Methods for recognizing multispectral images based on neural networks. Models for recognizing multispectral images based on neural network technologies

      Аннотация: This monograph presents the basic concepts, methods, and models of multispectral image recognition based on oscillatory and diagonalized neural networks in a logical sequence. In addition, the monograph contains sections on the modified Hecht-Nielsen theorem, optimizing the size of a neural network taking into account the nonlinear generalized error of the neural network model, creating a pseudo-spin neural network with reduced learning time, and creating a decision support model based on a diagonalized neural network.

Ключевые слова: recognizing multispectral images, neural networks, Decision-Making, Model, diagonalized three-layer neural network, matrices of synaptic connections, time of decision making, nonlinear oscillator neuron, Frequency Modulation, morphology of the information signal, Resonance Effect, encoding and decoding of information