            |
GRASP Seminar Series: Spring 2008
March 7, 11:00 a.m., Wu & Chen Auditorium, Levine Hall (3330 Walnut Street)
Ronen Basri
Weizmann Institute of Science
"Approximate Nearest Subspace Search with Applications to Pattern Recognition"
Abstract: Linear and affine subspaces are commonly used to describe the appearance of objects under different lighting, viewpoint, articulation, and even identity. A natural problem arising from their use is -- given a query image portion represented as a point in some high dimensional space -- find a subspace near to the query. This talk presents an efficient solution to the approximate nearest subspace problem for both linear and affine subspaces. Our method is based on a simple reduction to the problem of nearest point search, and can thus employ tree based search or locality sensitive hashing to find a near subspace. Further speedup is achieved by using random projections to lower the dimensionality of the problem. We provide theoretical proofs of correctness and error bounds of our construction and demonstrate its capabilities on synthetic and real data. Our experiments demonstrate that an approximate nearest subspace can be located significantly faster than the exact nearest subspace, while at the same time it can find better matches compared to a similar search on points, in the presence of variations due to viewpoint, lighting etc.
Joint work with Tal Hassner and Lihi Zelnik-Manor
Biography: Dr. Ronen Basri received his BSc in Mathematics
and Computer Science from Tel Aviv University in 1986 and his PhD in
Computer Science from the Weizmann Institute of Science in 1991. He later
was a post doctoral fellow at the Massachusetts Institute of Technology
in the Department of Brain and Cognitive Science and the Artificial Intelligence
Laboratory. Since 1992 he has been affiliated with the Weizmann Institute
of Science in the Department of Computer Science and Applied Mathematics,
where he is currently holding the position of Professor and lately has
been serving as the chair for this department. Between 1999 and 2000
he spent a sabbatical at NEC Research Institute in Princeton, New Jersey.
Starting August 2007 he is serving as a
visiting professor at the Toyota Technological Institute in Chicago.
Dr. Basri's research has focused on computer vision, especially in the areas of image segmentation, 3D shape reconstruction, and object recognition. Among others, he has developed methods for analyzing the effect of lighting on images, multiscale algorithms for image segmentation and edge detection, and region-based pose estimation techniques. Dr. Basri's work deals with the development of algorithms, analysis, and implications to human vision.
GRASP
Laboratory
Site maintained by graspadm@grasp.cis.upenn.edu
Last update:23 January, 2008