计算机科学

计算机科学 ›› 2025, Vol. 52 ›› Issue (6): 264-273.doi: 10.11896/jsjkx.241200197

• 计算机图形学&多媒体 • 上一篇    下一篇

基于多尺度注意力和不确定性损失的两阶段左心房疤痕分割

张鑫艳1,2, 唐振超3,4, 李一夫5, 刘振宇1,2   

  1. 1 安徽医科大学公共卫生学院 合肥 230032
    2 中国科学院自动化研究所分子影像重点实验室 北京 100190
    3 北京航空航天大学医学科学与工程学院/医工交叉创新研究院 北京 100191
    4 北京航空航天大学大数据精准医学重点实验室 北京 100191
    5 北京航空航天大学卓越工程师学院 北京 100191
  • 收稿日期:2024-12-27 修回日期:2025-03-18 出版日期:2025-06-15 发布日期:2025-06-11
  • 通讯作者: 刘振宇(zhenyu.liu@ia.ac.cn)
  • 作者简介:(zhangxinyan_2022@163.com)
  • 基金资助:
    中央高校基本科研业务费专项资金(YWF-23-Q-1074);国家重点研发计划(2021YFA1301603)

Two-stage Left Atrial Scar Segmentation Based on Multi-scale Attention and Uncertainty Loss

ZHANG Xinyan1,2, TANG Zhenchao3,4, LI Yifu5, LIU Zhenyu1,2   

  1. 1 School of Public Health,Anhui Medical University,Hefei 230032,China
    2 CAS Key Laboratory of Molecular Imaging,Institute of Automation,Beijing 100190,China
    3 Beijing Advanced Innovation Center for Big Data-Based Precision Medicine,School of Engineering Medicine,Beihang University,Beijing 100191,China
    4 Key Laboratory of Big Data-Based Precision Medicine,Beihang University,Ministry of Industry and Information Technology,Beijing 100191,China
    5 National Superior College for Engineers,Beihang University,Beijing 100191,China
  • Received:2024-12-27 Revised:2025-03-18 Online:2025-06-15 Published:2025-06-11
  • About author:ZHANG Xinyan,born in 1996,postgraduate.Her main research interests include deep learning and medical image analysis.
    LIU Zhenyu,born in 1986,professor,Ph.D supervisor.His main research interests include artificial intelligence,pattern recognition,medical imaging big data analysis and multimodal medical image big models.
  • Supported by:
    Fundamental Research Funds for the Central Universities(YWF-23-Q-1074) and National Key R&D Program of China(2021YFA1301603).

PDF (PC)

63

摘要： 心房颤动(AF)是临床上最常见的心律失常之一。左心房及其心肌梗死后疤痕区域的准确分割和面积评估,对于心肌梗死患者出现AF的早期诊断、治疗规划以及预后评估具有极其重要的临床意义。深度学习方法是进行左心房及其心肌梗死后疤痕区域自动分割的主流方向。但是由于心肌梗死后疤痕体积小且容易受到周围增强组织的影响,分割精度尚有待提高。为此,提出了一种基于多尺度注意力和不确定性损失的两阶段深度学习模型。一方面,在网络上采样之前引入多尺度注意力模块(MSAM),该模块能够编码丰富的多尺度语义信息并让模型更为关注重要的语义信息及空间信息。另一方面,引入不确定性损失(Uncertainty Loss)以增强模型对疤痕不确定性的建模能力。此外,还采用直方图匹配(HM)增强图像质量,提高网络的分割能力。将所提出的方法在验证集以及左心房和疤痕量化与分割挑战赛(LAScarQS++)验证平台上进行验证,实验结果均表明该方法分割的疤痕更加完整,分割精度也得到了提升。与nnU-Net相比,心肌梗死后疤痕分割骰子系数(Dice)提高了8.12%。

关键词: 心肌梗死后疤痕, 深度学习, 图像分割, 不确定性损失, nnU-Net, 多尺度注意力

Abstract: Atrial Fibrillation (AF) is one of the most common arrhythmias clinically.Accurate segmentation and area assessment of the left atrium and its scar area after myocardial infarction are of great clinical significance for the early diagnosis,treatment planning and prognosis assessment of AF in patients with myocardial infarction.The deep learning-based method is the mainstream direction for automatic segmentation of the left atrium and the scar area after myocardial infarction.However,as the scar after myocardial infarction is small in size and easily affected by the surrounding enhanced tissue,the segmentation accuracy still remains to be improved.Therefore,a two-stage deep learning model based on multi-scale attention and uncertainty loss is proposed.On the one hand,a Multi-Scale Attention Module (MSAM) is introduced before sampling on the network.This module can encode rich multi-scale semantic information and make the model pay more attention to important semantic and spatial information.On the other hand,uncertainty loss is introduced to enhance the model's ability to model scar uncertainty.In addition,this study also uses histogram matching (HM) to enhance image quality and improve the segmentation ability of the network.The proposed methodis verified on the validation set and the left atrial and scar quantification and segmentation (LAScarQS++) evaluation platform.The experimental results show that the scar segmented by this method is more complete and the segmentation accuracy is also improved.Compared with nnU-Net,the Dice coefficient (Dice) of scar segmentation after myocardial infarction is increased by 8.12%.

Key words: Post-Myocardial infarction scar, Deep learning, Image segmentation, Uncertainty loss, nnU-Net, Multi-scale attention

中图分类号: 

  • TP391.7
[1]SHI S,TANG Y,ZHAO Q,et al.Prevalence and risk of atrial fibrillation in China:A national cross-sectional epidemiological study [J].The Lancet regional health Western Pacific,2022,23:100439.
[2]AMERICAN HEART ASSOCIATION.Living With AFibGuide[EB/OL].https://www.hear-t.org/en/health-topics/atrial-fibrillation.
[3]KARIM R,HOUSDEN R J,BALASUBR-AMANIAM M,et al.Evaluation of current algorithms for segmentation of scar tissue from late Gadolinium enhancement cardiovascular magnetic resonance of the left atrium:an open-access grand challenge [J].Journal of Cardiovascular Magnetic Resonance,2013,15(1):105.
[4]LEFEBVRE A L,YAMAMOTO C A P,SHADE J K,et al.LASSNet:A Four Steps Deep Neural Network for Left Atrial Segmentation and Scar Quantification [C]//Proceedings of the Left Atrial and Scar Quantification and Segmentation:First Challenge,LAScarQS 2022 Held in Conjunction with MICCAI 2022,Singapore,September 18,2022,Proceedings.2023:1-15.
[5]LI L,WU F,YANG G,et al.Atrial scar quantification via multi-scale CNN in the graph-cuts framework [J].Medical Image Analysis,2020,60:101595.
[6]LI L,ZIMMER V A,SCHNABEL J A,et al.AtrialJSQnet:a new framework for joint segmentation and quantification of left atrium and scars incorporating spatial and shape information [J].Medical Image Analysis,2022,76:102303.
[7]ZHOU S,WANG K,ZHOU G.Edge-Enhanced Feature Guided Joint Segmentation of Left Atrial and Scars in LGE MRI Images [C]//Proceedings of the Left Atrial and Scar Quantification and Segmentation.2023:93-105.
[8]AREGA T W,BRICQ S,MERIAUDEAU F.Leveraging Uncertainty Estimates to Improve Segmentation Performance in Cardiac MR [C]//Proceedings of the Uncertainty for Safe Utilization of Machine Learning in Medical Imaging,and Perinatal Imaging,Placental and Preterm Image Analysis.2021:24-33.
[9]MA J.Histogram Matching Augmentation for Domain Adaptation with Application to Multi-centre,Multi-vendor and Multi-disease Cardiac Image Segmentation [C]//Proceedings of the Statistical Atlases and Computational Models of the Heart M&Ms and EMIDEC Challenges.2021:177-186.
[10]SU R,ZHANG D,LIU J,et al.MSU-Net:Multi-scale U-Net for 2D medical image segmentation [J].Frontiers inGenetics,2021,12:639930.
[11]KHAN A,ALWAZZAN O,BENNING M,et al.Sequential Segmentation of the Left Atrium and Atrial Scars Using a Multi-scale Weight Sharing Network and Boundary-Based Processing [C]//Proceedings of the Left Atrial and Scar Quantification and Segmentation.2023:69-82.
[12]LI F,LI W.Cross-Domain Segmentation of Left Atrium Basedon Multi-scale Decision Level Fusion [C]//Proceedings of the Left Atrial and Scar Quantification and Segmentation.2023:124-132.
[13]ISENSEE F,JAEGER P F,KOHL S A A,et al.nnU-Net:a self-configuring method for deep learning-based biomedical image segmentation [J].Nature Methods,2021,18(2):203-211.
[14]HU J,SHEN L,SUN G.Squeeze-and-Excitation Networks[C]//Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.2018:7132-7141.
[15]WOO S,PARK J,LEE J,et al.Cbam:Convolutional block attention module [C]//Proceedings of the European Conference on Computer Vision (ECCV).2018:3-19.
[16]PARK J,WOO S,LEE J,et al.BAM:Bottleneck AttentionModule [C]//Proceedings of the British Machine Vision Confe-rence.2018.
[17]NG M,GUO F,BISWAS L,et al.Estimating Uncertainty inNeural Networks for Cardiac MRI Segmentation:A Benchmark Study [J].IEEE Transactions on Biomedical Engineering,2023,70(6):1955-1966.
[18]SANDER J,VOS B D,WOLTERINK J M,et al.Towards increased trustworthiness of deep learning segmentation methods on cardiac MRI [C]//Proceedings of the Medical imaging 2019:image Processing.2019:324-330.
[19]YANG X,WANG N,WANG Y,et al.Combating Uncertainty with Novel Losses for Automatic Left Atrium Segmentation [C]//Proceedings of the Statistical Atlases and Computational Models of the Heart Atrial Segmentation and LV Quantification Challenges.2019:246-254.
[20]CHEN L C,ZHU Y,PAPANDREOU G,et al.Encoder-decoder with atrous separable convolution for semantic image segmentation [C]//Proceedings of the Proceedings of the European Conference on Computer Vision (ECCV).2018:801-818.
[21]OGBOMO-HARMITT S,GRZELAK J,QURESHI A,et al.TESSLA:Two-Stage Ensemble Scar Segmentation for the Left Atrium [C]//Proceedings of the Left Atrial and Scar Quantification and Segmentation.2023:106-114.
[22]BEINART R,ABBARA S,BLUM A,et al.Left atrial wallthickness variability measured by CT scans in patients undergoing pulmonary vein isolation [J].Journal of Cardiovascular Electrophysiology,2011,22(11):1232-1236.
[23]ZHANG X,YANG X,HUANG L,et al.Two Stage of Histogram Matching Augmentation for Domain Generalization:Application to Left Atrial Segmentation [C]//Proceedings of the Left Atrial and Scar Quantification and Segmentation.2023:60-68.
[24]SHAPIRA D,AVIDAN S,HEL-OR Y.Multiple histogrammatching [C]//Proceedings of the 2013 IEEE International Conference on Image Processing.2013:2269-2273.
[25]LI L,ZIMMER V A,SCHNABEL J A,et al.Medical imageanalysis on left atrial LGE MRI for atrial fibrillation studies:A review [J].MedicalImage Analysis,2022,77:102360.
[26]LI L,ZIMMER V A,SCHNABEL J A,et al.AtrialGeneral:domain generalization for left atrial segmentation of multi-center LGE MRIs [C]//Proceedings of the Medical Image Computing and Computer Assisted Intervention(MICCAI 2021).24th International Conference.2021:557-566.
[27]YANIV Z,LOWEKAMP B C,JOHNSON H J,et al.SimpleITK image-analysis notebooks:a collaborative environment for education and reproducible research [J].Journal of Digital Imaging,2018,31(3):290-303.
[28]LI L,ZIMMER V A,SCHNABEL J A,et al.Medical imageanalysis on left atrial LGE MRI for atrial fibrillation studies:A review [J].Medical Image Analysis,2022,77:102360.
[29]OKTAY O,SCHLEMPER J,FOLGOC L L,et al.Attention u-net:Learning where to look for the pancreas [J].arXiv:1804.03999,2018.
[30]DIAKOGIANNIS F I,WALDNER F,CACCETTA P,et al.ResUNet-a:A deep learning framework for semantic segmentation of remotely sensed data [J].ISPRS Journal of Photogrammetry Remote Sensing,2020,162:94-114.
[31]WU M,DING W,YANG M,et al.Multi-depth boundary-aware left atrial scar segmentation network [C]//Proceedings of the Challenge on Left Atrial and Scar Quantification and Segmentation.2022:16-23.
[32]WU J,JI W,FU H,et al.Medsegdiff-v2:Diffusion-based medical image segmentation with transformer [C]//Proceedings of the AAAI Conference on Artificial Intelligence.2023:6030-6038.
[33]ZHANG Y,MENG Y,ZHENG Y.Automatically segment the left atrium and scars from LGE-MRIs using a boundary-focused nnU-Net [C]//Proceedings of the Challenge on Left Atrial and Scar Quantification and Segmentation.2022:49-59.
[34]AREGA T W,BRICQ S,MERIAUDEAU F.Using Polynomial Loss and Uncertainty Information for Robust Left Atrial and Scar Quantification and Segmentation [C]//Proceedings of the Left Atrial and Scar Quantification and Segmentation.2023:133-144.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发