Cooperative Computation of Stereo Disparity

A cooperative algorithm is derived for extracting disparity information from stereo image pairs
  • D. Marr
  • T. Poggio


Perhaps one of the most striking differences between a brain and today’s computers is the amount of “wiring.” In a digital computer the ratio of connections to components is about 3, whereas for the mammalian cortex it lies between 10 and 10,000 (1).


Dimensional Image Subjective Contour Combinatorial Analysis Correspondence Problem Global Order 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References and Note

  1. 1.
    D. A. Sholl, The Organisation of the Cerebral Cortex (Methuen, London, 1956 ). The comparison depends on what is meant by a component. We refer here to the level of a gate and of a neuron, respectively.Google Scholar
  2. 2.
    A. N. Chomsky, Aspects of the Theory of Syntax ( MIT Press, Cambridge, Mass., 1965 ).Google Scholar
  3. 3.
    D. Marr and T. Poggio, Neurosci. Res. Prog. Bull.,in press (also available as Mass. Inst. Technol. Actif. Intell. Lab. Memo 357).Google Scholar
  4. 4.
    H. Haken, Ed. Synergetics-Cooperative Phenomena in Multicomponent Systems ( Teuber, Stuttgart, 1973 ).Google Scholar
  5. 5.
    H. Haken, Rev. Mod. Phys. 47, 67 (1975).CrossRefGoogle Scholar
  6. 6.
    J. D. Cowan, Prog. Brain Res. 17, 9 (1965).CrossRefGoogle Scholar
  7. 7.
    H. R. Wilson and J. D. Cowan, Kybernetik 13, 55 (1973).CrossRefGoogle Scholar
  8. 8.
    M. Eigen, Naturwissenschgßen 58, 465 (1971).CrossRefGoogle Scholar
  9. 9.
    P. H. Richter, paper contributed to a corn-petition of the Bavarian Academy of Science, Max-Planck-Institut filr Biophysikalische Chemie, 1974.Google Scholar
  10. 10.
    A. Gierer and H. Meinhardt, Kybernetik 12, 30 (1972).CrossRefGoogle Scholar
  11. 11.
    B. Julesz, Foundations of Cyclopean Perception ( Univ. of Chicago Press, Chicago, 1971 ).Google Scholar
  12. 12.
    K. Mori, M. Kidode, H. Asada, Comp. Graphics Image Process. 2, 393 (1973).CrossRefGoogle Scholar
  13. 13.
    D. Mart, Mass. Inst. Technol. Artif. Intell. Lab. Memo 327 (1974).Google Scholar
  14. 14.
    D. Mart, Philos. Trans. R. Soc. London Ser. B 275, 483 (1976).CrossRefGoogle Scholar
  15. 15.
    D. Mart, ibid 252, 23 (1971). See especially sec- tion 3. 1. 2.Google Scholar
  16. 16.
    D. Mart and T. Poggio, in preparation.Google Scholar
  17. 17.
    T. Poggio and W. Reichardt, Q. Rev. Biophys., in press.Google Scholar
  18. 18.
    W. Richards, J. Opt. Soc. Am. 62, 410 ( 1970.Google Scholar
  19. 19.
    H. B. Barlow, C. Blakemore, J. D. Pettigrew, J. Physiol. (London) 193, 327 (1967)Google Scholar
  20. J. D. Pettigrew, T. Nikara, P. O. Bishop, Exp. Brain Res. 6, 391 (1968)Google Scholar
  21. C. Blakemore, J. Physiol. (London) 209, 155 (1970).Google Scholar
  22. 20.
    B. Julesz and J.-J. Chang, Biocybernetics 22, 107 (1976).Google Scholar
  23. 21.
    D. H. Hubel and T. N. Wiesel, Nature (London) 225, 41 (1970).CrossRefGoogle Scholar
  24. 22.
    K. G. Wilson, Rev. Mod. Phys. 47, 773 (1975).CrossRefGoogle Scholar
  25. 23.
    Julesz (11), Cowan (6), and Wilson and Cowan (7) were the first to discuss explicitly the cooperative aspect of visual information processing. Much has been published recently on possible cooperative processes in nervous systems, ranging from the “catastrophe” literature [E. C Zeeman, Sci. Am. 234, 65 (April 1976)] to various attempts of more doubtful credibility. There has hitherto been no careful study of a cooperative algorithm in the context of a carefully defined computational problem [but see (15)], although algorithms that may be interpreted as cooperative were discussed, for instance, by P. Dev [Int. J. Man-Mach. Stud. 7, 511 (1975)] and by A. Rosenfeld, R. A. Hummel, and S. W. Zucker [Syst. Man. Cybern. 6, 420 (1976)]. In particular neither Dev nor J. I. Nelson [J. Theor. Biol. 49 1 (1975)] formulated the computational structure of the stereo-disparity problem. As a consequence, the resulting geometry of the inhibition between their disparity detectors does not correspond to ours (Fig. 2c) and apparently fails to provide a satisfactory algorithm.Google Scholar
  26. 24.
    S. Ullmann, Mass. Inst. Technol. Artif. Intell. Lab. Memo 367 (1967); also in Biol. Cybern.,in press.Google Scholar
  27. 25.
    We thank W. Richards for valuable discussions, H. Lieberman for making it easy to create stereograms, and K. Prendergast for preparing the figures. The research described was done at the Artificial Intelligence Laboratory of the Massachusetts Institute of Technology. Support for the laboratory’s artificial intelligence research is provided in part by the Advanced Research Projects Agency of the Department of Defense under Office of Naval Research contract N00014–75-C-0643; T.P. acknowledges the support of the Max-Planck-Gesellschaft during his visit to the Massachusetts Institute of Technology.Google Scholar

Copyright information

© American Association for the Advancement of Science 1976

Authors and Affiliations

  • D. Marr
    • 1
  • T. Poggio
    • 2
  1. 1.Artificial Intelligence LaboratoryMassachusetts Institute of TechnologyCambridgeUSA
  2. 2.Max-Planck Institut für Biologische Kybernetik74 Tubingen 1Germany

Personalised recommendations