Studies On Facial Expression Recognition Based On Gabor Wavelet

As China entered the aging society, the problem of empty-nesters increasingly brought to the attention of the country. EPIC laboratory's research of robots and cloud computing technology for applications of real-time intelligent electronic health care, is aim to improve the efficiency of health care services of empty nest elderly, to implement low cost health care and at the same time meet the needs of empty nest old people "psychological endowment". Identifying facial expressions can help understand the psychological condition of the elderly in real time, so that the robot can do better emotional interaction with the old people and complete the function of health care.Facial expression recognition, is to process the information of facial expressions and extract the features and then classify, so that computer can infer the state of people's mind according to the facial expression information, and realize human-machine intelligent interaction. In general, the recognition of facial expression mainly has three steps: face image preprocessing, feature extraction and expression classification.Research background and research status of facial expression recognition are introduced, and expounds the three basic part of the problems related to facial expression recognition. Adaboost face detection technique based on Haar features and the normalization method of the image Angle, dimension and gray scale in expression image preprocessing are introduced. In the part of facial expression feature extraction,summarizes some common expression feature extraction method, and introduces in detail the Gabor wavelet transformation, principal component analysis(PCA) and linear discriminant analysis(LDA) feature extraction method. In expression classification,discusses the commonly used expression classification algorithm K nearest neighbor(KNN) and support vector machine(SVM). A series of experiments on facial expression recognition method based on Gabor wavelet feature was conducted, the most suitable facecutting standard, scale normalization size, Gabor filter group, feature dimension reduction method and expression classifier were chosen, based on the experimental results a real-time face recognition system based on PC and a real-time facial expression recognition system based on WEB is designed.