Artificial Intelligence-Enabled Facial Privacy Protection for Ocular Diagnosis: Development and Validation Study.

IF 5.8 2区医学 Q1 HEALTH CARE SCIENCES & SERVICES

Journal of Medical Internet Research Pub Date : 2025-07-09 DOI:10.2196/66873

Haizhu Tan, Hongyu Chen, Zhenmao Wang, Mingguang He, Chiyu Wei, Lei Sun, Xueqin Wang, Danli Shi, Chengcheng Huang, Anping Guo

{"title":"Artificial Intelligence-Enabled Facial Privacy Protection for Ocular Diagnosis: Development and Validation Study.","authors":"Haizhu Tan, Hongyu Chen, Zhenmao Wang, Mingguang He, Chiyu Wei, Lei Sun, Xueqin Wang, Danli Shi, Chengcheng Huang, Anping Guo","doi":"10.2196/66873","DOIUrl":null,"url":null,"abstract":"Background: Facial biometric data, while valuable for clinical applications, poses substantial privacy and security risks.Objective: This paper aims to address the privacy and security concerns related to facial biometric data and support auxiliary diagnoses, in pursuit of which we developed Digital FaceDefender, an artificial intelligence-driven privacy safeguard solution.Methods: We constructed a diverse set of digitally synthesized Asian face avatars representing both sexes, spanning ages 5 to 85 years in 10-year increments, using 70,000 facial images and 13,061 Asian face images. Landmark data were separately extracted from both patient and avatar images. Affine transformations ensured spatial alignment, followed by color correction and Gaussian blur to enhance fusion quality. For auxiliary diagnosis, we established 95% CIs for pixel distances within the eye region on a cohort of 1163 individuals, serving as diagnostic benchmarks. Reidentification risk was assessed using the ArcFace model, applied to 2500 images reconstructed via Detailed Expression Capture and Animation (DECA). Finally, Cohen Kappa analyses (n=114) was applied to assess agreement between diagnostic benchmarks and ophthalmologists' evaluations.Results: Compared to the DM method, Digital FaceDefender significantly reduced facial similarity scores (FDface vs raw images: 0.31; FLAME_FDface vs raw images: 0.09) and achieved zero Rank-1 accuracy in Pose #2-#3 and Pose #2-#5, with near-zero accuracy in Pose #4 (0.02) and Pose #5 (0.04), confirming lower reidentification risk. Cohen Kappa analysis demonstrated moderate agreement between our benchmarks and ophthalmologists' assessments for the left eye (κ=0.37) and right eye (κ=0.45; both P<.001), validating diagnostic reliability of the benchmarks. Furthermore, the user-friendly Digital FaceDefender platform has been established and is readily accessible for use.Conclusions: In summary, Digital FaceDefender effectively balances privacy protection and diagnostic use.","PeriodicalId":16337,"journal":{"name":"Journal of Medical Internet Research","volume":"27 ","pages":"e66873"},"PeriodicalIF":5.8000,"publicationDate":"2025-07-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12266301/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Medical Internet Research","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.2196/66873","RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"HEALTH CARE SCIENCES & SERVICES","Score":null,"Total":0}

引用次数: 0

Abstract

Background: Facial biometric data, while valuable for clinical applications, poses substantial privacy and security risks.

Objective: This paper aims to address the privacy and security concerns related to facial biometric data and support auxiliary diagnoses, in pursuit of which we developed Digital FaceDefender, an artificial intelligence-driven privacy safeguard solution.

Methods: We constructed a diverse set of digitally synthesized Asian face avatars representing both sexes, spanning ages 5 to 85 years in 10-year increments, using 70,000 facial images and 13,061 Asian face images. Landmark data were separately extracted from both patient and avatar images. Affine transformations ensured spatial alignment, followed by color correction and Gaussian blur to enhance fusion quality. For auxiliary diagnosis, we established 95% CIs for pixel distances within the eye region on a cohort of 1163 individuals, serving as diagnostic benchmarks. Reidentification risk was assessed using the ArcFace model, applied to 2500 images reconstructed via Detailed Expression Capture and Animation (DECA). Finally, Cohen Kappa analyses (n=114) was applied to assess agreement between diagnostic benchmarks and ophthalmologists' evaluations.

Results: Compared to the DM method, Digital FaceDefender significantly reduced facial similarity scores (FDface vs raw images: 0.31; FLAME_FDface vs raw images: 0.09) and achieved zero Rank-1 accuracy in Pose #2-#3 and Pose #2-#5, with near-zero accuracy in Pose #4 (0.02) and Pose #5 (0.04), confirming lower reidentification risk. Cohen Kappa analysis demonstrated moderate agreement between our benchmarks and ophthalmologists' assessments for the left eye (κ=0.37) and right eye (κ=0.45; both P<.001), validating diagnostic reliability of the benchmarks. Furthermore, the user-friendly Digital FaceDefender platform has been established and is readily accessible for use.

Conclusions: In summary, Digital FaceDefender effectively balances privacy protection and diagnostic use.

查看原文本刊更多论文

基于人工智能的眼部诊断面部隐私保护：开发与验证研究。

背景：面部生物识别数据虽然具有临床应用价值，但存在重大的隐私和安全风险。目的：为解决面部生物识别数据相关的隐私和安全问题，支持辅助诊断，我们开发了一种人工智能驱动的隐私保护解决方案Digital FaceDefender。方法：我们使用70000张面部图像和13061张亚洲面部图像，构建了一组不同的数字合成亚洲面部化身，代表男女，年龄从5岁到85岁，以10年为单位。地标性数据分别从患者和化身图像中提取。仿射变换确保空间对齐，其次是颜色校正和高斯模糊，以提高融合质量。为了辅助诊断，我们在1163个个体的队列中建立了95%的眼睛区域像素距离ci，作为诊断基准。使用ArcFace模型评估再识别风险，应用于2500张通过详细表情捕捉和动画（DECA）重建的图像。最后，采用Cohen Kappa分析（n=114）来评估诊断基准与眼科医生评估之间的一致性。结果：与DM方法相比，Digital FaceDefender显著降低了面部相似性得分(FDface vs raw图像：0.31；FLAME_FDface vs原始图像：0.09)，并且在Pose #2-#3和Pose #2-#5中实现了零Rank-1精度，在Pose #4（0.02）和Pose #5（0.04）中实现了接近零的精度，证实了较低的重新识别风险。Cohen Kappa分析显示，我们的基准与眼科医生对左眼（κ=0.37）和右眼(κ=0.45；结论：总之，Digital FaceDefender有效地平衡了隐私保护和诊断使用。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Journal of Medical Internet Research 医学-卫生保健

CiteScore

14.40

自引率

5.40%

发文量

654

审稿时长

1 months

期刊介绍： The Journal of Medical Internet Research (JMIR) is a highly respected publication in the field of health informatics and health services. With a founding date in 1999, JMIR has been a pioneer in the field for over two decades. As a leader in the industry, the journal focuses on digital health, data science, health informatics, and emerging technologies for health, medicine, and biomedical research. It is recognized as a top publication in these disciplines, ranking in the first quartile (Q1) by Impact Factor. Notably, JMIR holds the prestigious position of being ranked #1 on Google Scholar within the "Medical Informatics" discipline.