DAX-Net: A dual-branch dual-task adaptive cross-weight feature fusion network for robust multi-class cancer classification in pathology images

Doanh C. Bui, Boram Song, Kyungeun Kim, Jin Tae Kwak

Abstract: Highlights•An effective dual-branch hybrid approach for combining a CNN-based model and a Transformer-based model.•A cross-weight feature fusion mechanism increases the accuracy and robustness of cancer classification.•Dual-task learning with tailored loss functions improves the learning capability of the proposed model.