Wavelet-odd prolate spheroidal wave functions in the problem of two-dimensional image segmentation

Аннотация

A wavelet in the form of the first odd prolate spheroidal wave function is proposed for the wavelet transform of a non-uniform 2D image and the formation of clusters of wavelet coefficients in it. Methods for calculating the wavelet function, clustering the field of wavelet coefficients, and constructing their corresponding optimal rectangular windows in an image are described. Modeling has shown the high efficiency of the methods and the algorithm implementing them under various real operating conditions of the optoelectronic device..