Existing studies of multi-modality medical image segmentation tend to aggregate all modalities without discrimination and employ multiple symmetric encoders or decoders for feature extraction and ...