Wilson S. Geisler

Biosketch

Wilson (Bill) Geisler holds the David Wechsler Regents Chair in Psychology at the University of Texas at Austin. He founded the Center for Perceptual Systems and was its director for over two decades. He is a vision scientist recognized for his experimental and theoretical studies of visual behavior, visual computation, and visual neuroscience in humans and non-human primates. Dr. Geisler was born in Palo Alto, California and grew up in the bay area. He obtained his undergraduate degree in psychology from Stanford University in 1971, and his doctoral degree in experimental and mathematical psychology from Indiana University in 1975. He joined the psychology faculty at the University of Texas in 1975.

Research Interests

I study perception and perceptual neuroscience, with an emphasis on vision in humans and non-human primates. My early years were directed at understanding the relationship between the retinal physiology of light and dark adaptation and the behavioral detection and discrimination performance of humans. This was followed by a series of studies directed at understanding the role of optical and retinal factors in limiting human spatial and color vision. These studies pioneered the application of (Bayesian) ideal observer theory to domains beyond simple photon detection and intensity discrimination. More recently the lab has been focused on spatial and contrast coding in the primate visual cortex, on natural tasks and natural scene statistics, and on the mathematics of how to perform perceptual tasks optimally. For example, we derived the theory for how to move the eyes optimally when searching for targets in natural texture and are using that theory to analyze human performance and eye movement statistics. For another example, we measured the statistical properties of contours in natural scenes, derived the theory of how to optimally use those properties to detect and interpolate contours in natural images and are using the theory to analyze human contour detection and interpolation performance.