Image Diagnosis of Breast Masses Based on Deep Learning

Abstract

With the change of lifestyle, the incidence rate of breast diseases is also increasing. In some cases, there will also be malignant diseases, that is, breast cancer. It is a malignant tumor occurring in the epithelial tissue of the breast, with a very high morbidity and mortality. At present, there is no medical means to completely eradicate the disease. Therefore, early screening plays a very important role in preventing the disease. There are many types of breast diseases, such as malignant tumors and benign tumors. Different types of lesions have different characteristics, which increases the difficulty and workload of doctors in diagnosis, thereby increasing the diagnostic time and cost of individual patients. CT imaging diagnosis accounts for a large proportion in breast disease screening. Therefore, this article proposes a deep learning model to classify breast CT images based on how to help doctors reduce workload and improve diagnostic accuracy. The model can label the CT images as normal breast, malignant breast tumors, and benign breast tumors based on the characteristics of the patient's breast CT images, in order to achieve early diagnosis of breast diseases and reduce the workload of doctors, Improve the recovery rate of patients. After completing the model construction, this article trained and evaluated the model to verify its effectiveness.