基于非零水平集保凸算法的左心室MRI分割

doi:10.6040/j.issn.1671-9352.0.2024.063

摘要/Abstract

摘要： 心脏左心室分割临床应用要求是分割的左心室保持凸形且包含左心室腔、小梁和乳头肌,提出一个包含改进距离正则项和非零水平集保凸项的心脏核磁共振成像分割模型,其利用水平集轮廓的曲率来保持凸性,从而使轮廓最终演化为凸形。使用ACDC MICCAI 2017数据集进行模型评估,该模型在心脏舒张末期和收缩末期阶段的平均Dice系数分别为0.961和0.936,平均豪斯多夫距离分别为4.89和5.79。同时该模型无需对训练数据进行人工标注和学习,分割精度和鲁棒性均可以达到与基于深度学习的左心室分割模型相同的分割性能。

关键词: 非零水平集, 保凸, 水平集方法, 左心室分割, 距离正则化, 双阱势函数

Abstract: Accurate segmentation of the left ventricle in clinical applications requires maintaining a convex shape that encompasses the left ventricle cavity, trabeculae, and papillary muscles. The nonzero level set convexity preserving model, a novel cardiac magnetic resonance imaging segmentation model incorporating an enhanced distance regularization term and a nonzero level set convexity preserving term is introduced. By leveraging the curvature of the level set contour, the model effectively promotes convexity, ensuring the contour evolves into a convex shape. Evaluated on the ACDC MICCAI 2017 datasets, the model achieved a mean Dice coefficient of 0.961 and 0.936 in end-diastole and end-systole phases, respectively, alongside a mean Hausdorff distance of 4.89 and 5.79. Notably, the model eliminates the need for manual annotation of training data and time-consuming learning processes, while achieving segmentation accuracy and robustness comparable to deep learning-based left ventricle segmentation models.

Key words: nonzero level set, convexity preserving, level set method, left ventricle segmentation, distance regularization, double well potential function

中图分类号:

TP391

李季,刘艾汶,秦柳. 基于非零水平集保凸算法的左心室MRI分割[J]. 《山东大学学报(理学版)》, 2025, 60(7): 32-47.

LI Ji, LIU Aiwen, QIN Liu. Left ventricular MRI segmentation based on nonzero level sets convexity preserving algorithm[J]. JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE), 2025, 60(7): 32-47.

参考文献

[1] GAIDAI O, CAO Y, LOGINOV S. Global cardiovascular diseases death rate prediction[J]. Current Problems in Cardiology, 2023, 48(5):101622.
[2] ZENG X, STAIB L, SCHULTZ R, et al. Volumetric layer segmentation using coupled surfaces propagation[C] //1998 IEEE Computer Society Conference on Computer Vision and Pattern Recognition(Cat. No.98CB36231). Santa Barbara:IEEE, 1998:708-715.
[3] PRATONDO A, CHUI C K, ONG S H. Robust edge-stop functions for edge-based active contour models in medical image segmentation[J]. IEEE Signal Processing Letters, 2016, 23(2):222-226.
[4] LYNCH M, GHITA O, WHELAN P F. Left-ventricle myocardium segmentation using a coupled level-set with a priori know-ledge[J]. Computerized Medical Imaging and Graphics, 2006, 30(4):255-262.
[5] CHAN T F, VESE L A. Active contours without edges[J]. IEEE Transactions on Image Processing, 2001, 10(2):266-277.
[6] LI C, KAO C, GORE J, et al. Minimization of region-scalable fitting energy for image segmentation[J]. IEEE Transactions on Image Processing, 2008, 17(10):1940-1949.
[7] LI Chunming, XU Chenyang, GUI Changfeng, et al. Distance regularized level set evolution and its application to image segmentation[J]. IEEE Transactions on Image Processing, 2010, 19(12):3243-3254.
[8] SHI Xue, LI Chunming. Convexity preserving level set for left ventricle segmentation[J]. Magnetic Resonance Imaging, 2021, 78:109-118.
[9] KIM Y C, KIM S M, CHOE Y H. Robust semi-automated quantification of cardiac MR perfusion using level set: application to hypertrophic cardiomyopathy patient data[J]. Computers in Biology and Medicine, 2016, 71:162-173.
[10] CIECHOLEWSKI M. An edge-based active contour model using an inflation/deflation force with a damping coefficient[J]. Expert Systems with Applications, 2016, 44:22-36.
[11] WALI S, LI C, IMRAN M, et al. Level-set evolution for medical image segmentation with alternating direction method of multipliers[J]. Signal Processing, 2023, 211:109105.
[12] WANG Lei, CHANG Yan, WANG Hui, et al. An active contour model based on local fitted images for image segmentation[J]. Information Sciences, 2017, 418-419:61-73.
[13] WANG Lei, CHEN Guangqiang, SHI Dai, et al. Active contours driven by edge entropy fitting energy for image segmentation[J]. Signal Processing, 2018, 149:27-35.
[14] WANG Lei, ZHANG Libo, YANG Xiaodong, et al. Level set based segmentation using local fitted images and inhomogeneity entropy[J]. Signal Processing, 2020, 167:107297.
[15] SHI Xue, LI Chunming. Anatomical knowledge based level set segmentation of cardiac ventricles from MRI[J]. Magnetic Resonance Imaging, 2022, 86:135-148.
[16] LEE H, CODELLA N, CHAM M, et al. Automatic left ventricle segmentation using iterative thresholding and an active contour model with adaptation on short-axis cardiac MRI[J]. IEEE Transactions on Biomedical Engineering, 2010, 57(4):905-913.
[17] FLETT A S, HASLETON J, COOK C, et al. Evaluation of techniques for the quantification of myocardial scar of differing etiology using cardiac magnetic resonance[J]. JACC: Cardiovascular Imaging, 2011, 4(2):150-156.
[18] BERNIER M, JODOIN P M, HUMBERT O, et al. Graph cut-based method for segmenting the left ventricle from MRI or echocardiographic images[J]. Computerized Medical Imaging and Graphics, 2017, 58:1-12.
[19] LI Lei, WU Fuping, YANG Guang, et al. Atrial scar quantification via multi-scale CNN in the graph-cuts framework[J]. Medical Image Analysis, 2020, 60:101595.
[20] BALAFAR M A. Gaussian mixture model based segmentation methods for brain MRI images[J]. Artificial Intelligence Review, 2014, 41(3):429-439.
[21] LIU Jie, ZHUANG Xiahai, WU Lianming, et al. Myocardium segmentation from DE MRI using multicomponent gaussian mixture model and coupled level set[J]. IEEE Transactions on Biomedical Engineering, 2017, 64(11):2650-2661.
[22] DONG Suyu, LUO Gongning, WANG Kuanquan, et al. A combined fully convolutional networks and deformable model for automatic left ventricle segmentation based on 3D echocardiography[J]. BioMed Research International, 2018, 2018:1-16.
[23] CHARTSIAS A, JOYCE T, PAPANASTASIOU G, et al. Factorised spatial representation learning: application in semi-supervised myocardial segmentation[M] //FRANGI A F, SCHNABEL J A, DAVATZIKOS C, et al. Medical Image Computing and Computer Assisted Intervention-MICCAI 2018: 11071. Cham: Springer, 2018:490-498.
[24] ZHANG D, ICKE I, DOGDAS B, et al. A multi-level convolutional LSTM model for the segmentation of left ventricle myocardium in infarcted porcine cine MR images[C] //2018 IEEE 15th International Symposium on Biomedical Imaging(ISBI 2018). Washington: IEEE, 2018:470-473.
[25] DU Xiuquan, SONG Yuhui, LIU Yueguo, et al. An integrated deep learning framework for joint segmentation of blood pool and myocardium[J]. Medical Image Analysis, 2020, 62:101685.
[26] XIE Liping, SONG Yi, CHEN Qiang. Automatic left ventricle segmentation in short-axis MRI using deep convolutional neural networks and central-line guided level set approach[J]. Computers in Biology and Medicine, 2020, 122:103877.
[27] TRAGAKIS A, KAUL C, MURRAY-SMITH R, et al. The fully convolutional transformer for medical image segmentation[C] //2023 IEEE/CVF Winter Conference on Applications of Computer Vision(WACV). Waikoloa: IEEE, 2023:3649-3658.
[28] KATO S, HOTTA K. Adaptive t-vMF dice loss: an effective expansion of dice loss for medical image segmentation[J]. Computers in Biology and Medicine, 2024, 168:107695.
[29] LI Chunming, HUANG Rui, DING Zhaohua, et al. A level set method for image segmentation in the presence of intensity inhomogeneities with application to MRI[J]. IEEE Transactions on Image Processing, 2011, 20(7):2007-2016.
[30] LIU Y, CAPTUR G, MOON J C, et al. Distance regularized two level sets for segmentation of left and right ventricles from cine-MRI[J]. Magnetic Resonance Imaging, 2016, 34(5):699-706.
[31] WIJNTJES M W A, SATO A, HAYWARD V, et al. Local surface orientation dominates haptic curvature discrimination[J]. IEEE Transactions on Haptics, 2009, 2(2):94-102.
[32] BERNARD O, LALANDE A, ZOTTI C, et al. Deep learning techniques for automatic MRI cardiac multi-structures segmentation and diagnosis: is the problem solved?[J]. IEEE Transactions on Medical Imaging, 2018, 37(11):2514-2525.
[33] HUTTENLOCHER D P, KLANDERMAN G A, RUCKLIDGE W J. Comparing images using the Hausdorff distance[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1993, 15(9):850-863.
[34] ISENSEE F, JAEGER P F, FULL P M, et al. Automatic cardiac disease assessment on cine-MRI via time-series segmentation and domain specific features[M] //POP M, SERMESANT M, JODOIN P M, et al. Statistical Atlases and Computational Models of the Heart. ACDC and MMWHS Challenges: Vol. 10663. Cham: Springer, 2018:120-129.
[35] KHENED M, ALEX V, KRISHNAMURTHI G. Densely connected fully convolutional network for short-axis cardiac cine MR image segmentation and heart diagnosis using random forest[M] //POP M, SERMESANT M, JODOIN P M, et al. Statistical Atlases and Computational Models of the Heart: ACDC and MMWHS Challenges. Cham: Springer, 2018:140-151.
[36] ZOTTI C, LUO Z, LALANDE A, et al. Convolutional neural network with shape prior applied to cardiac MRI segmentation[J]. IEEE Journal of Biomedical and Health Informatics, 2019, 23(3):1119-1128.
[37] BAUMGARTNER C F, KOCH L M, POLLEFEYS M, et al. An exploration of 2D and 3D deep learning techniques for cardiac MR image segmentation[M] //POP M, SERMESANT M, JODOIN P M, et al. Statistical Atlases and Computational Models of the Heart: ACDC and MMWHS Challenges. Cham: Springer, 2018:111-119.
[38] WOLTERINK J M, LEINER T, VIERGEVER M A, et al. Automatic segmentation and disease classification using cardiac cine MR images[M] //POP M, SERMESANT M, JODOIN P M, et al. Statistical Atlases and Computational Models of the Heart: ACDC and MMWHS Challenges. Cham: Springer, 2018:101-110.
[39] PAINCHAUD N, SKANDARANI Y, JUDGE T, et al. Cardiac segmentation with strong anatomical guarantees[J]. IEEE Transactions on Medical Imaging, 2020, 39(11):3703-3713.
[40] ROHÉ M M, SERMESANT M, PENNEC X. Automatic multi-atlas segmentation of myocardium with SVF-net[M] //POP M, SERMESANT M, JODOIN P M, et al. Statistical Atlases and Computational Models of the Heart: ACDC and MMWHS Challenges. Cham: Springer, 2018:170-177.
[41] PATRAVALI J, JAIN S, CHILAMKURTHY S. 2D-3D fully convolutional neural networks for cardiac MR segmentation[M] //POP M, SERMESANT M, JODOIN P M, et al. Statistical Atlases and Computational Models of the Heart: ACDC and MMWHS Challenges. Cham: Springer, 2018:130-139.
[42] GRINIAS E, TZIRITAS G. Fast fully-automatic cardiac segmentation in MRI using MRF model optimization, substructures tracking and B-spline smoothing[M] //POP M, SERMESANT M, JODOIN P M, at al. Statistical Atlases and Computational Models of the Heart: ACDC and MMWHS Challenges. Cham: Springer, 2018:91-100.
[43] YANG Xin, BIAN Cheng, YU Lequan, et al. Class-balanced deep neural network for automatic ventricular structure segmentation[M] //POP M, SERMESANT M, JODOIN P M, et al. Statistical Atlases and Computational Models of the Heart: ACDC and MMWHS Challenges. Cham: Springer, 2018:152-160.
[44] LI C, XU C, KONWAR K M, et al. Fast distance preserving level set evolution for medical image segmentation[C] //2006 9th International Conference on Control, Automation, Robotics and Vision. Singapore: IEEE, 2006:1-7.
[45] GUO F, NG M, GOUBRAN M, et al. Improving cardiac MRI convolutional neural network segmentation on small training datasets and dataset shift: a continuous kernel cut approach[J]. Medical Image Analysis, 2020, 61:101636.
[46] MA Dongdong, LIAO Qingmin, CHEN Ziqin, et al. Adaptive local-fitting-based active contour model for medical image segmentation[J]. Signal Processing: Image Communication, 2019, 76:201-213.
[47] RAHMAN M M, MARCULESCU R. Multi-scale hierarchical vision transformer with cascaded attention decoding for medical image segmentation[EB/OL].(2023-03-29)[2024-08-26]. https://doi.org/10.48550/arXiv.2303.16892.
[48] RAHMAN M M, MARCULESCU R. G-CASCADE: efficient cascaded graph convolutional decoding for 2D medical image segmentation[C] //2024 IEEE/CVF Winter Conference on Applications of Computer Vision(WACV). Waikoloa: IEEE, 2024:7713-7722.
[49] ZHOU Hong-Yu, GUO Jiansen, ZHANG Yinghao, et al. nnFormer: volumetric medical image segmentation via a 3D transformer[J]. IEEE Transactions on Image Processing, 2023, 32:4036-4045.

多维度评价

Viewed

Full text

Abstract

Cited

Shared

Discussed