Optimized Kernel Fuzzy Clustering for Breast Lesion Segmentation in Mammograms

Breast Cancer is one of the fatal diseases that is caused by the unnatural development of the tissues in the breasts, which are abnormal. The level of sarcoma or the stage of the cancer is mostly determined by the doctor’s analysis. In order to provide a technical contribution that supports the doctor in making a decision, this paper is intended to develop a framework that can help in determining the stage of the cancer at present. The major issues in the prediction of breast cancer through mammograms are the diverged artifacts, similar breast tissues, and lower contrast on the boundary between skin and air. To overcome these issues, the Optimized Kernel Fuzzy Clustering Algorithm (OKFCA) is used to determine the cancer portions in mammogram images. The OKFCA algorithm has been described to identify the segmented regions in the Mammogram Image Analysis Society (MIAS) database. The proposed segmentation algorithm is carried out with pre-processed mammogram images, a noise-free image that was obtained by using the Hybrid Denoising Filter (HDF) algorithm, and the proposed OKFCA is a significant approach to finding the cancer segment of the mammogram image. Data clustering facilitates placing data of similar types in one group and of dissimilar types in a different group. The results from the experiments, which were carried out on the MIAS data, confirm the efficiency of the proposed system in terms of accuracy when compared to that of the famous K-Means, OKFCA, and Otsu methods.