COBRC-MOS: Crayfish Optimization Algorithm for Breast Cancer Classification using Mammography Images

Breast cancer (BC) is one of the most common and deadly diseases necessitating early identification and effective diagnosis to increase survival chances. However, traditional diagnostic approaches suffer from high false-positive and false-negative rates and limited accessibility in resource-constrained settings. To overcome these challenges, a novel deep learning-based COBRC-MOS is proposed for the classification of BC using mammographic images. Initially, the input mammography images are gathered from the MIAS dataset. The input images are preprocessed using contrast-limited Adaptive Histogram Equalization (CLAHE). The EfficientNet is used for feature extraction, and the Crayfish Optimization Algorithm (CFO) is used to select the most discriminative and optimal features by efficiently exploring the search space to enhance model performance and interpretability. Then the SVM is used to classify eight types of breast cancer namely, malignant papillary carcinoma (MPC), malignant mucinous carcinoma (MMC), Malignant Ductal Carcinoma (MDC), malignant lobular carcinoma (MLC), Benign Tubular Adenoma (BTA), Benign Phyllodes Tumor (BPT), Benign Fibroadenoma (BF), Benign Adenosis (BA). Finally, the U-Net segmentation is used to identify affected regions accurately. Moreover, the proposed COBRC-MOS model classification accuracy was 98.35%. The proposed model improved total accuracy by 2.99%, 5.79%, and 5.37% better than BC 2 NetRF, BC 2 -MultiRad, and BCD-XGBoost, respectively.