Cancer in breast has emerged as the leading cause of death among women worldwide. Early detection of breast cancer is very important. CAD (computerassisted diagnostics) has emerged as a useful medical diagnostic tool that is very helpful in the protection of patients by decresing of false positive outcomes and allowing for rapid diagnosis. Rapid advances in the development of high-resolution imaging techniques have helped the computer to detect automatic breast cancer. The rapid development of in-depth learning, a family of machine learning techniques, has stimulated a great deal of interest in its application to the problems of medical illustrations. Here, we develop an in-depth study algorithm that can accurately detect breast cancer in mammograms tests using a “end-to-end” training method that utilizes training data sets with complete clinical definitions or only cancer status (label) of the whole image. The proposed Computer-Aided Diagnosis (CAD) program consists of four parts: mammograms reconstruction, extraction of characteristic using deep precision network, mass detection, and finally mass fragmentation using Fully Connected Neural Networks (FC-NNs).

Keywords : Deep Learning, Image Processing, Mammogram