Abstract
Optical systems of associative memory models are capable of reconstructing lost information from incomplete input patterns. In this paper, we evaluate the ability of associative memory models when used in practical pattern recognition systems. Effects of improvement techniques, including defocusing preprocessing, and the recursive use of the associative operation are also examined with computer simulations. Finally the effectiveness of the recognition system is tested experimentally with alphabetic characters in different illumination conditions.
© 1990 Optical Society of America
Full Article | PDF ArticleMore Like This
Jasprit Singh and Songcheol Hong
Appl. Opt. 29(11) 1682-1688 (1990)
Masatoshi Ishikawa, Naohisa Mukohzaka, Haruyoshi Toyoda, and Yoshiji Suzuki
Appl. Opt. 29(2) 289-295 (1990)
Haruyoshi Toyoda, Naohisa Mukohzaka, Yoshiji Suzuki, and Masatoshi Ishikawa
Appl. Opt. 32(8) 1354-1358 (1993)