Region-Based Segmentation of Dead Cancer Cells in Microscopic Cytological Image–A Reciprocal Study

In this paper the region-based image segmentation methods for segmenting dead cancer cells in microscopic cytological image of Swiss albino mice are presented. We have used three region-based segmentation methods: Thresholding based morphology; Region growing and distance transform based watershed transform. The morphological operators and watershed transform segment the dead cancer cells with modifications in the size and shape of the original cancer cells whereas the region growing method extracts the dead cancer cells with sharp edges and retains the useful information. The segmentation time is very low in watershed transform compared with other two methods.