DefinePK

Abstract

Breast cancer remains one of the most prevalent malignancies worldwide and a leading cause of mortality among women. Early detection is critical for improving survival outcomes, yet traditional mammography interpretation faces challenges such as inter-observer variability, dense breast tissues, and high false-positive rates. Recent advances in artificial intelligence (AI) and deep learning have enabled automated detection systems with significant diagnostic potential. This paper proposes TransResCNN, a hybrid framework integrating transfer learning with residual convolutional neural networks (ResNet) to improve automated breast cancer detection in mammographic images. The model leverages pre-trained ResNet weights for enhanced feature extraction, residual connections to mitigate vanishing gradients, and data augmentation strategies to address class imbalance. Evaluated on publicly available histopathology image datasets, TransResCNN achieved 90.76% accuracy and 93.56% F1-score, outperforming baseline CNN models and demonstrating robust generalization. Comparative analysis against state-of-the-art approaches confirms the superiority of the proposed framework. These results highlight the potential of TransResCNN to augment radiologists’ decision-making, reduce diagnostic errors, and support clinical breast cancer screening at scale.