Comparative Analysis of Local Adaptive LMS Image Filtration

In this paper the 2D part of the developed method for local adaptive 3D filtration of multidimensional grayscale images is compared with some advanced methods for image filtrations - BM3D, Average, Gaussian and etc. The adaptation of the presented algorithm is based on the information from a small local area of the neighbouring pixels of the processed images using the developed 2D Adaptive LMS filter. A comparative analysis of the quality of the processed images was made in terms of both peak signal-to-noise ratio calculations and subjective visual quality. The experimental results were carried out with a set of test halftone images that were noised with Additive White Gaussian Noise (AWGN) with different variance levels.