Rib Fracture Detection with Dual-Attention Enhanced U-Net

Figure 3

Our proposed CFSG U-Net. The model uses a U-Net structure consisting of four encoder blocks (E1-E4) and three decoder blocks (D1-D3). The number next to each block indicates the input size of each block. In the decoder block, denotes the feature map from the encoder path, and represents the feature map from the decoder path.