Causality-Inspired Single-Source Domain Generalization for Medical Image Segmentation
IEEE Transactions on Medical Imaging · 2023-04 · 42 (4) · 1095-1106
DOI: 10.1109/tmi.2022.3224067 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Medical Imaging, 2023-04. Verfasst von Cheng Ouyang und 6 weiteren Autor:innen. Hauptthemen laut MEDLINE: Pelvis, Prostate. Die Arbeit wurde bisher 255-mal zitiert (Crossref) und nennt 73 Literaturangaben. Gefördert von: Engineering and Physical Sciences Research Council (EPSRC) Programme, EPSRC, Innovate U.K.
Abstract (Originalsprache)
Deep learning models usually suffer from the domain shift issue, where models trained on one source domain do not generalize well to other unseen domains. In this work, we investigate the single-source domain generalization problem: training a deep network that is robust to unseen domains, under the condition that training data are only available from one source domain, which is common in medical imaging applications. We tackle this problem in the context of cross-domain medical image segmentation. In this scenario, domain shifts are mainly caused by different acquisition processes. We propose a simple causality-inspired data augmentation approach to expose a segmentation model to synthesized domain-shifted training examples. Specifically, 1) to make the deep model robust to discrepancies in image intensities and textures, we employ a family of randomly-weighted shallow networks. They augment training images using diverse appearance transformations. 2) Further we show that spurious correlations among objects in an image are detrimental to domain robustness. These correlations might be taken by the network as domain-specific clues for making predictions, and they may break on unseen domains. We remove these spurious correlations via causal intervention. This is achieved by resampling the appearances of potentially correlated objects independently. The proposed approach is validated on three cross-domain segmentation scenarios: cross-modality (CT-MRI) abdominal image segmentation, cross-sequence (bSSFP-LGE) cardiac MRI segmentation, and cross-site prostate MRI segmentation. The proposed approach yields consistent performance gains compared with competitive methods when tested on unseen domains.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Förderung
- Engineering and Physical Sciences Research Council (EPSRC) Programme
- EPSRC
- Innovate U.K
- European Research Council
Zitieren
APA
Ouyang, C., Chen, C., Li, S., Li, Z., Qin, C., Bai, W., & Rueckert, D. (2023). Causality-Inspired Single-Source Domain Generalization for Medical Image Segmentation. IEEE Transactions on Medical Imaging, 42(4), 1095-1106. https://doi.org/10.1109/tmi.2022.3224067
Vancouver
Ouyang C, Chen C, Li S, Li Z, Qin C, Bai W, et al. Causality-Inspired Single-Source Domain Generalization for Medical Image Segmentation. IEEE Transactions on Medical Imaging. 2023;42(4):1095-1106. doi:10.1109/tmi.2022.3224067
BibTeX
@article{Ouyang2023,
author = {Ouyang, Cheng and Chen, Chen and Li, Surui and Li, Zeju and Qin, Chen and Bai, Wenjia and Rueckert, Daniel},
title = {Causality-Inspired Single-Source Domain Generalization for Medical Image Segmentation},
journal = {IEEE Transactions on Medical Imaging},
year = {2023},
volume = {42},
number = {4},
pages = {1095--1106},
doi = {10.1109/tmi.2022.3224067}
}
EUVIMED