MS-TCNet: An effective Transformer-CNN combined network using multi-scale feature learning for 3D medical image segmentation

Yu Ao, Weili Shi, Bai Ji, Yu Miao, Wei He, Zhengang Jiang

Abstract: Highlights•Propose a 3D Transformer–CNN combined network with multi-scale feature learning.•Introduce a shunted Transformer into the encoder to capture 3D multi-scale features.•Utilize a CNN-based pyramid decoder to decode refined features from multiple scales.•Propose a lightweight module to adaptively fuse features at different scales.•Demonstrating the effectiveness and superiority of the model on three challenging tasks.