RSU-Net: U-net Based on Residual and Self-Attention Mechanism in the Segmentation of Cardiac Magnetic Resonance Images.
Yuan-Zhe Li,Yi Wang,Yin-Hui Huang,Ping Xiang,Wen-Xi Liu,Qing-Quan Lai,Yi-Yuan Gao,Mao-Sheng Xu,Yi-Fan Guo
DOI: https://doi.org/10.1016/j.cmpb.2023.107437
IF: 6.1
2023-01-01
Computer Methods and Programs in Biomedicine
Abstract:Background: Automated segmentation techniques for cardiac magnetic resonance imaging (MRI) are ben-eficial for evaluating cardiac functional parameters in clinical diagnosis. However, due to the character-istics 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 un-certainty 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 uncer-tain and discontinuous. Therefore, fast and accurate segmentation of cardiac tissue remains a challenging problem in medical image processing.Methodology: We collected cardiac MRI data from 195 patients as training set and 35patients from differ-ent 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 net-work 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 con-volutional 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.Results: 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 segmentation of the heart. New ideas for scientific research. Conclusion: Our proposed RSU-Net network combines the advantages of residual connections and self -attention. This paper uses the residual links to facilitate the training of the network. In this paper, a self -attention mechanism is introduced, and a bottom self-attention block (BSA Block) is used to aggregate global information. Self-attention aggregates global information, and has achieved good segmentation re-sults on the cardiac segmentation dataset. It facilitates the diagnosis of cardiovascular patients in the future.(c) 2023 Elsevier B.V. All rights reserved.