Enhancing Breast Cancer Classification in Mammography Images Using Multi-View Deep Convolutional Neural Networks

Early detection is crucial since breast cancer kills most women. Mammography imaging is used for early detection, although the data are hard to interpret. Deep learning systems excel at breast cancer classification. Most current studies employ one mammography image, which may omit significant features. This study presents a multi-view deep CNN-based breast cancer classification algorithm using mammography images. This approach uses information from the craniocaudal (CC) and mediolateral oblique (MLO) views. To improve breast cancer classification accuracy and robustness, the proposed model considers all these parameters. Researchers build and test the multi-view deep CNN model using mammography images from malignant and non-cancerous patients. This dataset's ground truth labels allow robust performance assessment. The multi-view deep CNN architecture learns unique features from each view, integrating complementing information and improving classification performance. Extensive trials compare the proposed multi-view deep CNN methodology to various state-of-the-art technologies. It outperforms single-view CNN models and feature-based classifiers in breast cancer classification. Multi-view information improves accuracy, sensitivity, and specificity, making breast cancer detection and discrimination more dependable. The proposed model has improved Accuracy 0.92, Sensitivity 0.89, Specificity 0.94, and Precision 0.91. This research uses Multi-View Deep Convolutional Neural Networks to improve mammography breast cancer classification. The model outperforms single-view and feature-based classifiers by adding craniocaudal (CC) and mediolateral oblique (MLO) viewpoints.