Synthesis of optimal differentiators for the locally oriented texture detection algorithm

A method is proposed for constructing noise-resistant mask differentiating filters minimizing error probability in locally oriented texture detection algorithm based on gradient structure tensor. A gaussian approximation of the components joint distribution of the gradient structure tensor is proposed to solve the problem of synthesizing. It is shown that a decrease in the systematic error leads to a considerable increase in the accuracy of directional field estimation. It is shown that the use of optimal mask differentiating filter greatly reduces the probability of error detection.