SD LMS L-Filters for Filtration of Gray Level Images in Timespatial Domain Based on GLCM Features
Robert Hudec, Miroslav Benco, Janko Krajcovic
DOI:
Abstract
In this paper, the new kind of adaptive signal-dependent LMS L-filter for suppression of a mixed noise in greyscale images is developed. It is based on the texture parameter measurement as modification of spatial impulse detector structure. Moreover, the one of GLCM (Gray Level Co-occurrence Matrix) features, namely, the contrast or inertia adjusted by threshold as switch between partial filters is utilised. Finally, at the positions of partial filters the adaptive LMS versions of L-filters are chosen.