My research focuses on developing computationally efficient algorithms to construct models of physical phenomena from massive amounts of noisy, ambiguous data. My approach is to pair theoretically sound, probabilistic, model-based reasoning with randomized approximation methods to attain efficiency.
I have exploited this theme in two major research projects prior to my thesis research: Texture-Based Tracking and
Monte Carlo Localization
(joint work with Dieter Fox). In my thesis research I exploit these same principles, but apply them to a problem of vastly greater complexity. In particular, I provide an efficient solution to the problem of building a 3D world model from 2D images, without correspondence.