BC-SwinNet: Swin transformer and CNN with multi-objective optimization for multi-class breast cancer detection using Histopathological images.

Breast cancer remains a major global health concern and is the leading cause of cancer-related deaths among women worldwide. It is estimated that by 2030, its incidence and mortality rates will rise due to population growth, aging, and lifestyle changes. Although histopathological examination is considered the gold standard for diagnosis, manual assessment is often time-consuming, subjective, and-particularly in the case of complex, multidimensional cancer classifications-prone to inter-observer variability. These limitations underscore the need for accurate, automated, and computationally efficient diagnostic systems to aid in clinical decision-making. To address these challenges, this study introduces BC-SwinNet-a hybrid deep learning framework that integrates the Conditional Swin Transformer (ConSwinTras), the Multi-Objective Elk Herd Optimization (MEHO) algorithm, and a Layered Attention-based Convolutional Neural Network (CA-CNN) to classify multiple breast cancer subtypes using histopathological images. ConSwinTras extracts hierarchical and contextual representations from tissue images, while MEHO is utilized to select subsets of high-resolution features, thereby reducing dimensionality and enhancing prediction performance. The CA-CNN module focuses on diagnostically relevant regions through layer-wise attention algorithms and performs the final sample analysis and classification. The proposed BC-SwinNet model was evaluated on two benchmark datasets-BreakHis and BACH-achieving classification accuracies of 99.91% and 99.854%, respectively. Experimental results demonstrate that the proposed framework outperforms numerous existing methods in terms of classification performance while maintaining computational efficiency. These findings suggest that BC-SwinNet offers a robust and efficient approach for automated breast cancer diagnosis and holds the potential to enhance diagnostic support systems in clinical practice.