基于形状注意力级联的CMR cine序列左心室运动估计方法

doi:10.3969/j.issn.0258-8021.2026.03.002

Abstract
Figure/Table
References
Related Citation (15)

Download: PDF (6414 KB) HTML (1 KB)
Export: BibTeX | EndNote (RIS)

Abstract Cardiac magnetic resonance cine (CMR cine) is a key non-invasive technique for assessing cardiac motion. However, its images often suffer from motion artifacts and improper imaging parameters, leading to blurring that compromises the boundary localization accuracy of existing left ventricular motion estimation methods, thereby affecting the reliability of cardiac function measurements. To address this issue, this study proposed a shape-attention cascade structure, consisting of a basic module and a sequence module, both of which was connected in cascade. The basic module enhanced boundary feature perception through a shape flow layer, enabling pixel-wise error correction of the initial prediction. The sequence module ensured motion estimation coherence using a bidirectional motion attention layer. In addition, this study introduced a novel data augmentation method targeting boundary noise and designs a weighted Dice loss function that integrates pixel confidence and frame confidence. Experimental results demonstratd that the proposed method significantly improved the estimation accuracy for the left ventricular motion. On the ACDC dataset, the Hausdorff distance of the proposed method is 3.83 mm, outperforming the baseline model at 7.51 mm; on the private dataset, the proposed method achieves 4.39 mm, outperforming the baseline model at 8.68 mm; on EchoNet-Dynamic, the proposed method achieves 4.86 mm, outperforming the baseline method at 5.87 mm. The above differences were all statistically significant as assessed by the Kruskal-Wallis test(P<0.05). This study provides an effective method for accurate cardiac motion estimation, with positive implications for advancing the clinical translation of automated cardiac function assessment.

Key words： cardiac magnetic resonance cine imaging motion estimation shape attention motion attention cascading method

Received: 24 March 2025

PACS:

R318

Corresponding Authors: ^*E-mail:zhangnan@ccmu.edu.cn

	Service

	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	Wang Yu
	Sun Zheng
	Zhang Nan

Cite this article:

Wang Yu,Sun Zheng,Zhang Nan. A Left Ventricular Motion Estimation Method for CMR Cine Sequences Based on Shape-AttentionCascading[J]. Chinese Journal of Biomedical Engineering, 2026, 45(3): 267-275.

URL:

http://cjbme.csbme.org/EN/10.3969/j.issn.0258-8021.2026.03.002 OR http://cjbme.csbme.org/EN/Y2026/V45/I3/267

[1] Qin C, Bai W, Schlemper J, et al. Joint learning of motion estimation and segmentation for cardiac MR image sequences [C]//Medical Image Computing and Computer Assisted Intervention. Cham: Springer, 2018: 472-480.
[2] Puyol-Antón E, Ruijsink B, Bai W, et al. Fully automated myocardial strain estimation from cine MRI using convolutional neural networks [C]//International Symposium on Biomedical Imaging. Piscataway: IEEE Computer Society, 2018: 1139-1143.
[3] Petitjean C, Rougon N, Cluzel P, et al. Assessment of myocardial function: a review of quantification methods and results using tagged MRI [J]. J Cardiovasc Magn Reson, 2005, 7(2): 501-516.
[4] Qin C, Wang S, Chen C, et al. Generative myocardial motion tracking via latent space exploration with biomechanics-informed prior [J]. Med Image Anal, 2023, 83: 102682.
[5] Morales MA, van den Boomen M, Nguyen C, et al. DeepStrain: a deep learning workflow for the automated characterization of cardiac mechanics [J]. Front Cardiovasc Med, 2021, 8:730316.
[6] Xue W, Li J, Hu Z, et al. Left ventricle quantification challenge: a comprehensive comparison and evaluation of segmentation and regression for mid-ventricular short-axis cardiac MR data [J]. IEEE J Biomed Health Inform, 2021, 25(9): 3541-3553.
[7] Vesal S, Gu M, Maier A, et al. Spatio-temporal multi-task learning for cardiac MRI left ventricle quantification [J]. IEEE J Biomed Health Inform, 2021, 25(7): 2698-2709.
[8] Uslu F, Varela M, Boniface G, et al. LA-Net: A multi-task deep network for the segmentation of the left atrium [J]. IEEE Trans Med Imaging, 2022, 41(2): 456-464.
[9] Lu P, Bai W, Rueckert D, et al. Dynamic spatio-temporal graph convolutional networks for cardiac motion analysis [C]//International Symposium on Biomedical Imaging. Piscataway: IEEE Computer Society, 2021: 122-125.
[10] Zheng Q, Delingette H, Ayache N. Explainable cardiac pathology classification on cine MRI with motion characterization by semi-supervised learning of apparent flow [J]. Med Image Anal, 2019, 56: 80-95.
[11] Zhang X, You C, Ahn S, et al. Learning correspondences of cardiac motion from images using biomechanics-informed modeling [C]//Medical Image Computing and Computer-Assisted Intervention - Statistical Atlases and Computational Models of the Heart: STACOM 2022 Workshop.. Cham: Springer, 2022: 13-25.
[12] Meng Q, Qin C, Bai W, et al. MulViMotion: shape-aware 3D myocardial motion tracking from multi-view cardiac MRI [J]. IEEE Trans Med Imaging, 2022, 41(8): 1961-1974.
[13] 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 Trans Med Imaging, 2018, 37(11): 2514-2525.
[14] Ouyang D, He B, Ghorbani A, et al. Video-based AI for beat-to-beat assessment of cardiac function [J]. Nature, 2020, 580(7802): 252-256.
[15] Sun J, Darbehani F, Zaidi M, et al. SAUNet: shape attentive U-Net for interpretable medical image segmentation [C]//Medical Image Computing and Computer-Assisted Intervention. Cham: Springer, 2020: 797-806.
[16] Takikawa T, Acuna D, Jampani V, et al. Gated-SCNN: gated shape CNNs for semantic segmentation [C]//International Conference on Computer Vision. Piscataway: IEEE Computer Society, 2019: 5229-5238.
[17] Zhang Z, Tian C, Bai H X, et al. Discriminative error prediction network for semi-supervised colon gland segmentation [J]. Med Image Anal, 2022, 79: 102458.
[18] Uslu F, Varela M. SA-net: a sequence aware network for the segmentation of the left atrium in cine MRI datasets [C]//International Symposium on Biomedical Imaging. Piscataway: IEEE Computer Society, 2021: 766-769.
[19] Zakeri A, Hokmabadi A, Ravikumar N, et al. A probabilistic deep motion model for unsupervised cardiac shape anomaly assessment [J]. Med Image Anal, 2022, 75:102276.
[20] Luo Y, Liu G, Guo Y, et al. Deep neural networks learn meta-structures from noisy labels in semantic segmentation [C]//Proceedings of the AAAI Conference on Artificial Intelligence. Palo Alto: AAAI Press, 2022: 1908-1916.
[21] Zhu H, Shi J, Wu J. Pick-and-Learn: automatic quality evaluation for noisy-labeled image segmentation [C]//Medical Image Computing and Computer Assisted Intervention. Cham: Springer, 2019: 576-584.
[22] Bertels J, Robben D, Vandermeulen D, et al. Theoretical analysis and experimental validation of volume bias of soft Dice optimized segmentation maps in the context of inherent uncertainty [J]. Med Image Anal, 2021, 67: 101833.
[23] Qin C, Wang S, Chen C, et al. Biomechanics-informed neural networks for myocardial motion tracking in MRI [C]//Medical Image Computing and Computer-Assisted Intervention. Cham: Springer, 2020: 296-306.
[24] Ye M, Kanski M, Yang D, et al. DeepTag: an unsupervised deep learning method for motion tracking on cardiac tagging magnetic resonance images [C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Computer Society, 2021: 16154-16163.
[25] Pavlidis T, Liow YT. Integrating region growing and edge detection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1990, 12(3): 225-233.
[26] Killoran JH, Gerbaudo VH, Mamede M, et al. Motion artifacts occurring at the lung/diaphragm interface using 4D CT attenuation correction of 4D PET scans [J]. J Appl Clin Med Phys, 2011, 12(4): 261-274.
[27] Sun Y, Zhang J, Li H, et al. Motion correction for liver DCE-MRI with time-intensity curve constraint [C]//Medical Image Computing and Computer Assisted Intervention. Cham: Springer, 2021: 733-742.
[28] Li L, Ding W, Huang L, et al. Multi-modality cardiac image computing: a survey [J]. Med Image Anal, 2023, 88: 102869.
[29] Yin FCP, Chan CCH, Judd RM. Compressibility of perfused passive myocardium[J]. American Journal of Physiology-Heart and Circulatory Physiology, 1996, 271(5): H1864-H1870.
[30] Fung YC. Foundations of Solid Mechanics[M]. Englewood Cliffs: Prentice-Hall, 1965.