大规模预训练用于改进基于深度学习的弥散加权成像几何畸变校正
Large-Scale Pretraining for Improving Deep Learning-Based Geometric Distortion Correction of Diffusion-Weighted Imaging
浏览论文内容
中文总结 AI 辅助
本研究通过大规模预训练策略改进基于深度学习的弥散加权成像几何畸变校正，发现预训练模型优于基线，但跨域迁移存在挑战，配准至标准空间可提升性能。
中文摘要 AI 辅助
弥散加权成像（DWI）在临床中广泛应用，但仍易受几何畸变影响。传统校正方法通常需要额外采集或依赖供应商特定解决方案，限制了其在高通量、资源受限环境中的可行性。本研究探讨大规模预训练策略能否改进基于深度学习的单相位编码DWI畸变校正。我们将任务表述为图像重建，并比较了未预训练基线模型与自监督预训练模型及生成式预训练模型的性能，评估采用定量图像相似性指标和定性专家评估。表现最佳的模型进一步在低中等收入国家（LMIC）环境下存在采集偏移的数据上测试了可迁移性。预训练模型优于未预训练基线，其中cWDM在定量和定性评估中均取得最强结果。然而，应用于LMIC数据时暴露出迁移挑战，包括对比度改变和对T1加权解剖结构的过度依赖。将图像配准到共同标准空间改善了预测，表明协调的预处理可能增强跨域部署能力。
英文摘要
Diffusion-weighted imaging (DWI) is widely used in clinical settings but remains vulnerable to geometric distortion. Conventional correction methods often require additional acquisitions or vendor-specific solutions, limiting their feasibility in high-throughput, resource-constrained settings. This study investigates whether large-scale pretraining strategies can improve deep learning-based distortion correction for single-phase-encoding DWI. We formulate the task as image reconstruction, and compare a non-pretrained baseline against a self-supervised and a generative pretrained model, evaluated using both quantitative image-similarity metrics and qualitative expert assessment. The best-performing model was further tested for transferability on data collected in an LMIC setting with acquisition shift. Pretrained models outperformed the non-pretrained baseline, with cWDM achieving the strongest results across both quantitative and qualitative evaluation. However, application to LMIC data revealed transferability challenges, including contrast alteration and over-reliance on T1-weighted anatomical structure. Registering images to a common standard space improved predictions, suggesting that harmonized preprocessing may enhance cross-domain deployment.
发表机构
- Nepal Applied Mathematics and Informatics Institute for Research (NAAMII)（尼泊尔应用数学与信息学研究所）
- University of Exeter（埃克塞特大学）
- Institute of Engineering, Purwanchal Campus（工程学院普尔万查尔校区）
- Institute of Engineering, Pulchowk Campus（工程学院普尔乔克校区）
- Gandaki Medical College Teaching Hospital and Research Center（甘达基医学院教学医院与研究中心）
- Madan Bhandari University of Science and Technology（马丹班达里科技大学）
- Medical Artificial Intelligence Laboratory（医学人工智能实验室）
- University of British Columbia（不列颠哥伦比亚大学）
- Montreal Neurological Institute, McGill University（麦吉尔大学蒙特利尔神经研究所）
- McGill University（麦吉尔大学）
- University of Lagos（拉各斯大学）
机构由 AI 辅助整理，请以论文原文为准。
