maintaining the control over the character in the user’s hands.
Paper Evaluation
Motion Texture: A
Two-Level Statistical Model
for Character
Motion Synthesis
Li, Wang, Shum
Ratings: Difficulty: 7/10
(Moderate)
Interest: 8/10 (High)
The paper aims on synthesizing complex human-figure motion. From original motion capture data the algorithm learns the maximum likelihood for motion to animate a character. New animations can be generated from the learnt motion and the user may direct the character interactively on different level.
Paper Evaluation
Interactive Motion
Generation from Examples
Arikan, Forsyth
Ratings: Difficulty: 4/10
(Moderate)
Interest: 8/10 (High)
This paper presents a framework “that generates human
motions by cutting and pasting motion capture data” which look natural. The
generation of motion runs in real-time automatically allowing complex
interactive authoring. This approach can easily generate various motions that
interact with each other and it can be adapted to non-human characters.
Paper Evaluation
Interactive
Control of Avatars Animated with Human Motion Data
Lee, Chai, Reitsma, Hodgins, Pollard
Ratings: Difficulty: 7/10
(Moderate)
Interest: 6/10 (Moderate)
Real-time
animation and control of characters in 3d is the focus of this paper. The
approach suggests a flexible way of how to choose appropriate motion sequences
and plausible transitions from a database with human motion data. Useful
applications are shown and a variety of user-interfaces tested.
Paper Evaluation
Motion Capture
Assisted Animation: Texturing and Synthesis
Pullen, Bregler
Ratings: Difficulty: 3/10 (Easy)
Interest: 8/10 (High)
The papers details on the combination of key framing and motion capture for animation purposes. An initially rough animation done with key frames is refined with motion capture data to fill in missing degrees of freedom and details. The result is a high quality sequence almost like live motion.