Applying the Neuronetchic Methodology to Text Images for Their Recognition

There is considered the methodology for recognizing text images based on neural networks, methods and algorithms for building a neuro-fuzzy system for recognizing text images, in particular methods for improving the quality of text images and reducing noise through linear and nonlinear filtration. Features of binarization of such images, fuzzy processing of images to allocate boundaries and segmentation of symbols, and the ability to implement grammar for the structural recognition of text images is shown. The simulation of the developed system is also carried out.