发表机构
Erasmus University Medical Center; Radboud University Medical Center; Institute of Molecular and Clinical Ophthalmology, University of Basel; University of Lausanne; Swiss Institute of Bioinformatics; University of Cape Town（伊拉斯姆斯大学医学中心; 拉德堡德大学医学中心; 巴塞尔大学分子与临床眼科学研究所; 洛桑大学; 瑞士生物信息学研究所; 开普敦大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究提出结合ConvNeXtV2与ArcFace、三元组损失的512维度量学习编码器，实现跨设备、长随访的鲁棒视网膜生物特征，可高效准确验证和检索患者身份，保障医疗记录完整性。
AI 中文摘要
患者身份错误会损害纵向医疗记录、研究数据库及下游临床决策。本文提出一种视网膜生物特征系统，用于从彩色眼底图像中验证声称的身份并检索正确身份。我们在鹿特丹研究（Rotterdam Study）的21851个患者眼身份的227004张图像上，训练了一个512维的度量学习编码器，该编码器结合了ConvNeXtV2骨干网络与ArcFace及三元组损失函数，这些图像涵盖多种成像设备和最长32.6年的随访时间。该系统在鹿特丹研究的保留数据上进行评估，并在英国生物银行（UK Biobank）和年龄相关性眼病研究（AREDS）上进行外部评估。评估前，我们使用该模型筛选身份不一致的图像并人工裁定标记的图像，发现鹿特丹研究图像中0.588%存在错误分配，英国生物银行图像中为0.259%，AREDS图像中为0.164%。在移除近重复图像后的仅回顾性验证中，该系统在鹿特丹研究、英国生物银行和AREDS中分别达到0.9998、0.9997和0.9998的AUROC。对于仅使用先前获取图像的身份检索，在平均包含4436至8510个身份的图库中，Recall@1分别为99.7%、97.2%和97.6%；至少98.6%的情况下正确身份出现在前五名结果中。该系统在跨成像设备和长随访间隔时表现稳健，而图像质量较低和视网膜视野不一致是大多数失败的原因。这些发现证实视网膜解剖结构是一种持久的生物特征信号，可用于保障纵向成像记录的完整性。
英文摘要
Patient identity errors can compromise longitudinal medical records, research databases, and downstream clinical decisions. We present a retinal biometric system for verifying claimed identities and retrieving the correct identity from color fundus images. We trained a 512-dimensional metric-learning encoder combining a ConvNeXtV2 backbone with ArcFace and triplet losses on 227,004 images from 21,851 patient-eye identities in the Rotterdam Study, spanning multiple imaging devices and up to 32.6 years of follow-up. The system was evaluated on held-out Rotterdam Study data and externally on the UK Biobank and Age-Related Eye Disease Study (AREDS). Before evaluation, we used the model to screen for identity inconsistencies and manually adjudicated flagged images, identifying incorrect assignments in 0.588% of Rotterdam Study images, 0.259% of UK Biobank images, and 0.164% of AREDS images. In retrospective-only verification after removing near-duplicate images, the system achieved AUROCs of 0.9998, 0.9997, and 0.9998 in the Rotterdam Study, UK Biobank, and AREDS, respectively. For identity retrieval using only previously acquired images, Recall@1 was 99.7%, 97.2%, and 97.6%, respectively, from galleries averaging 4436-8510 identities; the correct identity appeared among the top five results in at least 98.6% of cases. Performance remained robust across imaging devices and long follow-up intervals, while lower image quality and inconsistent retinal fields accounted for most failures. These findings establish retinal anatomy as a durable biometric signal, useful for safeguarding the integrity of longitudinal imaging records.
