[1]金丰庆,周月娥,索国栋,等.SSCU-UNet：融合空间位移与内容感知注意力的医学图像分割模型[J].中国医学物理学杂志,2026,43(6):787-797.[doi:DOI:10.3969/j.issn.1005-202X.2026.06.012]
　JIN Fengqing,ZHOU Yuee,et al.SSCU-UNet: a medical image segmentation network integrating spatial shift and content-aware attention[J].Chinese Journal of Medical Physics,2026,43(6):787-797.[doi:DOI:10.3969/j.issn.1005-202X.2026.06.012]
点击复制

SSCU-UNet：融合空间位移与内容感知注意力的医学图像分割模型()

《中国医学物理学杂志》[ISSN:1005-202X/CN:44-1351/R]

卷:
43卷
期数:
2026年第6期
页码:
787-797
栏目:
医学影像物理
出版日期:
2026-06-26

文章信息/Info

Title:
SSCU-UNet: a medical image segmentation network integrating spatial shift and content-aware attention
文章编号:
1005-202X（2026）06-0787-11
作者:
金丰庆1; 2; 周月娥1; 2; 索国栋1; 2; 杨建兰2
1.甘肃中医药大学医学信息工程学院， 甘肃 兰州 730000； 2.泉州市正骨医院， 福建 泉州 362000
Author(s):
JIN Fengqing1;  2;  ZHOU Yuee1;  2;  SUO Guodong1;  2;  YANG Jianlan2
1. School of Medicine Information Science and Engineering, Gansu University of Chinese Medicine, Lanzhou 730000, China 2. Quanzhou Orthopedic-Traumatological Hospital, Quanzhou 362000, China
关键词:
医学图像分割; SSCU-UNet; Swin Transformer; 空间位移模块; 内容感知上采样; 多器官分割
Keywords:
Keywords: medical image segmentation spatial shift-enhanced and content-aware upsampling UNet Swin Transformer spatial shift module content-aware upsampling multi-organ segmentation
分类号:
R318；TP391.41
DOI:
DOI:10.3969/j.issn.1005-202X.2026.06.012
文献标志码:
A
摘要:
目的：设计一种能兼具全局感知与高分辨率特征保持能力的分割框架SSCU-UNet，以提升多器官CT与心脏MRI的分割性能。方法：提出SSCU-UNet分割框架，包含3个核心设计：（1）残差卷积下采样模块采用双路径结构，通过Conv-GELU-Conv主路径与步长卷积快捷路径融合，并结合层归一化以保留高分辨率空间特征。（2）SSFormer模块，位于每个Swin Transformer块中的窗口注意力和多层感知器之间，沿残差分支运行，以增强局部空间建模，同时通过空间位移注意力保持全局上下文推理。（3）内容感知注意力上采样模块根据输入特征自适应生成重建权重，以实现更精确的边界恢复。结果：在Synapse数据集上，SSCU-UNet的平均Dice相似系数为82.62%，较U-Net、TransUNet与Swin-UNet分别提升5.77%、5.14%和3.49%，Hausdorff距离降至18.79 mm，显著改善胆囊、胰腺、胃等形态复杂器官的边界精度。在ACDC数据集上，平均Dice相似系数达92.34%，优于多种基线模型。结论：SSCU-UNet通过整合残差卷积下采样、空间位移增强局部建模和内容感知上采样，有效平衡全局语义推理与细粒度结构恢复，在多器官分割任务中取得先进性能。
Abstract:
Abstract: Objective To design a segmentation framework, namely spatial shift-enhanced and content-aware upsampling UNet (SSCU-UNet), which balances global context awareness and high-resolution feature retention to improve performance in multi-organ CT and cardiac MRI segmentation tasks. Methods The proposed SSCU-UNet segmentation framework incorporated 3 key components. (1) A residual convolutional downsampling module adopted a dual-branch structure consisting of a Conv-GELU-Conv main branch and a strided-convolution shortcut branch, and coupled with layer normalization to retain high-resolution spatial information. (2) A SSFormer module was inserted between the window attention and multilayer perceptron of each Swin Transformer block, and operates along the residual branch to enhance local spatial modeling while maintaining global contextual reasoning via spatial shift attention. (3) A content-aware attention upsampling module was used to enable finer boundary restoration, which adaptively generates feature reconstruction weights according to input features. Results The proposed SSCU-UNet achieved an average Dice similarity coefficient of 82.62% on the Synapse dataset, which was 5.77%, 5.14%, and 3.49% higher than U-Net, TransUNet, and Swin-UNet, respectively, and reduced the Hausdorff distance to 18.79 mm, with particularly prominent performance improvements for anatomically complex organs including the gallbladder, pancreas, and stomach. On the ACDC dataset, the SSCU-UNet yielded an average Dice similarity coefficient of 92.34%, outperforming multiple baseline models. Conclusion By integrating residual convolutional downsampling, spatial shift-enhanced local modeling, and content-aware upsampling, the SSCU-UNet effectively balances global semantic reasoning with fine-grained structural reconstruction, and achieves state-of-the-art performance in multi-organ segmentation tasks.

相似文献/References:

[1]万俊,聂生东,王远军,等.基于MRI的脑肿瘤分割技术研究进展[J].中国医学物理学杂志,2013,30(04):4266.[doi:10.3969/j.issn.1005-202X.2013.04.013]
[2]林江,戴齐,欧阳婷雪,等.一种边界和马尔可夫随机场相结合的脑MRI医学图像分割方法[J].中国医学物理学杂志,2015,32(05):717.[doi:doi:10.3969/j.issn.1005-202X.2015.05.023]
　[J].Chinese Journal of Medical Physics,2015,32(6):717.[doi:doi:10.3969/j.issn.1005-202X.2015.05.023]
[3]吴迪,胡胜,刘伟峰,等. 基于特征融合视觉显著性的医学图像分割[J].中国医学物理学杂志,2018,35(6):670.[doi:DOI:10.3969/j.issn.1005-202X.2018.06.010]
　WU Di,HU Sheng,LIU Weifeng,et al. Medical image segmentation based on visual saliency of feature fusion[J].Chinese Journal of Medical Physics,2018,35(6):670.[doi:DOI:10.3969/j.issn.1005-202X.2018.06.010]
[4]刘国才,官文静,田娟秀,等. 集成自适应回归核的肿瘤生物靶区随机游走勾画方法[J].中国医学物理学杂志,2018,35(7):758.[doi:DOI:10.3969/j.issn.1005-202X.2018.07.004]
　LIU Guocai,GUAN Wenjing,TIAN Juanxiu,et al. A random walk method with adaptive regression-kernel for delineation of biological target volumes of tumors[J].Chinese Journal of Medical Physics,2018,35(6):758.[doi:DOI:10.3969/j.issn.1005-202X.2018.07.004]
[5]宫进昌,赵尚义,王远军. 基于深度学习的医学图像分割研究进展[J].中国医学物理学杂志,2019,36(4):420.[doi:DOI:10.3969/j.issn.1005-202X.2019.04.010]
　GONG Jinchang,ZHAO Shangyi,WANG Yuanjun.Research progress on deep learning-based medical image segmentation[J].Chinese Journal of Medical Physics,2019,36(6):420.[doi:DOI:10.3969/j.issn.1005-202X.2019.04.010]
[6]张富利,杨安宁,路娜,等.基于FC_DenseNet深度学习网络自动分割肺癌放疗中的危及器官[J].中国医学物理学杂志,2021,38(2):259.[doi:DOI:10.3969/j.issn.1005-202X.2021.02.024]
　ZHANG Fuli,YANG Anning,LU Na,et al.FC_Densenet-based autosegmentation of organs-at-risk in lung cancer radiotherapy[J].Chinese Journal of Medical Physics,2021,38(6):259.[doi:DOI:10.3969/j.issn.1005-202X.2021.02.024]
[7]方新林,方艳红,王迪.基于多模态特征融合的脑瘤图像分割方法[J].中国医学物理学杂志,2022,39(6):682.[doi:DOI:10.3969/j.issn.1005-202X.2022.06.005]
　FANG Xinlin,FANG Yanhong,WANG Di.Brain tumor image segmentation method based on multi-modal feature fusion[J].Chinese Journal of Medical Physics,2022,39(6):682.[doi:DOI:10.3969/j.issn.1005-202X.2022.06.005]
[8]魏坤,沈记全,赵艳梅.MAUNet：用于皮肤病变分割的轻量级模型[J].中国医学物理学杂志,2023,40(5):555.[doi:DOI:10.3969/j.issn.1005-202X.2023.05.006]
　WEI Kun,SHEN Jiquan,ZHAO Yanmei.MAUNet: a lightweight model for skin lesion segmentation[J].Chinese Journal of Medical Physics,2023,40(6):555.[doi:DOI:10.3969/j.issn.1005-202X.2023.05.006]
[9]周金治,胡震,郭莉莉,等.基于GAN-DAUnet的肝脏CT图像肿瘤分割算法[J].中国医学物理学杂志,2023,40(8):971.[doi:DOI:10.3969/j.issn.1005-202X.2023.08.008]
　ZHOU Jinzhi,HU Zhen,et al.Liver CT image tumor segmentation algorithm based on GAN-DAUnet[J].Chinese Journal of Medical Physics,2023,40(6):971.[doi:DOI:10.3969/j.issn.1005-202X.2023.08.008]
[10]黄仟甲,张恒,李奇轩,等.医学图像分割的研究进展[J].中国医学物理学杂志,2024,41(8):939.[doi:DOI:10.3969/j.issn.1005-202X.2024.08.003]
　HUANG Qianjia,ZHANG Heng,,et al.Review on medical image segmentation methods[J].Chinese Journal of Medical Physics,2024,41(6):939.[doi:DOI:10.3969/j.issn.1005-202X.2024.08.003]

备注/Memo

备注/Memo:
【收稿日期】2025-12-10 【基金项目】泉州市科技项目（2023N002S） 【作者简介】金丰庆，硕士研究生，研究方向：医学图像分割与深度学习，E-mail: ff_001024@163.com 【通信作者】杨建兰，副教授，研究方向：医学影像识别与应用、卫生信息管理系统开发与研究、卫生信息数据挖掘，E-mail: fjyjl@gszy.edu.cn
更新日期/Last Update: 2026-06-29