RSU-Net: U-net based on residual and self-attention mechanism in the segmentation of cardiac magnetic resonance images

• Published in: Computer methods and programs in biomedicine 2023-04, Vol.231, p.107437-107437, Article 107437
• Main Authors: Li, Yuan-Zhe, Wang, Yi, Huang, Yin-Hui, Xiang, Ping, Liu, Wen-Xi, Lai, Qing-Quan, Gao, Yi-Yuan, Xu, Mao-Sheng, Guo, Yi-Fan
• Format: Article
• Language: English
• Subjects: Anisotropy; Benchmarking; Cardiac MRI; Deep learning; Entropy; Heart; Humans; Image Processing, Computer-Assisted; Image segmentation; Magnetic Resonance Imaging; Residual; Self-attention; U-net
• ISSN: 0169-2607; 1872-7565
• DOI: 10.1016/j.cmpb.2023.107437

•Two skip connections are introduced inside the Conv_Connet module, which combines the original input of the module with the output after feature extraction to enhance the transfer of image feature information.•A self-attention mechanism is introduced into the self-attention module at the bottom to integrate the input with global information.•Combine the advantages of CrossEntropyLoss and Dice loss function, improve the loss function of the model, and use the fusion loss function to improve the training efficiency and segmentation accuracy of the network. Automated segmentation techniques for cardiac magnetic resonance imaging (MRI) are beneficial for evaluating cardiac functional parameters in clinical diagnosis. However, due to the characteristics of unclear image boundaries and anisotropic resolution anisotropy produced by cardiac magnetic resonance imaging technology, most of the existing methods still have the problems of intra-class uncertainty and inter-class uncertainty. However, due to the irregularity of the anatomical shape of the heart and the inhomogeneity of tissue density, the boundaries of its anatomical structures become uncertain and discontinuous. Therefore, fast and accurate segmentation of cardiac tissue remains a challenging problem in medical image processing. We collected cardiac MRI data from 195 patients as training set and 35patients from different medical centers as external validation set. Our research proposed a U-net network architecture with residual connections and a self-attentive mechanism (Residual Self-Attention U-net, RSU-Net). The network relies on the classic U-net network, adopts the U-shaped symmetric architecture of the encoding and decoding mode, improves the convolution module in the network, introduces skip connections, and improves the network's capacity for feature extraction. Then for solving locality defects of ordinary convolutional networks. To achieve a global receptive field, a self-attention mechanism is introduced at the bottom of the model. The loss function employs a combination of Cross Entropy Loss and Dice Loss to jointly guide network training, resulting in more stable network training. In our study, we employ the Hausdorff distance (HD) and the Dice similarity coefficient (DSC) as metrics for assessing segmentation outcomes. Comparsion was made with the segmentation frameworks of other papers, and the comparison results prove that our RSU-Net network performs better and can make accurate segmen