An edge preserving noise smoothing technique using multiscale morphology

This paper presents a method for improving the quality of gray-level images by reducing the effect of noise using multiscale morphology, The underlying concept of the work is to assign progressively less importance to features of smaller scales as their possibilities of being noise particles are more. Features at various scales are extracted by means of morphological filtering. The proposed scheme is first illustrated in one dimension. Morphological towers are built to implement the method in two dimensions. The proposed algorithm has been tested on a set of real images corrupted with different types of noise. The results are compared with those of other standard noise removal algorithms based on some standard performance measures. A modification of the method considering noise statistics along with its results are also presented in this paper. (C) 2002 Elsevier Science B.V. All rights reserved.