Breast Tumor Segmentation Using K-Means Clustering and Cuckoo Search Optimization

Abstract

Today, there are various methods for detecting tumors in breasts. But researchers are still trying to find an exact automatic way to segment the tumors from breast images. In this paper we propose a clustering-based algorithm for automatic tumor segmentation in the MRI samples. In the proposed method, we use k-means clustering algorithm for segmentation and also we use cuckoo search optimization (CSO) algorithm to initialize centroids in the k-means algorithm. We have used RIDER breast dataset to evaluate the proposed method and results clearly show that our algorithm outperforms similar methods such as simple k-means clustering algorithm and Fuzzy C-Means (FCM).