Rethinking U-Net from an Attention Perspective with Transformers for Osteosarcoma MRI Image Segmentation

Figure 1

Illustration of the UATransNet medical image segmentation framework for osteosarcoma MRI. Among them, UATransNet successively performs dataset classification optimization and removes MRI image irrelevant background by averaging teacher models and normalized preprocessing. Then, UATransNet is designed with transformer self-attention component (TSAC) and global context aggregation component (GCAC) at the bottom of the encoder-decoder architecture to perform integration of local features and global dependencies and aggregation of contexts to learned features.