Seminars/Colloquia/Invited Talks

Seminars

Rene Vidal

Segmentation of Bilinear Surfaces

PLACE:	Clark 314
EVENT:	CIS Seminar Series
DATE:	April 26, 2005
TIME:	1:00 - 2:00

Abstract

In this talk, we look at the problem of segmenting a set of points lying on multiple bilinear surfaces.In particular, we look at three classes of surfaces that show up in 3-D motion segmentation problemsin computer vision: affine motion surfaces, essential manifolds and planar homographies. In each case, we show that the union of all surfaces is the zero set of a bi-homogeneous polynomial whose degree is the number of surfaces and whose factors are the bilinear forms defining each surface. The segmentation problem is then solved by fitting this polynomial to all data points and then factorizing this polynomial as a product of bilinear forms. Rather than performing polynomial factorization, we demonstrate that the factors can be obtained from the derivatives of the polynomial. This requires exploiting the geometric properties of the manifold associated with each model. Applications of the theory in 2-D and 3-D motion segmentation of dynamic scenes will also be presented. Joint work with Xiaodong Fan and Dheeraj Singaraju

Brief biography

Professor Vidal received his B.S. degree in Electrical Engineering (highest honors) from the Pontificia Universidad Catolica de Chile www.ing.puc.cl in 1997 and his M.S. and Ph.D. degrees in Electrical Engineering and Computer Sciences www.eecs.berkeley.edu from the University of California at Berkeley in 2000 and 2003, respectively. He was a research fellow at the National ICT Australia www.nicta.com.au since September 2003 and joined The Johns Hopkins University www.jhu.edu in January 2004 as an Assistant Professor in the Department of Biomedical Engineering www.bme.jhu.edu and the Center for Imaging Science www.cis.jhu.edu. His areas of research are *biomedical imaging* (DTI registration and clustering, heart motion analysis), *computer vision* (segmentation of static and dynamic scenes, multiple view geometry, omnidirectional vision), *machine learning* (generalized principal component analysis GPCA, kernel GPCA, dynamic GPCA), vision-based coordination and control of unmanned vehicles, and hybrid systems identification and control. Dr. Vidal is recipient of the 2005 NFS CAREER Award and the 2004 Best Paper Award Honorable Mention (with Prof. Yi Ma) for his work on "A Unified Algebraic Approach to 2-D and 3-D Motion Segmentation" presented at the European Conference on Computer Vision. He also received the 2004 Sakrison Memorial Prize www.coe.berkeley.edu/engnews/Spring04/EN16S/awards.html for "completing an exceptionally documented piece of research", the 2003 Eli Jury award www.coe.berkeley.edu/engnews/spring03/16S/otherawards.html for "outstanding achievement in the area of Systems, Communications, Control, or Signal Processing", the 2002 Student Continuation Award from NASA Ames, the 1998 Marcos Orrego Puelma Award from the Institute of Engineers of Chile, and the 1997 Award of the School of Engineering of the Pontificia Universidad Católica de Chile to the best graduating student of the school.