摘要
为了增强掌纹图像信息的分析和提取能力,提出了一种基于轮廓与边缘纹理特征融合的掌纹识别算法。该算法分别将灰度直方图和差分盒子维算法作用于掌纹图像的低频和高频成分,提取图像的轮廓和边缘纹理特征,然后并联融合二者特征向量进行匹配。通过香港理工大学Poly U掌纹图像数据库的验证,算法识别精度为99.56%,特征提取和匹配时间为56 ms,满足识别精度和实时性要求,且算法复杂度和造价低,具有较高的工程应用价值。
In order to enhance the ability of analysis and extraction of palm print image information, this paper proposes a novel palm print recognition method based on contour feature and edge texture feature fusion. We choose grey histogram and differential box counting to work on low-frequency features and high-frequency, and extract its contour feature and edge texture features. Then, making two features fusion in parallel. Experimental results on Poly U palm print experiment and compares with the traditional palm print recognition algorithm, the proposed method can obtain 99.56% recognition accuracy and the time of feature extraction and matching is only 56 ms, meeting the real-time requirements. At the same time, with low algorithm complexity and low cost in the actual demand, and has high value of engineering application.
引文
[1]Liu M H,Yap P T.Invariantrepresentation of orientation fields for fingerprint indexing[J].Pattern Recognition,2012,45(7):2532-2542.
[2]Jiang X B,You X,Yuan Y,et al.A method using long digital straight segments for fingerprint recognition[J].Neurocomputing,2012,77(1):28-35.
[3]Zhang D,Kong W K,You J,et al.Online palmprint identification[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2003,25(9):1041-1051.
[4]Zuo W M,Yue F,Zhang D.On accurate orientation extraction and appropriate distance measure for lowresolution palmprint recognition[J].Pattern Recognition,2011,44(4):964-972.
[5]Hong D F,Su J,Hong Q G,et al.Blurred palmprint recognition based on stable-feature extraction using a Vese–Osher decomposition model[J].PLOS ONE,2014,9(7):e101866.
[6]Hong D F,Pan Z K,Wu X.Improved differential box counting with multi-scale and multi-direction:A new palmprint recognition method[J].Optik-International Journal of Light Electron Optics,2014,125(15):4154-4160.
[7]Hong D F,Liu W Q,Su J,et al.A Novel Hierarchical Approach for Multispectral Palmprint Recognition[J].Neurocomputing,2015,151:511-521.
[8]魏伟波,洪丹枫,潘振宽,等.基于区域特征映射的模糊掌纹识别方法[J].自动化学报,2015,41(2):386-395.
[9]赵志刚,吴鑫,洪丹枫,等.基于信息熵的GLBP掌纹识别算法[J].计算机科学,2014,41(8):293-296.
[10]洪丹枫,潘振宽,苏健,等.VO图像分解模型的模糊掌纹识别方法[J].计算机辅助设计与图形学学报,2014,26(10):1737-1746.
[11]苑玮琦,范永刚,柯丽.基于灰度曲面匹配的掌纹识别方法[J].光电子·激光,2009,20(6):807-811.
[12]陈梓毅,康文雄,邓飞其.基于灰度差统计分析的掌纹线提取[J].计算机工程与设计,2011,32(2):653-655.S
[13]冯晓慧.基于局部分形维数的图像处理方法[J].数字技术与应用,2012,11:225-226.
[14]洪丹枫,凌青,彭娇娇.基于小波变换的ICA盲源分离算法[J].海军航空工程学院学报,2013,28(3):237-240.
[15]洪丹枫,苗俊,苏健,等.一种变步长凸组合LMS自适应滤波算法改进及分析[J].电子学报,2014,42(11):2225-2230.
[16]吴鑫,赵志刚,洪丹枫,等基于多尺度的掌纹辨识算法[J].青岛大学学报(自然科学版),2014,27(1):66-71.
[17]Zhao Y,Gao J,Chen G,et al.Multi-scale and multi-orientation texture feature extraction method based on fractal theory[J].Chinese Journal of Scientific Instrument,2008,29(4):787-191.
[18]Pan X,Ruan Q Q,Wang Y X,et al.Palmprint recognition using Gabor local relative features[J].Computer Engineering and Applications,2012,48(15):34-38.
[19]Poly UPalmprint Palmprint Database[R/OL].Availablefrom:http://www.comp.polyu.edu.hk/biometrics/S.
[20]王恩永.基于模糊聚类的灰度图像特征提取和识别研究[D].昆明:云南大学,2010:1-43.
[21]Lu G M,Zhang D,Wang K Q.Palmprint recognition using eigenpalms features[J].Pattern Recognition Letters,2003,24(9/10):1463-1467.
[22]郭金玉,苑玮琦.基于二维Fisher线性判别的掌纹识别方法[J].计算机工程,2008,34(6):212-213.