Image transformation using a feature map of multiply descent cost competitive learning

Summary form only given. It was shown that the feature map obtained by multiple descent cost competitive self-organization can be used for the transformation of images combined with the supervision of an outside intelligence. The example of the change of emotional expression of a face was considered. The task of the first module was to locate the prospective edges. Basically, two orthogonally oriented preliminary network (horizontal and vertical) are sufficient. This is true because a homogeneous region can be outlined using only vertical and horizontal edges. A prospective edge selection (preliminary) network consists of two types of neurons: image neurons and edge neurons. Each image neuron corresponds to a pixel in the image. Between each image neuron is an edge neuron, one corresponding to every possible edge location for the desired orientation. The goal of the vertical network is to find the prospective vertical edges in the horizontal direction