Convolutional Neural Networks based Breast Cancer Detection Using Feature Fusion

Currently, one of the most significant health issues affecting women is breast cancer. Breast cancer, which has the highest mortality and morbidity rates among diseases that affect women, poses a severe threat to their lives and health. It is crucial to diagnose breast cancer early. Recently, the technological and theoretical developments in innovative techniques such as machine learning made it possible to achieve early diagnoses of breast cancer. Using Convolutional Neural Networks (CNNs) and Support Vector Machines (SVMs) for breast cancer detection, we developed a Computer-Aided Diagnosis (CAD) system to identify breast tumors from mammogram images. The proposed system is composed of four stages. Firstly, CNN is used to classify mammogram images. Secondly, CNN-based features are extracted and used with a standard classifier, which is SVM, to identify potential tumors. Thirdly, SVM is used to distinguish between different types of tumors based on the extracted features. Finally, a data fusion method is employed to combine the results obtained from the first, second, and third methods to improve the overall performance of the system. The individual techniques, accuracy is increased, and it reaches around 99% by using fusion methods. Keywords: Breast Cancer, Convolutional Neural Network, Computer-Aided Diagnosis, Support Vector Machine, Fusion Methods