Detection of Breast Cancer in Mammography Using Deep Learning Models

Due to the growing need for specialized care in communities that do not have a local radiology service, and the long response times of diagnostic departments, there is a need to provide technologies that help medical diagnosis. In this project, we implement convolutional neural network models for the identification of mammography’s with data suggestive of breast cancer, considering the BI-RADS scale system in a binary classification based on its scales. We considered BI-RADS scales from 1–3 as benign and from 4–5 as malign, binary interpretation for both models (VGG16 and ResNet50) obtained results of 0.87 in accuracy.