Transfer and Ensemble Approach for Breast Cancer Detection and Classification Using Deep Learning

Abstract

Breast cancer is a serious disease that can cause significant health problems for women worldwide. It is crucial to detect and classify breast cancer early stage so that doctors can promptly treat it and aid patients in their recovery. Many investigators have used various deep learning (DL) strategies to detect and classify breast cancer. However, due to the complexity of the problem, relying on a single DL model may not suffice to achieve high accuracy. Therefore, this study suggests a transfer and ensemble deep model for breast cancer detection and classification. The suggested model involves using pre-trained models such as Sequential, Xception, DenseNet201, VGG16, and InceptionResNetV2. The top three models are selected to collaborate and deliver the most accurate results. The proposed DL model was tested on publicly available breast BUSI datasets, demonstrating its superiority over single DL models, achieving an accuracy of 87.9% on the BUSI dataset. Additionally, the model proved to be adapTABLE to different amounts of data, making it potentially valuable in hospitals and clinics.