Graph Matching for Visual Correspondence
Event details
| Date | 01.12.2008 |
| Hour | 16:15 |
| Speaker | Dr. Lorenzo Torresani, Microsoft Research Cambridge, UK |
| Location |
INM202
|
| Category | Conferences - Seminars |
In this talk I will present a novel energy minimization approach to the problem of establishing correspondences between sparse image features related by an unknown non-rigid mapping and corrupted by clutter and occlusion. I will describe a correspondence objective formulated as a function of the appearance and the spatial arrangement of the features. Optimization of this energy is an instance of graph matching, which is in general a NP-hard problem. I will present a novel graph matching optimization technique based on the dual decomposition approach. In our evaluation this method outperforms existing graph matching techniques and in the majority of our examples is able to find the global minimum of our objective within a minute.
This talk reports on joint work with Vladimir Kolmogorov and Carsten Rother.
Bio:
Lorenzo Torresani is a Post Doc Researcher at Microsoft Research Cambridge. He received a Laurea Degree in Computer Science with summa cum laude honors from the University of Milan (Italy) in 1996, and an M.S. and a Ph.D. in Computer Science from Stanford University in 2001 and 2005, respectively. He worked at Riya, Inc., at the Courant Institute of New York University, and at Digital Persona, Inc. His research interests are in computer vision, machine learning, and computer animation. He was the recipient of the Best Student Paper Prize at the IEEE Conference on Computer Vision and Pattern Recognition 2001. In January 2009, he will join the Computer Science Department at Dartmouth College as an assistant professor.
L. Torresani's homepage
Practical information
- General public
- Free