A Hierarchical Approach to Motion Analysis and Synthesis for Articulated Figures 2000 . 2 . 18

Animating human-like characters is a recurring issue in computer graphics. Recently, capturing live motion has become one of the most promising technologies in character animation. Due to the success of this technology, realistic and highly detailed motion data are rapidly spread in computer graphics applications. Crafting animation with such data requires a rich set of specialized tools such as interactive editing, retargetting, blending, stitching, smoothing, enhancing, up-sampling, downsampling and so on. With those tools, animators combine motion clips together into animations of arbitrary length with great variety in character size, environment geometry, and scenario. A typical process of producing animation from live-captured data includes three major steps: The first step is to filter the raw input signal received from a motion capture system to obtain smoother, pleasing motion data. The next is to adapt the motion data to fit into a specific character or a virtual environment that may different from the live puppeteer or the environment, respectively, where the motion takes place. The motion segments thus obtained are combined into a seamless animation in the final step. In the thesis, we elaborate fundamental techniques that facilitate such tasks with proper consideration on orientation components and articulated structures. Specifically, three issues are addressed: We first present a general scheme of constructing spatial filters that perform smoothing and sharpening on orientation data. The next issue is to adapt an existing motion of a human-like character to have the desired features specified by a set of constraints. An efficient algorithm based on a hierarchical curve fitting technique and a fast inverse kinematics solver is provided. Finally, we present a multiresolution analysis scheme which decomposes a motion signal into a base level and a hierarchy of detail coefficients, followed by a synthesis scheme that produces a new motion by hierarchically combining detail coefficients of prescribed example motions.