Next: Stereoscopic Perception of Up: Computer Vision Previous: ShapesShocks, and

Optimal Control Methods in Computer Vision

Authors: [tex2html_wrap4134]B. Kimia (Brown University), A. Tannenbaum (U. of Minnesota), S.W. Zucker

Investigator username: zucker

Category: perception

Subcategory: computer vision

We are studying the employment of methods from optimal control for problems in computer vision and image processing. The underlying principle will be that of dynamic programming and the associated Hamilton-Jacobi equation which allows a unified approach to tackle a number of different issues in vision. In particular, we will consider problems concerning shape theory, morphology, optical flow, nonlinear scale-spaces, and shape-from-shading. We will also consider some numerical issues concerned with computer implementations of this methodology.


baron@cim.mcgill.ca