摘要
研究虹膜识别问题,由于在图像采集过程中,噪声影响准确性,为提高虹膜识别准确率,针对一维Log-Gabor滤波特征提取方法丢失二维信息的缺陷,提出一种改进Log-Gabor滤波的虹膜识别算法。首先采用小波变换对虹膜图像进行分解,获取虹膜图像低频子带信息,然后采用不同方向尺度的2维Log-Gabor滤波器组提取虹膜纹理特征,最后采用支持向量机对生成的虹膜特征码进行匹配。采用CASIA和UBIRIS虹膜库对算法性能进行测试,测试结果表明,改进Log-Gabor滤波的虹膜识别算法提高了虹膜识别准确率,加快识别速度,更加适合于实时虹膜识别,为虹膜准确识别提供了依据。
In order to improve the accuracy of iris recognition,this paper proposed an improved Log-Gabor filter algorithm for iris recognition.Wavelet transform was adopted in the iris image decomposition to capture the low frequency sub-band information of iris images.Then the 2-dimensional Log-Gabor filters with a different direction was used to extract iris texture features.Finally,support vector machine was used to generate the iris code matching.The algorithm performances were tested with CASIA and UBIRIS iris database,and the test results show that,compared with the traditional iris recognition algorithm,this algorithm can improve iris recognition accurate rate,accelerate the recognition speed,and is more suitable for real-time iris recognition.
出处
《计算机仿真》
CSCD
北大核心
2012年第5期283-286,共4页
Computer Simulation
关键词
虹膜识别
特征提取
支持向量机
虹膜分割
Iris recognition
Feature extraction
Support vector machine（SVM）
Iris segmentation
