Modified watershed transform algorithm for cancer cell segmentation counting

Abstract

In cancer research, the migration-invasion assay is one of the most important experiments used to evaluate the spreading potential of cancer cells from primary site of tumor to other organs. The process needs to read the number of cells spread to target organs on the pictures taken under microscopic in order to evaluate the performance of clinical trails on various cancer treatments. The number is usually counted, manually, under microscope or using Image J software as an assisting tool. However, the cells are difficult to identify since they are usually overlapped, hence, the process requires experienced researchers. Moreover, the process is time consuming and subjective, which could lead to an inaccurate results.

      In this paper, an alternative approach to segment and count the number of cells is presented. The proposed algorithm is based on digital image processing technique, specifically, Watershed transform. The studied pictures are from migration-invasion assay bile duct cancer cell line. The results are compared to the numbers reported by three experienced researchers. The results show that the proposed method can achieve 97.01% average on accuracy. The average time of processing per picture is 4.28 minute.