Parametric comparison for breast cancer detection using machine learning techniques for mammogram imaging
V. Pathak, S.S. Kesavaraju, S. Sinha, R. Chatterjee · A.A. Balkema eBooks · 2025
This research compares three Convolutional Neural Network (CNN) classifiers against an Artificial Neural Network (ANN) and Bayesian Neural Network (BNN) with Support Vector Machine (SVM) for breast cancer detection using mammogram images from the mini-MIAS database. After preprocessing steps like noise removal and segmentation, the CNN achieved a sensitivity of 82.68% through four hidden layers. The ANN utilized adaptive mean filters, CLAHE, and SVM classification after BNN feature extraction, reaching a sensitivity of 98%. The study confirms that the ANN-SVM hybrid model outperforms others in sensitivity and accuracy. This highlights the potential of hybrid models in enhancing automated diagnostic systems, suggesting a direction for future research in improving detection accuracies.