Ce trebuie să știi

The contributors bring a wide range of methodologies to bear on the common problem of image-based object recognition. These interconnected essays on three-dimensional visual object recognition present cutting-edge research by some of the most creative neuroscientific, cognitive, and computational scientists in the field. Cassandra Moore and Patrick Cavanagh take a classic demonstration, the perception of "two-tone" images, and turn it into a method for understanding the nature of object representations in terms of surfaces and the interaction between bottom-up and top-down processes. Michael J. Tarr and Isabel Gauthier use computer graphics to study whether viewpoint-dependent recognition mechanisms can generalize between exemplars of perceptually defined classes. Melvyn A. Goodale and G. Keith Humphrey use innovative psychophysical techniques to investigate dissociable aspects of visual and spatial processing in brain-injured subjects. D.I. Perrett, M.W. Oram, and E. Ashbridge combine neurophysiological single-cell data from monkeys with computational analyses for a new way of thinking about the mechanisms that mediate viewpoint-dependent object recognition and mental rotation. Shimon Ullman also addresses possible mechanisms to account for viewpoint-dependent behavior, but from the perspective of machine vision. Finally, Philippe G. Schyns synthesizes work from many areas, to provide a coherent account of how stimulus class and recognition task interact.