Advanced Deep Learning Approach for Multiclass Breast Cancer Detection from Mammogram Images
Zaid Bin Tariq Baig, Xiaojuan Han, Muhammad Azam, Rehan Ahmad · 2023
Breast cancer continues to be a significant health concern worldwide, the leading cause of cancer-related deaths among women. However, early detection of breast cancer can significantly enhance survival rates and the efficacy of treatment plans. This study harnesses the power of artificial intelligence (AI) to aid in the accurate and prompt breast cancer diagnosis from histopathological images. We implement a two-stage approach involving image preprocessing using MediCorrectNet, an innovative deep-learning model designed to enhance the diagnostic quality of medical images, and a Convolutional Neural Network (CNN) for classifying the preprocessed images. Specifically, we utilize the EfficientNetB0 model, a pre-trained network known for its effectiveness in image classification tasks. With the combination of MediCorrectNet and EfficientNetB0, our approach achieved an impressive accuracy of 93%, precision of 92%, recall of 91%, F1-Score of 0.915, and an AUC-ROC of 0.98. These results demonstrate our proposed approach's feasibility and high potential for effective breast cancer diagnosis, setting a promising direction for future research and real-world applications.