Based On OpenCV Face Recognition System Research And Implementation

The face recognition is face detection and location in the image or video stream,including the location of the face in the image or video stream, the size, shape, and thenumber of information in recent years due to the rapid computing speed makes thedevelopment of image processing technology has been widely applied in many fields, whichincludes intelligent monitoring, secure transactions, safer and more friendly andhuman-computer interaction. Today, many companies or research as a separate subject tostudy and explore.This thesis focuses on face recognition technology from the exposition of the theory tothe demonstration and completed a face recognition system including the acquisition phase ofthe face image through DirectShow camera image acquisition, image pre-processing stage byhistogram equalization for image enhancement, the AdaBoost stage to create a trainingsample is divided into the positive samples and negative samples to the sample of the faceimage training, training the classifier and face detection Camshift stage skin color clusteringanalysis method based on the HSV color space, and to establish a model for the color in theimage or video stream is not necessarily face problems, the algorithm detects a match, use acolor space vector to determine the candidate interval, and then face discrimination, andimproved face for single and multiplayer face the PCAphase of the sample images projectionin the feature space, the average face and the face of principal components, and paper for eachstage of processing to make appropriate theoretical arguments and experimental results, andappropriate improvements in the algorithm.This thesis is developed by Intel-based machine vision library OpenCV, the Visual C++2008 integrated development environment, based on the MFC on a PC on the systemdevelopment. The test results table, mentioning improve system stability and the recognitionaccuracy rate also need to continue research and exploration.