Cancer is a significant global health concern, claiming a high number of lives annually and causing death rates to increase globally. The most common type of cancer worldwide is breast cancer, followed by lung, bronchial, and prostate cancer. Early detection and identification of cancer are crucial steps toward improving medical diagnosis and increasing survival rates. Imaging plays a vital role in detecting breast cancer and determining the stage of the disease, thereby enabling prompt biopsies and appropriate treatment decisions. The primary imaging technique for detecting breast cancer is mammography, although other techniques such as ultrasound, tomosynthesis (DBT), and contrastenhanced mammography (CEM) may also be used. In this study, the focus will be on the development and implementation of deep learning models to classify mammogram images and to compare and evaluate their performance. Furthermore, the study will include a second part that involves the segmentation of ultrasound mammogram images, testing the models, and determining the best among them. The model's performance will then be evaluated through the classification of predicted masks using a convolutional neural network designed specifically for this purpose. The study aims to develop and implement deep learning models to detect and classify mammogram images for the early diagnosis of breast cancer, thereby improving medical diagnosis and increasing survival rates. Additionally, the study will evaluate the performance of the models in segmenting ultrasound mammogram images, which may aid in the development of more accurate and effective diagnostic tools