Ph.D., Massachusetts Institute of Technology
Professor
Department of Cognitive and Linguistic Sciences
229 Metcalf Research Laboratory
Tel: (401) 863-1148
Email: Michael_Tarr@Brown.edu
Go to lab web page

My research examines mechanisms of visual object representation, recognition and categorization using psychophysical, neuroscientific, and computation methods. Questions I am interested in include: How does the brain visually represent and recognize three-dimensional objects across dramatic changes in the two-dimensional retinal image, including variations produced by changes in viewpoint and illumination? What cognitive and neural mechanisms are used in acquisition of visual expertise and the ability to recognize objects at both highly specific and general categorical levels? How do we mentally represent and utilize information about the complex configurations of objects in scenes? As an example, consider that almost every pixel in an image of an object changes if there is a change in lighting direction. Yet, human observers are incredibly good at discounting this variation and successfully recognizing objects regardless of such changes. My research suggests that high-level mental representations of objects include information about the effects of lighting. The reason for this is straightforward: lighting effects tell us something about the 3D structure of objects and, thus, this information is valuable for recognition, particularly in the case of novel objects. We are presently pursuing how the visual system, which is encoding this highly variable information, compensates for the changes due to lighting that are likely to occur in the image. In another project, several collaborators, and I have been exploring how observers acquire visual expertise as a means for understanding the flexibility of the human visual system across a wide range of visual recognition tasks.

Tarr, M. J., & Kriegman, D. J. (2001). What defines a view? Vision Research, 41(15), 1981-2004.

Tarr, M. J., & Gauthier, I. (2000). FFA: A Flexible Fusiform Area for subordinate-level visual processing automatized by expertise. Nature Neuroscience, 3 (8), 764-769.

Tarr, M. J. (In press). Visual Object Recognition: Can a Single Mechanism Suffice? In M. A. Peterson & G. Rhodes (eds.), Analytic and Holistic Processes in the Perception of Faces, Objects, and Scenes, JAI/Ablex.