Improved Butterfly Optimization Algorithm for Automated Breast Cancer Detection and Classification using Deep Learning

Breast cancer is a worldwide health issue affecting women, and research is primarily done in medical images for early diagnosis and detection. The problem statement is to train model strategies during pre-processing with the absence of identification and specific consideration for development, leading to a biased model. The MIAS dataset is used in mammography scans with the three classes of malignant, normal, and benign. The proposed method goes with the feature extraction process with the vector, the Improved Butterfly Optimization algorithm (IBOA) for feature selection, deep adaptive spatial-based feature fusion, and finally the feature vector selection process. The numerical validation of the proposed method results in an accuracy of 99.98%, a sensitivity of 98.79%, a specificity of 99.50%, and an F1-score of 98.90%. Comparing the existing methods, like Improved Multi-fractal Dimension (M-FD), Improved Marine Predator Algorithm (IMPA-ResNet50), and Infinite Genetic Algorithm (IFSGA-DNN), can resolve the problems overcome by the proposed method.