Royal Eye Infirmary, Plymouth, UK.
Eye (Lond). 2010 Jun;24(6):1006-10. doi: 10.1038/eye.2009.275. Epub 2009 Nov 13.
The purpose of this study was to investigate the effect of cataract surgery and pupil dilation on iris pattern recognition for personal authentication.
Prospective non-comparative cohort study. Images of 15 subjects were captured before (enrolment), and 5, 10, and 15 min after instillation of mydriatics before routine cataract surgery. After cataract surgery, images were captured 2 weeks thereafter. Enrolled and test images (after pupillary dilation and after cataract surgery) were segmented to extract the iris. This was then unwrapped onto a rectangular format for normalization and a novel method using the Discrete Cosine Transform was applied to encode the image into binary bits. The numerical difference between two iris codes (Hamming distance, HD) was calculated. The HD between identification and enrolment codes was used as a score and was compared with a confidence threshold for specific equipment, giving a match or non-match result. The Correct Recognition Rate (CRR) and Equal Error Rates (EERs) were calculated to analyse overall system performance.
After cataract surgery, perfect identification and verification was achieved, with zero false acceptance rate, zero false rejection rate, and zero EER. After pupillary dilation, non-elastic deformation occurs and a CRR of 86.67% and EER of 9.33% were obtained.
Conventional circle-based localization methods are inadequate. Matching reliability decreases considerably with increase in pupillary dilation. Cataract surgery has no effect on iris pattern recognition, whereas pupil dilation may be used to defeat an iris-based authentication system.
本研究旨在探讨白内障手术和瞳孔扩张对虹膜模式识别进行个人认证的影响。
前瞻性非对照队列研究。15 名受试者在（入组）前、散瞳后 5、10 和 15 分钟以及常规白内障手术前采集图像。白内障手术后 2 周后采集图像。入组和测试图像（瞳孔扩张后和白内障手术后）被分割以提取虹膜。然后将其展开到矩形格式以进行归一化，并应用一种使用离散余弦变换的新方法将图像编码为二进制位。计算两个虹膜码之间的数字差异（汉明距离，HD）。将识别和入组代码之间的 HD 用作分数，并与特定设备的置信阈值进行比较，给出匹配或不匹配的结果。计算正确识别率（CRR）和等错误率（EER）以分析整体系统性能。
白内障手术后，实现了完美的识别和验证，零错误接受率、零错误拒绝率和零 EER。瞳孔扩张后，会发生非弹性变形，获得 86.67%的 CRR 和 9.33%的 EER。
传统的基于圆的定位方法是不够的。随着瞳孔扩张的增加，匹配可靠性大大降低。白内障手术对虹膜模式识别没有影响，而瞳孔扩张可能被用于破坏基于虹膜的认证系统。
