Location
CEWIT 200
Event Description
We are very excited to start our Distinguished lecture series this Friday with Ramin Zabih.
Title: Solving Medical Deconvolution
Abstract: Graph cut methods have proved very successful for many problems in computer vision and graphics, but they are largely restricted to problems with first-order (pairwise) interactions between pixels. I will describe a difficult and important deconvolution problem that arises in medical imaging. Problems like this clearly require significant advances in inference algorithms. I will present several higher-order graph cut techniques that my group has recently developed, and show some preliminary experimental results. This is joint work with several co-authors, primarily Alex Fix and Ashish Raj.
Short bio: Ramin Zabih is a professor of Computer Science at Cornell's new NYC campus, and also holds a joint appointment in the Radiology department at Weill Cornell Medical College. His research interests are in discrete optimization techniques for computer vision, and in applications to medicine. The graph cut methods that his research group developed, which are widely used in both academia and industry, have received several awards. He has held a number of positions in the vision community, including Program Chair for CVPR (2007) and General Chair for both CVPR (2013) and ECCV (2018). Zabih served as Editor-in-Chief of the IEEE Transactions on Pattern Analysis and Machine Intelligence from 2009-2012. He is a fellow of the ACM and the IEEE.
