Professor of Computer Science at University of Illinois at Urbana-Champaign. Biggest past contributions are in single-view geometry and 3D scene understanding, object recognition and attribute-based representations, and multitask representation and learning.

My current research directions:
XVG (miXed-View Geometry): How can we recover complete and precise 3D geometry for scenes viewed partly by multiple images and partly by single images? Our goal to develop a unifying framework for recognition-based single-view methods and correspondence-based multiview methods to produce compact, precise, complete, structured, and semantically rich 3D models.
General Purpose Vision: How can we make vision systems that easily learn and adapt to new tasks and environments? Our goal is to reformulate visual learning as continual, opportunistically supervised, self-directed, and minimizing asymptotic error and sample complexity for unknown future tasks.