摘要
为提高虹膜识别系统的识别率和识别时效性,设计了基于多通道Gabor滤波和二维主分量分析(2DPCA)的虹膜识别算法。利用多通道Gabor滤波器对采集的虹膜进行特征提取。由于得到的特征向量矩阵的维数通常较大,会影响特征匹配和虹膜识别的运算速度,因而需要对特征矩阵进行降维处理。2DPCA算法能有效地克服传统虹膜识别系统中的"维数危机"问题,在保留虹膜特征主分量的基础上,降低虹膜匹配运算量,提高虹膜系统的识别效率。虹膜识别采用差异度匹配法,通过阈值比较得到识别结果,对容量不等的各类虹膜图库均具有良好的适应性。实验中对容量为50的虹膜图库进行了算法测试,系统的最低识别率达到了88%,识别时间仅为传统非降维识别方法的一半。理论分析和实验结果表明,该算法对虹膜纹理的特征提取精度高,识别率高,识别速度快。
In order to improve the recognition rate and time efficiency of the iris recognition system, an iris recognition algorithm based on multi-channel Gabor filter and two-dimensional principal component analysis （2DPCA） is designed in this paper. The multi-channel Gabor which shares the same characteristics with the biological visual system has the ability of co-location in the airspace and frequency domain. The extraction of iris texture feature is high when using multi-channel Gabor filters to analysis the iris obtained. Since the arithmetic speed of feature matching and iris recognition are usually affected by the fact that the dimensions of the characteristic vector matrix are usually large, it is necessary to reduce the dimension of the feature matrix. The 2DPCA can effectively overcome the ＂dimension crisis＂ which is popularly encountered in traditional iris recognition systems, and it can reduce the iris matching computation and improve the efficiency of the iris identification system based on the retention of the main component of the iris texture. Difference matching method is utilized in iris recognition. Experimental results are carried out by comparing the threshold value obtained in the experiment. This method shares the characteristic that it can adjust every iris image database whose capacity range is not the same. In the experiment, the algorithm presented is tested based on an iris image database whose capacity is 50. The lowest recognition rate reaches 88%. Moreover, the recognition time is just about one half of the traditional methods. Both the theory analysis and experimental results have shown that the algorithm proposed is highly precise in feature extraction, and has a high recognition rate and recognition speed.
出处
《实验室研究与探索》
CAS
北大核心
2013年第6期21-24,120,共5页
Research and Exploration In Laboratory
基金
贵州省教育厅自然科技基金(20090028)
贵州大学人才引进基金(2009026)
贵州大学大学生创新性实验计划项目(2009053)
关键词
多通道Gabor
二维主分量分析
差异度匹配
虹膜识别
multi-channel Gabor
two-dimensional principal component analysis
difference matching
iris recognition
