Research And Design Of The Face Recognition System Based On Cloud Computing

Face recognition and cloud computing are the most popular research areas currently.Face recognition has been rapid popular on video surveillance and other way, as the usernon-fit state, a distance of rapid identification technology. The cloud computing is theanother great change after the one in the1980s which is from mainframe computer to theclient-server. The combination of cloud computing and face recognition is a good way tosolve the problem that the face recognition process needs a large amount of computing andstorage space.The goal of this article is to achieve a human face recognition system which clientsoftware get the face image and preprocess, cloud platform is going to analyze, process andidentify. And I deeply discussed and analyzed the algorithm of face recognition based onPCA and the Hadoop distributed cluster system. It researches the distributed File System ofthe Hadoop cloud platform and PCA algorithm.The article builds a three node of Hadoopframework, discusses the four processes of face detection, processing, feature extraction andrecognition, analyzes three face detection method based on the featureextraction, template matching and statistics, elaborates on the pretreatment methods offiltering, histogram equalization and gray stretch, focuses on the research of facerecognition algorithm based on principal components analysis, realizes the algorithm ofimage acquisition, processing and recognition based on the OpenCV computer vision library,Studies on Face detection method of Haar classifier in OpenCV, develop the client based onVS2010development environment configured the OpenCV libraries and Map and Reducemodule based on eclipse development environment configured the JavaCV libraries.Through the research of this paper,It’s great influence for lightand background complexity factors on face recognition base on PCA,Hadoopcloud platform for large data processing performance is great, but the real-time performanceis poor.Therefore, this paper deal with the improvement of recognition algorithmsandreal-time cloud framework for further study.