云计算平台下的人脸识别  

Face recognition on cloud computing platform

作　　者：张泊平[1] ZHANG Boping(School of Information Engineering, Xuchang University, Xuchang 461000, China)

机构地区：[1]许昌学院信息工程学院,河南许昌461000

出　　处：《现代电子技术》2016年第18期88-90,95,共4页Modern Electronics Technique

基　　金：国家自然科学基金资助项目(61163034);河南省科技攻关资助项目:基于视觉测量的GPS导航技术研究(132102210491)

摘　　要：对云计算平台下的人脸识别方法进行研究,在Hadoop平台上建立基于支持向量机分类模型的人脸识别方法,以发挥Map Reduce并行计算优势,提高识别效率。由于常规LBP算子和深度LBP算子识别人脸特征不同,所以该文使用加权方式结合两种算子,以发挥各自的优点。最后,使用人脸识别领域应用最为广泛的Yale B人脸数据库、ORL人脸数据库以及FERET人脸数据库对该文研究的云计算平台下的人脸识别算法进行实例分析。实验结果表明,所研究的识别方法的识别准确率要高于使用传统方法的识别率。在相同云计算平台下,使用BP神经网络和RBF神经网络建立分类器与该文研究的人脸识别方法进行对比,结果表明,在云计算平台下,使用SVM分类器进行人脸识别的效果优于BP神经网络和RBF神经网络分类器。The face recognition method on cloud computing platform is studied. A face recognition method based on supportvector machine classification model was established on Hadoop platform to develop the parallel computing advantages of MapRe?duce, and improve the recognition efficiency. The features of face recognition of the conventional LBP operator and depth LBPoperator are different, so the two operators are combined with weighted method to give play to their respective advantages. TheYale B face database, ORL face database and FERET face database widely used in face recognition field are adopted to analyzethe face recognition algorithm on cloud computing platform with instance. The results show that the recognition accuracy of thestudied method is higher than that of the traditional methods. The face recognition method studied in this paper is compared withthe classifier established with BP neural network and RBF neural network on the same cloud computing platform. The resultsshow that, on cloud computing platform, the face recognition effect of using SVM classifier is better than that of using BP neuralnetwork and RBF neural network.

关 键 词：云计算 HADOOP平台 人脸识别 支持向量机 神经网络 

分 类 号：TN911.34[电子电信—通信与信息系统] TP317.4[电子电信—信息与通信工程]