[1]陈沐萱,袁金丽,郭志涛,等.融合改进Transformer和卷积通道注意力模块的U-Net用于双心室分割[J].中国医学物理学杂志,2024,41(1):32-42.[doi:DOI:10.3969/j.issn.1005-202X.2024.01.005]
　CHEN Muxuan,YUAN Jinli,GUO Zhitao,et al.Biventricular segmentation using U-Net incorporating improved Transformer and convolutional channel attention module[J].Chinese Journal of Medical Physics,2024,41(1):32-42.[doi:DOI:10.3969/j.issn.1005-202X.2024.01.005]
点击复制

融合改进Transformer和卷积通道注意力模块的U-Net用于双心室分割()
分享到：

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

卷:
41卷
期数:
2024年第1期
页码:
32-42
栏目:
医学影像物理
出版日期:
2024-01-23

文章信息/Info

Title:
Biventricular segmentation using U-Net incorporating improved Transformer and convolutional channel attention module
文章编号:
1005-202X（2024）01-0032-11
作者:
陈沐萱; 袁金丽; 郭志涛; 卢成钢
河北工业大学电子信息工程学院， 天津 300401
Author(s):
CHEN Muxuan;  YUAN Jinli;  GUO Zhitao;  LU Chenggang
School of Electronic Information Engineering, Hebei University of Technology, Tianjin 300401, China
关键词:
双心室分割; 图像处理; Transformer; 注意力机制; 特征提取
Keywords:
Keywords: biventricular segmentation image processing Transformer attention mechanism feature extraction
分类号:
R318；TP391
DOI:
DOI:10.3969/j.issn.1005-202X.2024.01.005
文献标志码:
A
摘要:
设计一种融合改进Transformer和卷积通道注意力模块的U-Net用于MRI图像双心室分割。通过在U-Net的高层卷积部分基础融合改进Transformer，有效增强全局特征信息的提取能力以应对右心室复杂的形态变化造成低分割性能的难题。改进的Transformer在自注意力模块部分中加入固定窗口注意力进行位置定位，随后对其输出特征图进行聚合以缩小特征图尺寸；同时通过改进多层感知器来加深网络深度以提高网络学习能力。为解决组织边缘模糊造成的分割性能不理想问题，引入特征聚合模块进行多层次底层特征的融合，利用卷积通道注意力模块对底层特征进行重标定，实现自适应地学习特征权重。此外，针对编解码结构中通道衰减造成特征丢失导致的低分割性能，网络集成一个即插即用的特征增强模块，保证空间信息同时增加有用通道信息的比重。在ACDC数据集对本文算法进行测试，结果表明本文方法对左右心室的分割精度均优于近年其他算法，尤其是右心室分割结果，相比于其他方法，DSC系数提高至少2.83%，证明本文方法对双心室分割的有效性。
Abstract:
Abstract: A U-Net incorporating improved Transformer and convolutional channel attention module is designed for biventricular segmentation in MRI image. By replacing the high-level convolution of U-Net with the improved Transformer, the global feature information can be effectively extracted to cope with the challenge of poor segmentation performance due to the complex morphological variation of the right ventricle. The improved Transformer incorporates a fixed window attention for position localization in the self-attention module, and aggregates the output feature map for reducing the feature map size and the network learning capability is improved by increasing network depth through the adjustment of multilayer perceptron. To solve the problem of unsatisfactory segmentation performance caused by blurred tissue edges, a feature aggregation module is used for the fusion of multi-level underlying features, and a convolutional channel attention module is adopted to rescale the underlying features to achieve adaptive learning of feature weights. In addition, a plug-and-play feature enhancement module is integrated to improve the segmentation performance which is affected by feature loss due to channel decay in the codec structure, which guarantees the spatial information while increasing the proportion of useful channel information. The test on the ACDC dataset shows that the proposed method has higher biventricular segmentation accuracy, especially for the right ventricle segmentation. Compared with other methods, the proposed method improves the DSC coefficient by at least 2.83%, proving its effectiveness in biventricular segmentation.

相似文献/References:

[1]田 龙,李明辉,胡逸民,等.利用EPID图像金球位置自动跟踪算法研究[J].中国医学物理学杂志,2014,31(04):5038.[doi:10.3969/j.issn.1005-202X.2014.04.014]
[2]刘玉红,王志芳,杨佳仪,等.彩色图像二值化算法及应用[J].中国医学物理学杂志,2013,30(01):3873.[doi:10.3969/j.issn.1005-202X.2013.01.009]
[3]杜永兴,桑路路,秦岭,等.一种基于图像匹配的微波热疗方法[J].中国医学物理学杂志,2016,33(10):1026.[doi:10.3969/j.issn.1005-202X.2016.10.011]
　[J].Chinese Journal of Medical Physics,2016,33(1):1026.[doi:10.3969/j.issn.1005-202X.2016.10.011]
[4]朱翔宇,葛中芹,张冰清,等.基于图像处理的医学影像处理平台系统设计[J].中国医学物理学杂志,2017,34(4):388.[doi:DOI:10.3969/j.issn.1005-202X.2017.04.014]
[5]金京,孟雅丹,徐琦,等.二维分形维数在皮肤癌光学相干断层成像技术诊断中的应用[J].中国医学物理学杂志,2017,34(7):698.[doi:10.3969/j.issn.1005-202X.2017.07.010]
　[J].Chinese Journal of Medical Physics,2017,34(1):698.[doi:10.3969/j.issn.1005-202X.2017.07.010]
[6]王耿媛,赖官铨,段铮昱,等. 基于裂隙灯的眼微血管形态学参数分析在眼科疾病诊断的应用[J].中国医学物理学杂志,2018,35(3):323.[doi:DOI:10.3969/j.issn.1005-202X.2018.03.014]
　WANG Gengyuan,LAI Guanquan,DUAN Zhengyu,et al. Application of morphological parameters analysis of ophthalmic microvasculature based on slit lamp in the diagnosis of ophthalmic diseases[J].Chinese Journal of Medical Physics,2018,35(1):323.[doi:DOI:10.3969/j.issn.1005-202X.2018.03.014]
[7]王力,孟庆民,郭永新,等. 基于Hessian矩阵的DSA图像冠状动脉直径的测量[J].中国医学物理学杂志,2019,36(10):1182.[doi:DOI:10.3969/j.issn.1005-202X.2019.10.013]
　WANG Li,MENG Qingmin,GUO Yongxin,et al. Measurement of coronary artery diameter on DSA image based on Hessian matrix[J].Chinese Journal of Medical Physics,2019,36(1):1182.[doi:DOI:10.3969/j.issn.1005-202X.2019.10.013]
[8]孙凯,姚旭峰,马风玲,等.基于机器学习的血细胞分类研究进展[J].中国医学物理学杂志,2020,37(1):127.[doi:DOI:10.3969/j.issn.1005-202X.2020.01.023]
　SUN Kai,YAO Xufeng,et al.Blood cell classification based on machine learning[J].Chinese Journal of Medical Physics,2020,37(1):127.[doi:DOI:10.3969/j.issn.1005-202X.2020.01.023]
[9]惠钊,黄慧明.基于比率距离的自适应超声图像去噪方法[J].中国医学物理学杂志,2020,37(2):174.[doi:DOI:10.3969/j.issn.1005-202X.2020.02.008]
　HUI Zhao,HUANG Huiming.Adaptive ultrasound image denoising method based on ratio distance[J].Chinese Journal of Medical Physics,2020,37(1):174.[doi:DOI:10.3969/j.issn.1005-202X.2020.02.008]
[10]王尧,黄钢.深度学习在细胞图像分析中的应用进展[J].中国医学物理学杂志,2022,39(8):963.[doi:DOI:10.3969/j.issn.1005-202X.2022.08.008]
　WANG Yao,HUANG Gang,Advances in application of deep learning in cell image analysis[J].Chinese Journal of Medical Physics,2022,39(1):963.[doi:DOI:10.3969/j.issn.1005-202X.2022.08.008]

备注/Memo

备注/Memo:
【收稿日期】2023-08-06 【基金项目】河北省教育厅重点项目（ZD2022115） 【作者简介】陈沐萱，硕士研究生，研究方向：计算机视觉、机器学习、医疗图像处理，E-mail: cmxstudent@163.com 【通信作者】袁金丽，博士，副教授，研究方向：智能信息处理、计算机视觉、机器学习，E-mail: jinli_yuan@hebut.edu.cn
更新日期/Last Update: 2024-01-23