摘要
提出了一种基于信息熵(information entropy)的GLBP掌纹识别算法(EGLBP),首次将该算法运用到掌纹中。同时,为了提高识别精度、降低算法复杂度,引入信息熵来度量掌纹所含的信息量,熵越大,所含信息量越多。首先对图像进行Gabor变换,分别计算变换后图像的信息熵,去除熵较小的几幅图像;然后对剩余的图像使用分块思想,对每块进行LBP特征提取,并联融合所有特征;最后使用卡方距离对掌纹所属类别进行判定。经过PolyU掌纹中心区域图像的验证,与传统掌纹识别算法相比,EGLBP算法识别率达到99.89%,识别时间为113.9ms,具有有效性和优越性。
A novel palmprint recognition algorithm based on information entropy under GLBP （EGLBP） was proposed in this paper,and it was first introduced into palmprint recognition. At the same time,in order to improve the recogni- tion accuracy and reduce the complexity of the algorithm, information entropy was used to measure the information of palmprint. The bigger the entropy is, the more information is contained. Firstly, the images were decomposed though Gabor transform,and their information entropies were calculated. Then several small images were removed. Secondly, LBP feature extraction for each block was processed using the block idea, and the features were parallelly fused. Finally, chi-square distance was used to determine palmprint category. After PolyU palmprint database＇s verification of the re- gion of interest, compared with the traditional palmprint recognition algorithm, EGLBP algorithm has a recognition rate of 99.89% and a recognition time of 113.9ms,demonstrating its superiority and effectiveness.
出处
《计算机科学》
CSCD
北大核心
2014年第8期293-296,共4页
Computer Science
基金
国家自然基金(61170106)
山东省科技发展计划项目(2012YD01058)资助
关键词
掌纹识别
GABOR变换
局部二值模式
信息熵
分块
EGLBP算法
Palmprint recognition, Gabor trans{orm, Local binary pattern, Information entropy, Block thought, EGLBPalgorithm
