DR-Net: dual-rotation network with feature map enhancement for medical image segmentation

Abstract To obtain more semantic information with small samples for medical image segmentation, this paper proposes a simple and efficient dual-rotation network (DR-Net) that strengthens the quality of both local global feature maps. The key steps DR-Net algorithm are as follows (as shown in Fig. 1). First, number channels each layer is divided into four equal portions. Then, different rotation strategies used to map multiple directions subimage. multiscale volume product dilated convolution learn features Finally, residual strategy integration fuse generated Experimental results demonstrate method can higher segmentation accuracy on CHAOS BraTS data sets compared state-of-the-art methods.