arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2601.19939cs.LGcs.CV
收藏

oculomix:基于视网膜的系统性疾病的分层采样

oculomix: Hierarchical Sampling for Retinal-Based Systemic Disease Prediction

Institute of Ophthalmology, University College London, United Kingdom(伦敦大学学院眼科研究所);NIHR Moorfields Biomedical Research Centre, London, United Kingdom(英国 Moorfields 生物医学研究中心);University of Cologne, Faculty of Medicine, Department of Ophthalmology, Germany(科隆大学医学院眼科系);Hanyang University Guri Hospital, Guri City, South Korea(翰阳大学Guri医院)

Hyunmin Kim, Yukun Zhou, Rahul A. Jonas, Lie Ju, Sunjin Hwang, Pearse A. Keane, Siegfried K. Wagner

AI总结:

Oculomix通过分层采样策略提升视网膜图像在预测系统性疾病中的性能,优于传统混合样本增强方法。

摘要:

视网膜医学(oculomics)的概念,即通过视网膜成像预测系统性疾病,如心血管疾病和痴呆症,已因基于变压器的基础模型(如RETFound)的数据效率而迅速发展。图像层面的混合样本数据增强,如CutMix和MixUp,常用于训练变压器,但这些技术会扰动患者特定属性,如医疗共病和临床因素,因为它们仅考虑图像和标签。为了解决这一限制,我们提出了一种混合样本增强的分层采样策略,即Oculomix。我们的方法基于两个临床先验。首先(检查层面),同一患者在同一时间点获取的图像共享相同的属性。其次(患者层面),同一患者在不同时间点获取的图像具有软时间趋势,因为患病率通常随时间增加。受这些先验的指导,我们的方法将混合空间限制在患者和检查层面,以更好地保留患者特定特征,并利用其层次关系。所提出的方法在ViT模型上进行了验证,用于对大规模种族多样化的群体(Alzeye)进行五年主要不良心血管事件(MACE)的预测。我们显示,Oculomix在AUROC上比图像层面的CutMix和MixUp高出多达3%,证明了所提出方法在视网膜医学中的必要性和价值。

英文摘要:

Oculomics - the concept of predicting systemic diseases, such as cardiovascular disease and dementia, through retinal imaging - has advanced rapidly due to the data efficiency of transformer-based foundation models like RETFound. Image-level mixed sample data augmentations, such as CutMix and MixUp, are frequently used for training transformers, yet these techniques perturb patient-specific attributes, such as medical comorbidity and clinical factors, since they only account for images and labels. To address this limitation, we propose a hierarchical sampling strategy, Oculomix, for mixed sample augmentations. Our method is based on two clinical priors. First (exam level), images acquired from the same patient at the same time point share the same attributes. Second (patient level), images acquired from the same patient at different time points have a soft temporal trend, as morbidity generally increases over time. Guided by these priors, our method constrains the mixing space to the patient and exam levels to better preserve patient-specific characteristics and leverages their hierarchical relationships. The proposed method is validated using ViT models on a five-year prediction of major adverse cardiovascular events (MACE) in a large ethnically diverse population (Alzeye). We show that Oculomix consistently outperforms image-level CutMix and MixUp by up to 3% in AUROC, demonstrating the necessity and value of the proposed method in oculomics.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑