Breast Cancer CAD: African Buffalo Optimization, Clustering and XGBoost for Histopathological Breast Cancer Detection

A hybrid computer-aided diagnosis system for breast cancer histopathology. The pipeline combines Macenko stain normalisation and tiling, 118 handcrafted descriptors (GLCM, LBP, Gabor, morphological and first-order intensity), African Buffalo Optimization for swarm-intelligence feature selection, K-means and Fuzzy C-Means clustering as an unsupervised cross-check, XGBoost classification with patient-grouped splitting and Youden-J threshold selection, probability calibration, SHAP explainability, and a Django REST deployment layer with role-based access control. Developed and validated on a Nigerian breast histopathology dataset, with BreakHis used for external validation. If you use this software, please cite it as below.