CSWin-UNet: Transformer UNet with cross-shaped windows for medical image segmentation

被引:8
作者
Liu, Xiao [1 ]
Gao, Peng [1 ,3 ]
Yu, Tao [1 ]
Wang, Fei [2 ]
Yuan, Ru-Yue
机构
[1] Qufu Normal Univ, Sch Cyber Sci & Engn, Qufu, Peoples R China
[2] Harbin Inst Technol, Sch Integrated Circuits, Shenzhen, Peoples R China
[3] Yuntian Educ Grp, Hangzhou 253700, Peoples R China
关键词
Medical image segmentation; Deep learning; Attention mechanism; Neural network;
D O I
10.1016/j.inffus.2024.102634
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Deep learning, especially convolutional neural networks (CNNs) and Transformer architectures, have become the focus of extensive research in medical image segmentation, achieving impressive results. However, CNNs come with inductive biases that limit their effectiveness in more complex, varied segmentation scenarios. Conversely, while Transformer-based methods excel at capturing global and long-range semantic details, they suffer from high computational demands. In this study, we propose CSWin-UNet, a novel U-shaped segmentation method that incorporates the CSWin self-attention mechanism into the UNet to facilitate horizontal and vertical stripes self-attention. This method significantly enhances both computational efficiency and receptive field interactions. Additionally, our innovative decoder utilizes a content-aware reassembly operator that strategically reassembles features, guided by predicted kernels, for precise image resolution restoration. Our extensive empirical evaluations on diverse datasets, including synapse multi-organ CT, cardiac MRI, and skin lesions demonstrate that CSWin-UNet maintains low model complexity while delivering high segmentation accuracy.
引用
收藏
页数:12
相关论文
共 55 条
[21]  
Fang YX, 2021, ADV NEUR IN
[22]   SOTR: Segmenting Objects with Transformers [J].
Guo, Ruohao ;
Niu, Dantong ;
Qu, Liao ;
Li, Zhenbo .
2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021), 2021, :7137-7146
[23]   UNETR: Transformers for 3D Medical Image Segmentation [J].
Hatamizadeh, Ali ;
Tang, Yucheng ;
Nath, Vishwesh ;
Yang, Dong ;
Myronenko, Andriy ;
Landman, Bennett ;
Roth, Holger R. ;
Xu, Daguang .
2022 IEEE WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV 2022), 2022, :1748-1758
[24]   HiFormer: Hierarchical Multi-scale Representations Using Transformers for Medical Image Segmentation [J].
Heidari, Moein ;
Kazerouni, Amirhossein ;
Soltany, Milad ;
Azad, Reza ;
Aghdam, Ehsan Khodapanah ;
Cohen-Adad, Julien ;
Merhof, Dorit .
2023 IEEE/CVF WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV), 2023, :6191-6201
[25]  
Ho JAT, 2019, Arxiv, DOI arXiv:1912.12180
[26]   CCNet: Criss-Cross Attention for Semantic Segmentation [J].
Huang, Zilong ;
Wang, Xinggang ;
Wei, Yunchao ;
Huang, Lichao ;
Shi, Humphrey ;
Liu, Wenyu ;
Huang, Thomas S. .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2023, 45 (06) :6896-6908
[27]   nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation [J].
Isensee, Fabian ;
Jaeger, Paul F. ;
Kohl, Simon A. A. ;
Petersen, Jens ;
Maier-Hein, Klaus H. .
NATURE METHODS, 2021, 18 (02) :203-+
[28]   Learning Calibrated Medical Image Segmentation via Multi-rater Agreement Modeling [J].
Ji, Wei ;
Yu, Shuang ;
Wu, Junde ;
Ma, Kai ;
Bian, Cheng ;
Bi, Qi ;
Li, Jingjing ;
Liu, Hanruo ;
Cheng, Li ;
Zheng, Yefeng .
2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021, 2021, :12336-12346
[29]   Swin Transformer: Hierarchical Vision Transformer using Shifted Windows [J].
Liu, Ze ;
Lin, Yutong ;
Cao, Yue ;
Hu, Han ;
Wei, Yixuan ;
Zhang, Zheng ;
Lin, Stephen ;
Guo, Baining .
2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021), 2021, :9992-10002
[30]  
Long J, 2015, PROC CVPR IEEE, P3431, DOI 10.1109/CVPR.2015.7298965