Construction of complex images by an adaptive neuronlike research system

We describe the architecture of an adaptive recognition system that permits one to record, describe, and identify various information signals. The proposed architecture of the recognition systems performs automatic encoding and image recovery operations in accordance with the established sets of models for the behavior of the environment, images, and events. This makes it possible to tune the system to the feature of the image being examined or to the problem being solved using such an identification system. Examples of operation of the basic units of the system are presented.