My current research interests include computer vision, machine learning and improving the expressive nature of motion capture based animation. Extensive undergraduate work in applied mathematics and visual arts helped to shape my disparate interests. In particular, I worked on an undergraduate thesis on Bayesian techniques including the use of Bayesian nets for learning.
Graduate work has been quite varied. Research has been conducted briefly numerical optimization, bioinformatics, computer vision and mocap based animation. However, the underlying theme in all areas has been my original fascination with the application of statistical techniques for physical problems.
While studying at NYU, my life was permanently altered after taking a course in color theory at Parsons. This usually monotonous gut class changed how I saw the world and forged a sustained interest in human perception. The complexity and relative nature of vision fascinated me. I began to forge ideas of new ways to view certain computer vision problems.
Current projects include extensive work with the Condensation Algorithm (Condition Density Propagation), a tracking method that is successful in cluttered scenes and based on the old statistical technique of particle filters. Additional work have been completed studying SVMs, pattern recognition methods, numerical optimization and basic computer vision areas (stereo, edge detection, shape contexts). A long term collaborative project is in progress studying the use of LMA in the analysis motion capture data.
Below are some random projects.
- - ------------------- - - - -- ------------- --- --------------- -- --- --- ---------- - -- -- - ----- ------
speaking with hands :: siggraph 2004
scalability for kernel based methods
Reading List - coming soon