A Novel Rough Set Based Image Denoising Algorithm

The primary focus of the paper is towards image enhancement via removal of salt and pepper noise from images. In this paper, a novel statistical approach based on the properties of rough set theory is proposed, where noisy pixel identification and removal are controlled by decision rough parameters. Each enhancement decision is directly governed by four parameters – the pixel is noisy or not, the pixel has any non-noisy neighbor compatible enough to replace it, the deviation of the neighboring pixel from the central pixel value, and matching of the threshold criterion. The four phase decision making algorithm fetches highly accurate results and with consecutive iterations and upgradation, the algorithm is able to remove all noisy pixels while maintaining fine details of the image for even 95% corruption levels.