Нейроподібні методи ущільнення зображень

Abstract

In scientific literature different approaches are examined to application of networks of neurons for the compression of images, however deserved the special attention is by approaches which lean against principles of vector quantization, as they provide the high speed of compression at saving of good quality of the picked up thread image. Ideal for the decision of these tasks are networks of neurons, that selforganized, offered the Finnish scientist T. Kohonen, namely, networks which self-organized, as a two dimensional Kohonen map. The devoted is given questions to consideration of these work.