摘要
在自动虹膜识别系统的图像采集过程中,不符合标准的虹膜图像可能被系统拒绝登录(“注册失败”)。本文提出了“图像清晰度”、“内外偏心度”和“虹膜可见度”三个衡量图像质量的指标,实现了对虹膜图像进行实时质量评价的要求。实验结果表明,该质量评价方法反映了虹膜图像的真实质量水平,能够显著提高识别系统的整体性能。
During the image acquisition process of an automatic iris recognition system, unqualified iris images are probably rejected (Failure to Enroll). Three performance metrics, Image Definition, Inside and Outside Eccentricity, and Iris Visibility are introduced, which makes possible a real-time evaluation to the acquired images. The experiment results show that this evaluation system can reflect the real quality of images and thus significantly improve the overall performance of the iris recognition system.
出处
《中国体视学与图像分析》
2003年第2期108-113,共6页
Chinese Journal of Stereology and Image Analysis
关键词
图像质量评价
虹膜识别
图像识别
Image Quality Evaluation
Iris Recognition
Image Recognization
