Dr. William (Bill) T. Freeman


Current title: Research Scientist

Mitsubishi Electric Research Labs
201 Broadway
Cambridge, MA 02139
freeman@merl.com

Research/Job interests
Various topics in computer vision and image processing. Bayesian models of perception. Hand gesture recognition. Color and shape perception. Exploiting the generic viewpoint assumption. Steerable filters and steerable pyramids. Image representation.

Vismod Advisor: Ted Adelson

Education
Ph.D., Media Arts and Sciences, MIT, 1992
M.S., Applied Physics, Cornell University, 1981
M.S., Electrical Engineering, Stanford University, 1979
B.S., Physics, Stanford University, 1979

Experience
1992 - present:
Research Scientist,
Mitsubishi Electric Research Labs.

1986 - 1987:
Foreign Expert, Taiyuan University of Technology, P. R. China.

1981 - 1986:
Principal Engineer, Polaroid Corporation.


Project/Research Description

The charter of Mitsubishi Electric Research Labs (MERL) is to explore uses of computers 10 and 20 years in the future. At MERL, I study problems in computer vision, and their applications. I am interested in Bayesian models of visual perception, currently focusing on color and shape perception. I am also interested in exploiting the generic viewpoint assumption, and related assumptions, in computational algorithms.

At MERL, I have developed real-time algorithms for hand gesture recognition, and for the control of appliances by hand gestures. I am interested in related problems of the human-computer interface.

Selected Publications

W. T. Freeman, "The generic viewpoint assumption in a framework for visual perception", Nature, vol. 368, p. 542 - 545, April 7, 1994.

D. H. Brainard and W. T. Freeman, "Bayesian method for recovering surface and illuminant properties from photosensor responses", in Human Vision, Visual Processing and Digital Display V, SPIE Proceedings Series, vol. 2179, 1994.

W. T. Freeman and M. Roth, "Orientation Histograms for Hand Gesture Recognition", MERL Technical Report 94-03, Mitsubishi Electric Research Labs, 201 Broadway, Cambridge, MA 02139, 1994.

W. T. Freeman, "Exploiting the generic view assumption to estimate scene parameters," IEEE International Conference on Computer Vision, Berlin, Germany, 1993.

W. T. Freeman, "Steerable filters and local analysis of image structure", Ph.D. thesis, Massachusetts Institute of Technology, 1992. Also as MIT Media Lab Vision and Modeling Group Technical Report 190.

E. P. Simoncelli, W. T. Freeman, E. H. Adelson, and D. J. Heeger, "Shiftable multi-scale transforms", IEEE Trans. Information Theory special issue on wavelets, vol. 38, no. 2, pp. 587 - 607, March, 1992.

W. T. Freeman and E. H. Adelson, "The design and use of steerable filters," IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 13, no. 9, pp. 891 - 906, September, 1991.

W. T. Freeman, E. H. Adelson, and D. J. Heeger, "Motion without movement," ACM Computer Graphics, vol. 25, no. 4, (SIGGRAPH '91), pp. 27 - 30, July, 1991.

W. T. Freeman and E. H. Adelson, "Steerable filters for early vision, image analysis, and wavelet decomposition," IEEE International Conference on Computer Vision, Osaka, Japan, 1990.

E. H. Adelson, E. P. Simoncelli, and W. T. Freeman, "Pyramids and multiscale representations," Proc. 13th European Conference on Visual Perception, Paris, 1990.

Project(s) in Vismod

Developed "steerable filters," a class of oriented filters with applications in image processing and computer vision. Designed a "steerable pyramid", a multi-scale, orientation-tuned image decomposition and explored applications for shape-from-shading and noise removal. Used steerable filters to classify junctions as X, T, or L. Based on that classification, developed algorithms to label contours as being caused by shading, occlusion or transparency. Developed "motion without movement", a novel motion display technique using steerable filters. Analyzed the generic viewpoint assumption in a Bayesian framework, which allows computer vision algorithms to exploit that assumption.


Personal Web Page


Vismod Home Page