Share Email Print

Proceedings Paper

Terminal attractor optical associative memory for pattern recognition
Author(s): Xin Lin; Masahiko Mori; Junji Ohtsubo; Masanobu Watanabe
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Optical associative memory for pattern recognition using the Hopfield model has the advantage of simplicity for its network structure. But further investigation reveals that the basins of the attractors for stored patterns are small in the Hopfield model and the recalling ability is not so great. Moreover, the memory capacity is low because there are many spurious states and oscillations in the network. In order to obviate the spurious states and low storage capacity in the Hopfield model, a type of attractor called a terminal attractor (TA), which represents singular solutions of a neural dynamic system, was introduced. These terminal attractors are characterized by having finite relaxation times, no spurious states, and infinite stability.

Paper Details

Date Published: 28 June 1999
PDF: 3 pages
Proc. SPIE 3864, Joint International Symposium on Optical Memory and Optical Data Storage 1999, 38641A (28 June 1999); doi: 10.1117/12.997584
Show Author Affiliations
Xin Lin, Electrotechnical Lab. (Japan)
Masahiko Mori, Electrotechnical Lab. (Japan)
Junji Ohtsubo, Shizuoka Univ. (Japan)
Masanobu Watanabe, Electrotechnical Lab. (Japan)

Published in SPIE Proceedings Vol. 3864:
Joint International Symposium on Optical Memory and Optical Data Storage 1999
Shigeo R. Kubota; Ryuichi Katayama; Douglas G. Stinson, Editor(s)

© SPIE. Terms of Use
Back to Top