Mammography scans for breast cancer detection using CNN
S. Anandhi, Sonali Jagadish Mahure, Thupakula Yogesh Royal, V Nikhil, Uppalapati Viswanadh · 2024
This study suggests the collection of studies with state-of-the-art technologies to improve breast cancer screening techniques. With appreciable potential for improvements, one study introduces a non-invasive thermography screening method with Convolutional Neural Networks (CNNs) to classify breast graphical records into normal and pathological categories. Another presents an ensemble-based machine learning system with 98.83% accuracy in the automatic diagnosis and prognostication of mammary cancer, highlighting importance of early observation in research and medical settings.Translation invariance are achieved by a fresh approach to deep NN design that combines a group CNN with a special Euclidean motion group and Discrete Cosine Transform (DCT). This network architecture exhibits high computational performance and inference generalization with a noteworthy 94.84% accuracy on mammography images. Finally, a “two-view classifier” for mammary cancer diagnosis is introduced using a deep learning approach. It makes use of Efficient Net and achieves an AUC of 0.9344, displaying excellent performance and offering a useful tool for scientific community.