Details

Breast cancer is the leading cause of death among women, with the American Cancer Society reporting over 40,000 deaths each year. Given the critical importance of early detection in improving outcomes, advancements in diagnostic technologies are crucial. This thesis presents the development of the ReAttentionNet model, an enhancement of the ResNet50 architecture integrated with a Convolutional Block Attention Module (CBAM). This model is specifically designed for classifying histopathological images as either malignant or benign at a 40x magnification level. Demonstrating outstanding performance metrics—99.25% accuracy, 100% precision, 98.92% recall, and a 99.56% F1-score—the ReAttentionNet model significantly outperforms existing methods, highlighting its potential as a powerful tool for early breast cancer detection. Further validation of our model on the IDC dataset, utilizing the foundational BreakHis dataset, confirms its adaptability and potential for real-world clinical settings, setting the stage for transformative advances in early detection practices for breast cancer.