A Novel Deep Learning Approach for Breast Cancer Detection on Screening Mammography
Carson Kai-Sang Leung, Hoang Hai Nguyen · 2023
Breast cancer is one of the most common cancers in women. Currently, mammography is one of the best methods for early detection of breast cancer. The rapid development of deep learning has attracted much attention from researchers in an effort to develop automatic support tools for breast cancer diagnosis. There are many published works on deep learning models supporting the diagnosis of breast cancer using mammography. However, most—if not all—deep learning models perform well on certain or similar datasets but fail with unknown datasets that come with different distribution. Thus, there is a critical need of deterministic models, which can work for any datasets. In this paper, we propose a novel approach to develop deterministic models, in which we emphasize the role of data mining as a prerequisite step that can help model detect and locate region of interest (ROIs) accurately. Furthermore, our approach can perform mammogram image segmentation and classification in parallel.