基于Gabor小波和NBP算法的手掌静脉识别

发布时间：2018-03-21 20:54

本文选题：图像处理　切入点：手掌静脉识别　出处：《激光与光电子学进展》2017年05期 　论文类型：期刊论文

【摘要】：为了提取手掌静脉图像的纹理特征,并有效提高其识别率,在联合Gabor小波和近邻二值模式(NBP)的基础上提出了一种纹理特征提取方法。该方法利用静脉结构中血管粗细与延伸方向不同的特点,将掌静脉图像感兴趣区域与4尺度、4方向的Gabor小波卷积获得多个幅值特征,并在4个不同的尺度下分别求取均值,获得Gabor尺度均值模式(GSP),在每个GSP分块上使用NBP描述算子来提取局部邻域关系模式(GSPNBP)。然后将这些多尺度、多方向的GSPNBP分块区域的编码序列的总和作为掌静脉特征向量。最后通过求特征向量间汉明距离衡量静脉图像的相似程度来计算识别率,并在PolyU图库和自建图库中进行实验。实验结果显示,该算法获得的识别率最高可分别可达99.7935%和99.3965%,识别时间都在1s以内,有效增强了算法稳健性。
[Abstract]:In order to extract the texture feature of palm vein image and improve the recognition rate effectively, Based on the combination of Gabor wavelet and nearest neighbor binary mode, a texture feature extraction method is proposed, which makes use of the different direction of blood vessel thickness and extension in vein structure. The region of interest of palmar vein image is convoluted with Gabor wavelet in 4 scales and 4 directions, and the mean value is obtained under four different scales. The Gabor scale mean value model (GSPN) is obtained, and the NBP description operator is used on each GSP block to extract the local neighborhood relation pattern (GSPNBPN). The sum of coding sequences in the multi-directional GSPNBP block region is taken as the palmar vein feature vector. Finally, the recognition rate is calculated by calculating the hamming distance between the feature vectors to measure the similarity of the vein image. Experiments are carried out in the PolyU library and the self-built map library. The experimental results show that the recognition rate of the algorithm can reach 99.7935% and 99.3965%, respectively, and the recognition time is within 1 s, which effectively enhances the robustness of the algorithm.
【作者单位】：辽宁工程技术大学电子与信息工程学院;中国农业科学院蜜蜂研究所;
【基金】：辽宁省教育厅科学研究一般项目(L2014132) 辽宁省科技厅面上项目(2015020100)
【分类号】：TP391.41

【相似文献】