On seeing stuff: the perception of materials and surfaces

Objects are «things» but they are made of «stuff.» In both human and machine vision, there has been much research on object recognition, but rather little on material perception. My lab is exploring material perception at various levels. At the low-level, we have identified certain image statistics, such as subband skewness, that are correlated with surface properties such as gloss. We suggest that there are mechanisms in early vision that are sensitive to such statistics, which could be easily computed with neural mechanisms. For our high-level studies, we are evaluating material recognition (categorization): the ability to look at something and decide whether it is, say, leather or plastic or cloth. We have assembled some image databases for use in exploring material recognition. In humans, material recognition can occur at high speed, even when the images are quite diverse. We have used machine vision and machine learning techniques in an attempt to achieve automated categorization of images in our database. While our system outperforms prior systems, it still falls far short of human performance.

Speaker Bios

Edward Adelson is Professor of Vision Science at the Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, and is a member of the Computer Science and Artificial Intelligence Lab. He has a B.A. in Physics and Philosophy from Yale University, and a Ph. D. in Experimental Psychology from the University of Michigan. After a post-doc at NYU, he worked on human and computer vision at RCA Sarnoff Labs in Princeton. He joined the MIT Media Lab in 1987, and moved to the Department of Brain and Cognitive Sciences in 1995. He has published numerous papers in the fields of human perception, computer vision, visual neuroscience, and image processing, and holds a dozen patents on image processing and video coding. His current interests include mid-level vision, motion perception, and the perception of materials and surfaces. His work has been honored with the Optical Society of America¹s Adolph Lomb Medal, and the Rank Prize in Electro-optics.