Vision is to know what is where by looking. I will focus on visual recognition, i.e. the classification of objects, scenes and behaviors from images and video. Starting from a brief history of the subject I will present current approaches to visual recognition and highlight algorithmic and theoretical challenges. I will demonstrate the current recognition capabilities of computer vision systems and discuss how they can interact with communities of experts to achieve better-than-human performance. I will point out current limitations, and avenues for progress towards building intelligent machines.
Pietro Perona is interested in artificial intelligence and computational vision. His work includes partial differential equations for image processing (anisotropic diffusion), modeling attention and visual search, visual categorization, analyzing behavior using vision, and designing "communities of knowledge" composed of data, people and machines (the Visipedia project). Perona is currently the Allen E. Puckett Professor of Electrical Engineering and of Computation and Neural Systems with the California Institute of Technology (Caltech), and Amazon Fellow. Perona graduated in Electrical Engineering at the University of Padova, Italy. He received a Ph.D. in Electrical Engineering and Computer Science from the University of California at Berkeley. He has been a faculty at Caltech since 1991.
More infos are available here: