Regularization of automatic training of Mahalanobis neurons for small samples of examples of the “Own” image

Abstract

The paper considers the problem of regularization of automatic stable training of networks of artificial Mahalanobis neurons that simultaneously perform recognition of the “Own” image.