Watershed and Clustering Based Segmentation of Chromosome Images

Abstract

Chromosome images are captured by microscopicimaging and hence analysis of these chromosomes is very difficulttask. To correctly analyse the chromosomes, segmentation ofchromosome images is necessary. Segmentation is necessary foridentification genetic diseases. For accurate diagnosis, M-FISHlabelling technique is used. This technique helps in correctlyidentifying the genetic diseases. This technique uses 5 colorthat labels each of chromosomes and DNA stain. In thisproject, chromosome images are segmented using fuzzy cmeans algorithm and by using watershed algorithm. Thesegmentation results of both the algorithms are compared byfinding segmentation accuracy. The algorithms were tested onM-FISH database which is available online.