A Deep Learning Approach to Detect Breast Cancer Disease from Mammogram Images

Abstract

Breast cancer is a significant cause of mortality among women. Breast cancer is regarded as the most severe concern among women. The early detection of breast cancer can potentially impact mortality rates significantly. The implementation of deep learning techniques has the potential to mitigate the death rate associated with breast cancer by the early detection of the disease. This study used several transfer learning models for a specific task and assessed the specificity of each model’s predictions. The dataset was collected from a publicly available resource that contains mammogram images of breasts. We have used several image preprocessing techniques to enhance image quality. The primary objective is to evaluate the precision and usefulness of each algorithm in accurately classifying data, using factors such as efficiency, accuracy, recall, specificity, and F1-score. The MobileNet model showed remarkable performance compared to other models. Our proposed model’s training and validation accuracy were reported as 99.54%, which indicates that this model is a best-fit model