The Research Of Facial Expression Recognition Technology Based On The Video Stream

In recent years, with the furthering research on facial expression recognition theoryand the development of human-computer interaction technology, facial expressionrecognition based on the video stream has been a research focus in the field ofcomputing vision at the present. In this paper, some studies have been done to the facialexpression recognition technology. According to the classic algorithms in the field offacial expression recognition, an improved facial expression recognition method basedon2D Gabor wavelet transform and the classification tree is presented.The method based on Haar features and cascade Boosted machine learning isapplied in the face detection. A new algorithm using the relevance of the imagesequences in the video stream is proposed. The2D Gabor wavelet transform is appliedin the grid facial expression, then Gabor features of each grid together form the vectorof the whole facial expression image; An efficient facial expression extraction methodbased on the optimal frequency Gabor wavelets is presented to distinguish these threeconfused expressions(disgust, anger, sadness).For the seven basic expressions, theclassification tree method is used to divided the classification process into multiplelevels, the elastic template matching and weighted K-nearest neighbor(WKNN)classification algorithm are adopted in each level classification.Finally, a facial expression recognition system based on the video stream is built.The system can not only do a real-time face detection and expression recognition, butalso can validate facial expression recognition method presented in this paper.Experimental results show that the method proposed in the paper has a higherrecognition rate for the seven basic expressions and it can improve the real-time andaccuracy in face detection.