AI-Driven Breast Cancer Detection: An Empirical Analysis of Deep Learning and Machine Learning Techniques

Breast cancer is regarded as one of the most prevalent and perilous diseases affecting women globally. A fast and accurate diagnosis is crucial for improving patient survival. Mammography, biopsy, and pathologist interpretation are essential for breast cancer diagnosis; nevertheless, they are fraught with deficiencies such as human error, exorbitant costs, and limited accessibility. This research presents deep learning and machine learning techniques for automated breast cancer screening to overcome these challenges. Models including CNN, ResNet50, U-Net, Random Forest, and SVM were evaluated using critical datasets such as WDBC, BreakHis, and BCDR-D01. The empirical analysis demonstrates that the deep learning architecture ResNet50 outperformed others, achieving an accuracy of 0.97, an F1 score of $\mathbf{0. 9 6}$, and an AUC of $\mathbf{0. 9 8}$.