Amit K. Roy-Chowdhury (UC Riverside)
Title: Dynamic Scene Analysis in A Camera Network
As cameras get cheaper, large numbers of them are being installed in many applications. However, most of them transmit their videos to a central location where they are interpreted by human observers. This is infeasible as the number of cameras get larger. Therefore, it is very important to automatically extract meaningful patterns of dynamic events from the video sequences observed over the network.
In this talk, we will start with a high-level presentation of the multi-camera networks we are building at UCR. Thereafter, I will highlight the main issues that need to be addressed for this endeavor to be successful and focus on some of them. Specifically, I will describe two new results - (i) an appearance manifold of objects that is derived using a combination of analytically derived geometrical models and statistical data analysis, and (ii) a closed-loop tracking framework that can track people through different activities. I will conclude by giving an overview of some ongoing research in this area.
Bio: Amit K. Roy-Chowdhury has been an Assistant Professor of Electrical Engineering at the University of California, Riverside since January 2004. He completed his PhD in 2002 from the University of Maryland, College Park, where he also worked as a Research Associate in 2003. His research interests are in the broad areas of image processing and analysis, computer vision, video communications and machine learning. Currently, he is working on problems of pose and illumination invariant video-based object recognition, event analysis in large video networks, and multi-terminal video compression. Dr. Roy-Chowdhury has over fifty papers in peer-reviewed journals, conferences and edited books. He is an author of the book titled ``Recognition of Humans and Their Activities Using Video. He is a PI on a number of research grants from Federal and private agencies. He is on the program committee of most major conferences in computer vision and image/signal processing and is a regular reviewer for the main journals in these areas.