My lab is hosting Richard Souvenir this week. He will be giving a talk:
Abstract: Current approaches to evaluating designed spaces are limited by the time and expense of manual observation. In collaboration with architects and ethnographers, we are developing a system for sensing, storing, and analyzing human activity data over long time scales using a network of indoor cameras. The system relies on automated methods for people tracking and action recognition. In this talk, I will describe our approach to human detection and action recognition in multi-camera networks, which is a hybrid between (efficient) single-camera approaches and (accurate) multi-view methods. Our work is motivated by the observation that not all viewpoints in a multi-camera network are equal for recognizing actions and introduces dynamic viewpoint selection to improve processing time without sacrificing accuracy.
Bio: Richard Souvenir is an Associate Professor in the Department of Computer Science at The University of North Carolina at Charlotte and directs the Video and Image Analysis Lab. His research interests lie in understanding natural video without developing highly constrained models to improve computer vision tasks, such as segmentation, recognition, and tracking in various domains, including biomedical imaging and human motion analysis. Dr. Souvenir received his D.Sc. from Washington University in St. Louis in 2006.