K. M. Singh

Improved rank conditioned median filter for removal of impulse noise from images

This paper presents an improved rank-conditioned median filter, with the ability of the median filter to remove the impulse noise and the rank conditioned median filter to preserve edges and fine image details. The center pixel that lies outside the trimming range after ranking in the rank-conditioned median algorithm may not be the corrupted pixel, but it may be a pixel on the edge or image detail and therefore, it is further verified if the central sample is a possible candidate for being an impulse by thresholding. The better performance of the new filter is demonstrated.

Features preserving filters using fuzzy Kohonen clustering network in detection of impulse noise

It is always advisable to apply filtering only on corrupted pixels of images leaving untouched the uncorrupted ones to preserve image features and to avoid blurring effects. We present an algorithm which can detect the corrupted pixels in texture images. It uses a fuzzy Kohonen clustering network that integrates with the fuzzy c-means (FCM) model utilizing the updating strategies of the first and the learning rate of the second.

Multi-stage detection based filtering for removal of impulse noise from images

This paper presents a new technique for removal of impulse noise from images. By incorporating the rank-conditioned median filtering, a threshold operation on the deviation of the extreme ranks from the healthy pixel-values in the filtering window and the center-weighted median (CWM) based adaptive impulse detection mechanism into the detection framework. An improved impulse detection scheme is developed. The proposed detection scheme followed by rank-ordered mean (ROM) filtering of corrupted pixels effectively reduces impulse noise while preserving the image details. Simulations show that the proposed scheme works better than the existing schemes in suppressing both fixed-valued and random-valued impulses at different noise ratios.