An ecological approach to computational vision is developed which investigates the relation between the statistical structure of the visual environment and the properties of the mammalian visual pathway. A specific theory will be developed and tested based on the premise that the goal of early visual processing is to produce a response to the visual environment that is both sparse (a minimum number of 'active' cells) and distributed (all cells have equal response probability). The proposed studies will be directed along three parallel lines. 1) A general theory will be developed to account for how sparse distributed codes can be produced given an environment with statistical regularity. 2) Using computational models of the mammalian visual system this theory will be tested by investigating whether natural scenes have the necessary and sufficient conditions to produce sparse distributed activity 3) Using computational modeling and experiments involving human psycho-physics, specific hypotheses about the advantages of sparse distributed codes will be tested. By relating the behavior of cell populations to the statistical structure of the visual environment, we expect to gain a better insight into the nature of how the mammalian visual system functions under natural conditions. The computational studies along with proposed studies related to the perception of blur in natural environments, will provide a means for relating visual dysfunction (e.g., low vision) to underlying clinical abnormalities.

Agency
National Institute of Health (NIH)
Institute
National Institute of Mental Health (NIMH)
Type
First Independent Research Support & Transition (FIRST) Awards (R29)
Project #
5R29MH050588-03
Application #
2249942
Study Section
Cognitive Functional Neuroscience Review Committee (CFN)
Project Start
1993-07-01
Project End
1998-06-30
Budget Start
1995-07-01
Budget End
1996-06-30
Support Year
3
Fiscal Year
1995
Total Cost
Indirect Cost
Name
Cornell University
Department
Psychology
Type
Schools of Arts and Sciences
DUNS #
City
Ithaca
State
NY
Country
United States
Zip Code
14850
Kingdom, F A; Hayes, A; Field, D J (2001) Sensitivity to contrast histogram differences in synthetic wavelet-textures. Vision Res 41:585-98
Olshausen, B A; Field, D J (1997) Sparse coding with an overcomplete basis set: a strategy employed by V1? Vision Res 37:3311-25
Field, D J; Brady, N (1997) Visual sensitivity, blur and the sources of variability in the amplitude spectra of natural scenes. Vision Res 37:3367-83
Field, D J; Hess, R F (1996) Uncalibrated distortions vs undersampling. Vision Res 36:2121-4
Brady, N; Field, D J (1995) What's constant in contrast constancy? The effects of scaling on the perceived contrast of bandpass patterns. Vision Res 35:739-56