基于FREAK算符及双向最近汉明距的非理想虹膜识别算法
NON-IDEAL IRIS RECOGNTION ALGORITHM BASED ON FREAK OPERATOR AND BIDIRECTIONAL CLOSEST HAMMING DISTANCE
季毕胜 1叶学义 1廖奕艺 1邹茹梦1
作者信息
- 1. 杭州电子科技大学通信工程学院 浙江杭州 310018
- 折叠
摘要
针对低约束条件的非理想虹膜识别性能的显著下降,以及尺度不变特征变换(SIFT)和加速稳健特征(SURF)方法实时性不高的问题,提出FREAK算符配合双向最近汉明距的识别算法.通过高斯核构建多尺度特征点检测算子提取稳健的特征点集合,并引入FREAK算符改进SIFT算符,以提高特征点的表征和匹配速度;利用双向最近汉明距的匹配策略增强特征匹配对的稳定性,以降低非同源虹膜的匹配点数.实验结果表明,识别等错误率和正确识别率均有改善且单次验证时间均在0.3 s左右.该方法能够有效应对非理想虹膜类内纹理质量的变化,与SIFT与SURF算法相比具有更好的实时性.
Abstract
Aimed at the significant degradation of non-ideal iris recognition performance under low-constraint conditions,and the low real-time performance of scale invariant feature transform(SIFT)and speeded up robust feature(SURF)methods,a recognition algorithm based on FREAK operator and bidirectional nearest Hamming distance is proposed.A Gaussian kernel was used to construct a multi-scale feature point detection operator to extract a robust set of feature points,and the FREAK operator was introduced to improve the SIFT operator to enhance the characterization and matching speed of feature points.The bidirectional closest Hamming distance matching strategy was used to enhance the stability of feature matching pairs so as to reduce the number of mismatched points of non-homologous irises.The experimental results show that the recognition error rate and the correct recognition rate are improved and the recognition time for a single verification is about 0.3s.The proposed method can effectively cope with the texture quality changes in non-ideal iris classes and has better real-time ability compared with SIFT and SURF algorithms.
关键词
非理想虹膜识别/多尺度检测/FREAK算符/双向汉明距/实时性Key words
Non-ideal iris recognition/Multi-scale detection/FREAK operator/Bidirectional nearest hamming distance/Real-time ability引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据