Reprints 101-120

Date:
1980-1992
Reference:
PP/CRI/M/1/6
Catalogue details

Licence: In copyright

Credit: Reprints 101-120. Source: Wellcome Collection.

    1/194
    ^*1 An Information-Processing Approach to Understanding the Visual Cortex Francis H. C. Crick, David C. Marr, and Tomaso Poggio abstract An outline description is given of the experimental work on the visual acuity and hyperacuity of human beings. The very high resolution achieved in hyperacuity corresponds to a fraction of the spacing between adja cent cones in the fovea. We briefly outline a computational theory of early vision, according to which (a) the retinal image is filtered through a set of approximately band-pass spatial filters and (b) zero-crossings are detected independently in the outputs of these channels. Zero-crossings may contain sufficient information for much of the subsequent processing. Consideration of the optimal filter leads to one which is equivalent to a cell with a particular center-surround type of response. An edge in the visual field then corre sponds to a line of zero-crossings in the filtered image. The mathematics of sampling and of Logan's zero-crossing theorem are briefly explained. In this framework we suggest, similarly to Barlow (1979), that the fine grid of small cells in layer TVCß of the striate cortex perform an approximate recon struction of the filtered image with the goal of representing the position of zero-crossings with a very high accuracy (around a few seconds of arc). How this might be achieved is discussed in outline. Finally it is mentioned that this picture is probably too static, as hyperacuity can be achieved with moving targets. To make for easier comprehension much of the mathematics is set out in an appendix. Introduction That we see the world as well as we do is something of a miracle. What seems so direct and effortless turns out, on close consideration, to in volve many rapid and complex processes the details of which we are only beginning to glimpse. In a series of papers Marr and his col laborators have outlined a general strategy for approaching some of these problems (Marr, 1976,1977; Marr and Nishihara, 1978b; Marr and Poggio, 1977; Ullman, 1979). Two useful introductions are the general account by Marr and Nishihara (1978a) and a sympathetic review in Nature by Sutherland (1979). Other papers by Marr and his coworkers f. h. c. crick, d. c. marr , and t. poggio The Salk Institute for Biological Studies, P. O. Box 1809, San Diego, California 92112; M.I.T. Psychology Department, 79 Amherst Street, Cambridge, Massachusetts 02139; Max-Planck-Institut fuer biologische Kybernetik, 74 Tuebingen 1, Spemannstrasse 38, Germany