Event Location: NSH 1507
Bio: Vijay Mahadevan is Research Scientist at Yahoo Labs in Sunnyvale, CA. He received the Ph.D degree in Electrical Engineering from UC San Diego. He also has an M.S. from Rensselaer, and a B.Tech from Indian Institute of Technology, Madras. His interests are in the areas of computer vision, machine learning and the understanding of human vision.

Abstract: The talk will be in two parts. In the first, I will provide an overview of some important computer vision problems that we are trying to solve at Yahoo Labs. In the second part, I will describe a novel approach for salient object detection, when the image is represented as a graph over the set of its superpixels. In this approach, the saliency of a set of seed locations is propagated through the graph via a diffusion process. But unlike previous heuristic approaches for seed selection, an optimal set of salient seeds is learned using a large margin formulation that best discriminates the object superpixels from those of the background.