A Nuclei Segmentation Method Based on Optimal Fuzzy Clustering Using Salp Swarm Algorithm for Histopathological Images

Abstract

Automatic segmentation of nuclei in the H&E stained histopathological images is an open-ended research problem. In this paper, a new salp swarm algorithm based fuzzy clustering method is proposed which is used for nuclei segmentation in the histopathological images. The salp swarm algorithm finds the optimal clusters by the objective function defined over intra-cluster distances or compactness. The performance of the proposed segmentation method is evaluated in terms of F1 score and aggregated jaccard index on the histopathological image dataset of TNBC patients. The experimental results depict the efficacy of the proposed method over the other considered clustering-based segmentation methods, namely K-means and Fuzzy C-means.