EUVIMEDEUVIMED entdecken
Crossref

DCL: Dynamic Causal Learning for Cross-Modality Cardiac Image Segmentation

Saidi Guo, Xinlong Liu, Qixin Lin, Weijie Cai, Guohua Zhao, Mingyi Wu, Qiujie Lv, Laurence T Yang

IEEE Transactions on Image Processing · 2026 · 35 · 3099-3112

DOI: 10.1109/tip.2026.3673293 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Image Processing, 2026. Verfasst von Saidi Guo und 7 weiteren Autor:innen. Hauptthemen laut MEDLINE: Heart, Machine Learning, Cardiac Imaging Techniques, Image Processing, Computer-Assisted. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 60 Literaturangaben. Gefördert von: Natural Science Foundation of Henan Province, Henan Provincial Science and Technology Research Project, Open Project of State Key Laboratory of Phytochemistry and Natural Medicines.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate cross-modality cardiac image segmentation is essential for effectively diagnosing and treating heart disease. Different imaging modalities help to determine suitable pre-procedure planning. However, most methods face the difficulty of spatial-temporal confounding, where the anatomy element and modality element of cardiac images are intertwined across both spatial and temporal dimensions. It is derived from the imaging diversity and structure diversity of cardiac images. The spatial-temporal confounding hinders knowledge transfer between cardiac images on different modalities. In this paper, we propose a novel dynamic causal learning (DCL) to solve spatial-temporal confounding. The DCL explores multi-dimensional causal intervention to consider not only the causal relationship between images and labels, but also the causality in time dimension and space dimension. It integrates historical optimal interventions and facilitates the transfer of this knowledge across temporal contexts. In addition, the DCL utilizes the diffusion mechanism to further ensure that the extracted anatomy element remains causal invariant, improving model performance across multiple imaging modalities. Extensive experiments on cross-modality cardiac images (MR, CT, and US) demonstrate the effectiveness of the DCL (mean Dice = 0.951), outperforming other advanced segmentation methods. DCL is freely accessible at https://github.com/asdww0721ww/DCL.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

HeartMachine LearningCardiac Imaging TechniquesImage Processing, Computer-AssistedHumansAlgorithmsTomography, X-Ray ComputedMultimodal Imaging

Förderung

  • Natural Science Foundation of Henan Province
  • Henan Provincial Science and Technology Research Project
  • Open Project of State Key Laboratory of Phytochemistry and Natural Medicines
  • National Youth Talent Support Program, Strategic Priority Research Program of Chinese Academy of Sciences
  • Strategic Priority Research Program of Kunming Institute of Botany, Chinese Academy of Sciences

Zitieren

APA

Guo, S., Liu, X., Lin, Q., Cai, W., Zhao, G., Wu, M., Lv, Q., & Yang, L. T. (2026). DCL: Dynamic Causal Learning for Cross-Modality Cardiac Image Segmentation. IEEE Transactions on Image Processing, 35, 3099-3112. https://doi.org/10.1109/tip.2026.3673293

Vancouver

Guo S, Liu X, Lin Q, Cai W, Zhao G, Wu M, et al. DCL: Dynamic Causal Learning for Cross-Modality Cardiac Image Segmentation. IEEE Transactions on Image Processing. 2026;35:3099-3112. doi:10.1109/tip.2026.3673293

BibTeX
@article{Guo2026,
  author = {Guo, Saidi and Liu, Xinlong and Lin, Qixin and Cai, Weijie and Zhao, Guohua and Wu, Mingyi and Lv, Qiujie and Yang, Laurence T},
  title = {DCL: Dynamic Causal Learning for Cross-Modality Cardiac Image Segmentation},
  journal = {IEEE Transactions on Image Processing},
  year = {2026},
  volume = {35},
  pages = {3099--3112},
  doi = {10.1109/tip.2026.3673293}
}
Kostenlos starten →