Exploiting Low-Dimensional Structure in Image and Video Analysis

Dr. Lawrence Carin

Department of Electrical and Computer Engineering
Duke University

12 September 2011
Shelby Center 121
2:30 (Refreshements at 2:00)

Abstract

We examine leveraging low-dimensional latent structure for applications in image and video processing. The low-dimensional structure may be manifested in terms of sparse signal models, as well as low-rank representations. To address this problem, we develop nonparametric Bayesian models based on the Dirichlet and beta processes with techniques developed for fast inference. In addition to develop the theory and models, practical example results are presented on several challenging real-world applications.