Self organizing neural network method and system for general classification of patterns

作者: Michael Kuperstein

DOI:

关键词:

摘要: A neural network system and method that can adaptively recognize each of many pattern configurations from a set. The learns maintains accurate associations between signal classes with training teaching mechanism. classifying consists distributed input processor an adaptive association processor. decomposes into modules localized contextual elements. These elements in turn are mapped onto using self-organizing associative scheme. mapping determines which class best represents the pattern. computation is done through gating correspond to Learning achieved by modifying true/false response computed probabilities for all parallel fault tolerant process. It easily be extended accommodate arbitrary number patterns at degree precision. classifier applied automated recognition inspection different types signals patterns.

参考文章(6)
Leon N. Cooper, Douglas L. Reilly, Charles Elbaum, Christopher L. Scofield, Parallel, multi-unit, adaptive, nonlinear pattern class separator and identifier ,(1986)
Bruce S. Buckley, Self-organizing circuits ,(1986)
Douglas L. Reilly, Charles Dr. Elbaum, Leon N. Dr. Cooper, Self organising general pattern class separator and identifier ,(1981)
Hiroyuki Tsuboi, Yoichi Takebayashi, Kenichi Maeda, Pattern recognition device ,(1985)
Kurosu Yasuo, Okazawa Koichi, Yokoyama Yoshihiro, Kadota Akizo, Niisato Naoyoshi, PATTERN RECOGNITION DEVICE ,(1988)