Sharypanov A. Structural method of image and signal processing using variable resolution

The Ph.D. thesis presents new solution for actual scientific task of halftone image segmentation. For the first time on the basis of existing mathematical description of changes in sizes of the excitation zones of visual neuron's receptive field the method for finding discontinuities of a piecewise smooth function using discrete implementation of this function with variable resolution was presented. For the first time the new algorithms of image and signal processing that used variable resolution were proposed and experimentally verified for noisy examples. No a priori information about noise was used. The decision was made taking into account all segmentation answers for all resolutions that were used for processing. The sequence of resolutions that was used for answer generation was detected automatically. The information technology of signal processing with variable resolution was developed. The practical value of obtained results was confirmed by acts of implementation of created information technology from virtual organization Medgrid and Cardiolyse Ltd.