Abstract:

The technologies for detecting and classifying breast cancer (CAD) have improved, however there are several problems and restrictions that still need to be looked into further. The development of breast cancer CAD systems was significantly impacted by the considerable advancements in machine learning and image processing techniques over the past 10 years, particularly with the advent of deep learning models. In addition to the traditional machine learning-based approaches, this study offers the current deep learning-based CAD system to identify and categorise masses in mammography in an organised manner. The survey offers the most modern approaches and the most popular assessment measures for the breast cancer CAD systems, as well as the publicly available mammographic datasets currently in use. The research highlighted the benefits and drawbacks of the present body of literature while providing a discussion of it. The survey also sheds insight on the difficulties and limits of the existing methods for identifying and categorising breast cancer.