A Collection of Contemporary Research Articles in Electronics, Communication & Computation- Book ISBN- 978-81-948050-7-6

Authors: Rajeshwari S. Patil, Ambaji S. Jadhav, Nagashettappa Biradar

Abstract: Breast cancer is a major disease, which is usually seen in women. Early researches have shown that early detection and suitable treatment might increase the life span. Those researches have also proven that detection of small lesions at an early stage improves prognosis, which leads to a decrease in the mortality rate. Mammography is the best approach used for screening the disease. The present paper plans to introduce an automatic breast cancer detection approach using four phases as "pre-processing, segmentation, feature extraction, and classification." Here, the median filtering approach is used for eliminating the noise present in the mammogram image. Later, the segmentation of the tumor is done by the optimized region growing approach, which is the advanced version of the traditional region growing algorithm. Furthermore, the features like “Grey Level Co-occurrence Matrix (GLCM)†and Gray-Level “Run-Length Matrix (GRLM)†are extracted from the segmented tumor during feature extraction. Once the feature extraction is done, the features are subjected to a classifier named Fuzzy logic classifier. The threshold of the region growing algorithm and the membership function of the fuzzy classifier is optimally tuned with the help of the Crow Search Algorithm (CSA) named as “Improved Awareness Probability-based CSA†(IAP-CSA). The analysis shows that the proposed IAP-CSA is acquiring the best results in breast cancer detection and classifying the normal, benign, and malignant images.

Keywords: Mammogram Image, Breast Cancer Detection, Optimized Region Growing Algorithm, Optimized Fuzzy Classifier, Improved Awareness Probability-based Crow Search Algorithm (IAP-CSA).