人脸图像ICM时间序列识别方法  

Method of face image recognize by ICM time serials

作　　者：李建锋[1,2] 

机构地区：[1]吉首大学数学与计算机科学学院,湖南吉首416000 [2]中南大学信息科学与工程学院,湖南长沙410083

出　　处：《计算机工程与设计》2011年第4期1398-1400,1405,共4页Computer Engineering and Design

基　　金：国家自然科学基金项目(60803024);湖南省教育厅科研基金项目(09C745)

摘　　要：提出一种仿生物视觉算法模型的彩色人脸图像识别方法——视觉交叉皮层时间序列人脸特征提取算法。将彩色人脸图像从RGB空间转换至HSI空间,对HSI空间的各个图像分量分别提取时间序列,将各个分量的时间序列连接形成整体的人脸图像特征。该序列对不同人脸图像具有较高的区分度,而对于不同角度的相同人脸图像却表现出一致性。用第一范式距离作为判据进行人脸图像识别,并与PCAI、CA以及基本的PCNN进行比较,实验结果表明,提出的方法识别速度明显高于比较的几种方法,识别准确率可以达到86.73%。A intersecting cortical model time series facial feature extraction algorithm is present which is a color face image recognition algorithm based on bionic vision algorithms model is proposed.Color face image is transformed from RGB space to HSI space.Time series is extracted from each image component in HSI space respectively,then time series of each image component is connected which forms the face image feature.Experimental result shows that high degree of discrimination can be obtained from the time series of different face image,while consistency can be achieved from different face images of the same person captured from different angle.Finally,distance of the first paradigm is adopted as a criterion for face image recognition.The comparison with PCA,ICA and basic PCNN demo-nstrate that the proposed algorithm is much faster than the others.The precision rate reaches 86.73%.

关 键 词：人脸识别 交叉视觉皮层 脉冲耦合神经网络 时间序列 彩色空间 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]