Fig. 1From: Multi-modality self-attention aware deep network for 3D biomedical segmentationComparisons of encoder and decoder structuresBack to article page