Abstract

This article describes a system for pattern recognition made up of electronic logic elements which has many of the characteristics of humans for recognizing patterns. In particular, the recognition is invariant for size and will tolerate a specified amount of tilt or figure rotation. The electronic components or organs which make up this system consist of delay lines, logical elements such as “and” circuits, and photoreceptors.

© 1961 Optical Society of America

Full Article  |  PDF Article

References

  • View by:
  • |
  • |
  • |

  1. W. Pitts and W. McCulloch, Bull. Math. Biophys. 9, 127–147 (1947).
    [Crossref] [PubMed]
  2. J. R. Singer, “A Computer Model of Visual Pattern Recognition in Humans,” paper given at the URSI Meeting, Washington, D. C., May 25, 1957.
  3. J. R. Singer, J. Opt. Soc. Am. 47, 205–207 (1957).
    [Crossref] [PubMed]
  4. J. R. Singer, “Information Theory Applied to the Human Visual System,” paper given at the First National Biophysics Conference at Columbus, Ohio, 1957.
  5. J. R. Singer, J. Opt. Soc. Am. 49, 639–640 (1959).
    [Crossref] [PubMed]
  6. J. von Neumann, Automata Studies, edited by Shannon and McCarthy (Princeton University Press, 1956), p. 43.
  7. Frank Rosenblatt, Proc. I.R.E.,  48, 301–309 (1960).
    [Crossref]
  8. O. G. Selfridge and D. Gold, and Lincoln Laboratory Group 54, Quart. Prog. Rept., Information Processing Report of Lincoln Laboratory (December1959) (also see some proceeding reports by this group).
  9. O. G. Selfridge, Proc. Western Joint Computer Conf. (March, 1955).
  10. L. A. Kamentsky, Proc. Western Joint Computer Conf. 304–309 (March, 1959).

1960 (1)

Frank Rosenblatt, Proc. I.R.E.,  48, 301–309 (1960).
[Crossref]

1959 (2)

L. A. Kamentsky, Proc. Western Joint Computer Conf. 304–309 (March, 1959).

J. R. Singer, J. Opt. Soc. Am. 49, 639–640 (1959).
[Crossref] [PubMed]

1957 (1)

1955 (1)

O. G. Selfridge, Proc. Western Joint Computer Conf. (March, 1955).

1947 (1)

W. Pitts and W. McCulloch, Bull. Math. Biophys. 9, 127–147 (1947).
[Crossref] [PubMed]

Gold, D.

O. G. Selfridge and D. Gold, and Lincoln Laboratory Group 54, Quart. Prog. Rept., Information Processing Report of Lincoln Laboratory (December1959) (also see some proceeding reports by this group).

Kamentsky, L. A.

L. A. Kamentsky, Proc. Western Joint Computer Conf. 304–309 (March, 1959).

McCulloch, W.

W. Pitts and W. McCulloch, Bull. Math. Biophys. 9, 127–147 (1947).
[Crossref] [PubMed]

Pitts, W.

W. Pitts and W. McCulloch, Bull. Math. Biophys. 9, 127–147 (1947).
[Crossref] [PubMed]

Rosenblatt, Frank

Frank Rosenblatt, Proc. I.R.E.,  48, 301–309 (1960).
[Crossref]

Selfridge, O. G.

O. G. Selfridge, Proc. Western Joint Computer Conf. (March, 1955).

O. G. Selfridge and D. Gold, and Lincoln Laboratory Group 54, Quart. Prog. Rept., Information Processing Report of Lincoln Laboratory (December1959) (also see some proceeding reports by this group).

Singer, J. R.

J. R. Singer, J. Opt. Soc. Am. 49, 639–640 (1959).
[Crossref] [PubMed]

J. R. Singer, J. Opt. Soc. Am. 47, 205–207 (1957).
[Crossref] [PubMed]

J. R. Singer, “A Computer Model of Visual Pattern Recognition in Humans,” paper given at the URSI Meeting, Washington, D. C., May 25, 1957.

J. R. Singer, “Information Theory Applied to the Human Visual System,” paper given at the First National Biophysics Conference at Columbus, Ohio, 1957.

von Neumann, J.

J. von Neumann, Automata Studies, edited by Shannon and McCarthy (Princeton University Press, 1956), p. 43.

Bull. Math. Biophys. (1)

W. Pitts and W. McCulloch, Bull. Math. Biophys. 9, 127–147 (1947).
[Crossref] [PubMed]

J. Opt. Soc. Am. (2)

Proc. I.R.E. (1)

Frank Rosenblatt, Proc. I.R.E.,  48, 301–309 (1960).
[Crossref]

Proc. Western Joint Computer Conf. (1)

O. G. Selfridge, Proc. Western Joint Computer Conf. (March, 1955).

Proc. Western Joint Computer Conf. 304–309 (1)

L. A. Kamentsky, Proc. Western Joint Computer Conf. 304–309 (March, 1959).

Other (4)

O. G. Selfridge and D. Gold, and Lincoln Laboratory Group 54, Quart. Prog. Rept., Information Processing Report of Lincoln Laboratory (December1959) (also see some proceeding reports by this group).

J. von Neumann, Automata Studies, edited by Shannon and McCarthy (Princeton University Press, 1956), p. 43.

J. R. Singer, “Information Theory Applied to the Human Visual System,” paper given at the First National Biophysics Conference at Columbus, Ohio, 1957.

J. R. Singer, “A Computer Model of Visual Pattern Recognition in Humans,” paper given at the URSI Meeting, Washington, D. C., May 25, 1957.

Cited By

OSA participates in Crossref's Cited-By Linking service. Citing articles from OSA journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (10)

Fig. 1
Fig. 1

The numeral in the circle indicates the number of pulses necessary to obtain an output pulse from the circular delay element. (a) a simple delay; (b) an “and” circuit; (c) another type of “and” circuit; at least three of the four inputs must be active to obtain an output pulse at x; (d) an “or” circuit; (e) a bridge circuit; input line a must be stimulated while b is not stimulated in order to obtain an output at x.

Fig. 2
Fig. 2

A method of combining the output from four photoreceptors (a, b, c and d) so as to obtain an output signal if one, two, or three receptors are stimulated, but not if all four are stimulated. This is a type of differentiator to emphasize image borders. In the model sets of 16 photoreceptors are differentiated also; the logical analysis is shown in Appendix A.

Fig. 3
Fig. 3

A specific model of 72 photoreceptors (represented by circles) differentiated four at a time, and also in clusters of 16, to provide increased sensitivity to image borders. The solid black circles represent 36 optic fibers to carry the differentiated signal pulses. The logical analysis of this system is given in Appendix A.

Fig. 4
Fig. 4

A specific model for a delay transformer to transform 36 optic fibers into delay space. Note that signals having spatial relationships in the optic fibers retain these relationships with a transformation into time intervals. Each unit of length for an image transforms into a specified unit of time. The dark disks represent optic fibers, and the delay transformer is interior to the cog-shaped borderline. Delay space is outside the borderline.

Fig. 5
Fig. 5

Delay space without the delay transformer. A distributed parameter system (delay line) would also serve to define a delay space. Note that the system may be bent into any desired physical configuration.

Fig. 6
Fig. 6

A circle recognition circuit. It should be noted that the c and d lines are spaced differently in their radial distances.

Fig. 7
Fig. 7

A symbol which will be named “theta.”

Fig. 8
Fig. 8

A set of recognition coincidences for both a circle and a theta.

Fig. A1
Fig. A1

A logical combination of 16 photoreceptors to obtain the desired differentiation. This is the iterated scheme used to obtain the system shown in Fig. 3. The black disks are differentiated output lines.

Fig. B1
Fig. B1

An example of a delay transformer which completely isolates the optic fibers. (Optic fibers are indicated by black disks.) The delays are unidirectional elements, and no interior signal will appear on surrounding optic fibers. This may be a more practicable arrangement than the delay transformer of Fig. 3.

Equations (1)

Equations on this page are rendered with MathJax. Learn more.

a b ¯ c ¯ d ¯ + ā b c ¯ d ¯ + ā b ¯ c d ¯ + ā b ¯ c ¯ d + a b c ¯ d ¯ + a b ¯ c d ¯ + a b ¯ c ¯ d + ā b c d ¯ + ā b c ¯ d + ā b ¯ c d + a b c d ¯ + a b c ¯ d + a b ¯ c d + ā b c d = ( a b c d ¯ ) ( a + b + c + d ) = X