Mediated morphological filters

Abstract

This paper presents new morphological operators based on a special combination of median filtering and classical gray-scale morphological operators. The newly introduced operators demonstrate a superb performance compared with classical ones and even weighted morphological operators (Sedaaghi and Wu (1998)) introduced by the author, for signals/images buried in speckle, salt and pepper, and Gaussian noise. Their efficiency is comparable with convolutional filtering for speckle and Gaussian noise removal, where classical morphological operators fail to be applicable. The proposed algorithms have been applied for off-line biomedical signal/image processing successfully.