Segmentation and Classification Of Mammograms For Breast Cancer Diagnosis

Abstract

Breast cancer is still a major health problem, which has led to a lot of research into cutting-edge detection techniques. This research explores the complex process of categorizing and segmenting mammograms in order to improve breast cancer diagnosis. We incorporate a thorough preprocessing workflow into our system, starting with careful image cropping. Using cutting-edge segmentation techniques along with pre-processing techniques and combining it with classification technique to enhance the accuracy of the prediction of cancer detection. Combining these processes will not only help to increase the accuracy but also helps in better segregating the right type of tumors. Our search for the best segmentation model demonstrates our dedication to accuracy and predictability in the detection of cancer affected area. The segmented features from the segmentation model are then combined with the actual mammogram and further fed to the pretrained classification model. The classification categories are categorized in 2–5 Breast Imaging Reporting and Data System (BI-RADS) classification