Bayesian approaches have enjoyed a great deal of recent success in their application to problems in computer vision. This success has led to an emerging interest in applying Bayesian methods to modeling human visual perception.
We consider the implications of a Bayesian view of visual information processing for experimentally investigating human visual perception. We have outlined the elements of a general program of empirical research which results from taking the basic Bayesian formulation seriously not only as a means for objectively modeling image information through ideal observer analysis (e.g. see our work in object recognition), but also as a framework for characterizing human perceptual inference. A major advantage of following such a program is that, because its structure is the same as that of the Bayesian framework for computational modeling, it supports a strong integration of psychophysics and computational theory. In particular, it provides the foundation for a psychophysics of constraints in which one tests hypotheses regarding quantitative and qualitative constraints used in human perceptual inferences. The Bayesian approach also suggests new ways to conceptualize the general problem of perception and to decompose it into isolatable parts for psychophysical investigation; that is, it not only provides a framework for modeling solutions to specific perceptual problems; it also guides the definition of the problems.
More information can be found in the following papers from our publication list:
Early work is here:
Kersten, D. (1990). Statistical limits to image understanding. In C. Blakemore (Ed.), Vision: Coding and Efficiency Cambridge: Cambridge University Press.
Kersten, D. (1991) Transparency and the cooperative computation of scene attributes.
Computational Models of Visual Processing, M.I.T. Press, Landy M and Movshon
A., Eds..
Knill, D. C., Kersten, D. K. (1991). Ideal Perceptual Observers for Computation, Psychophysics, and Neural Networks. In R. J. Watt (Ed.), Pattern Recognition by Man and Machine MacMillan Press.
For more recent work, see:
Schrater, P. R., & Kersten, D. (submitted). The Role of Task Specification in Optimal Cue Integration. International Journal of Computer Vision, under review. (draft pdf)
Kersten, D. & Schrater, P. R., (submitted). Pattern Inference Theory: A
Probabilistic Approach to Vision. In R. Mausfeld, & D. Heyer (Ed.), Perception
Theory: Conceptual Issues Chichester: John Wiley & Sons, Ltd. (Draft
pdf)
For an application to the perception of color and shape, see:
Bloj, M., Kersten, D., & Hurlbert, A. C. 3D Shape Perception Influences Colour Perception via Mutual Illumination. Nature. (pdf)
e-mail: kersten@eye.psych.umn.edu