Switching-based clustering algorithms for segmentation of low-level salt-and-pepper noise-corrupted images

This paper presents new clustering-based segmentation algorithms. The proposed switching-based clustering algorithms can minimize salt-and-pepper noise during segmentation without degrading the images' fine details. The proposed algorithms incorporate the salt-and-pepper noise detection stage into the clustering algorithm, producing an adaptive technique specifically for segmentation of noisy images. Experimental results show that the proposed switching-based clustering algorithms produce better segmentation with fewer noise effects than conventional clustering algorithms. Quantitative and qualitative analyses show positive results for the proposed switching-based clustering algorithms, which consistently outperform conventional clustering algorithms in segmenting up to 50 % of salt-and-pepper noise density. Thus, these switching-based clustering algorithms can be used as pre- or post-processing task (i.e., segmenting images into regions of interest) in electronic products such as televisions and monitors.