Анализ алгоритма самоорганизации нейронной сети типа «неокогнитрон» при распознавании образов
Федяев О.И., Труханов Т.И.
This paper studies the principles of setting a multilayer neural network neocognitron to recognize different patterns. For this purpose, a software model of neocognitron with self-organization learning algorithm was developed. Neocognitron with competitive learning algorithm acquires the ability to correctly recognize patterns at various distortions in their form.