Research Of Facial Expression Recognition Method Based On Wavelet Decomposition And Selected VLBP Feature

Facial expression recognition is one of the most challenging problems in the fields of pattern recognition,machine vision,affective computing and psychology. Facial expression recognition is an important part of affective computing and intelligent human-machine interactive,which has a wide range of applications and potential market value.Facial expression recognition is to analyze and detect the special expression state from given expression images or video frames and then to ascertain the subject's specific inborn emotion,achieving smarter and more natural interaction between human beings and computers.The study of facial expression recognition has found its values in economy and society.In this thesis,facial expression feature extraction,feature selection are studied. Several improved algorithms and methods for these tasks are developed.The performances of our methods are illustrated by experiment results.The main work is described as bellows:(1) Facial expression images obtained method based on wavelet decomposition is presented.Four sub-images are obtained by wavelet decomposition on facial expression image,which contain different frequency of original facial image.The first sub-image contain the detail information,the second one contain the horizontal edge information and the third one contain the vertical edge information of the original facial image.Thus we select the first three sub-images as the obtained facial images, because the facial expression information we wanted is the texture information and the deformation.Comparing with not using our method,the expression information reduce 1/4 for process of facial expression feature extraction and recognition,and the speed of feature extraction and recognition will upped properly.(2) Blocking improved-VLBP feature extraction method based on wavelet decomposition is presented.In this method,these obtained images with different frequency divided by different block size.Then facial expression feature extracted on these images by our presented feature extraction method.It can reduce the feature dimensions effectively and accelerate the speed of facial expression recognition by extracting feature after wavelet decomposition and blocking.Experiments show that this method can get a better expression recognition rate with high speed of expression recognition.(3) Feature selection method based on contribution analysis algorithm of neural network is presented.The dimension of the VLBP feature is very large,including redundant features and irrespective features,so we present feature selection method based on contribution analysis algorithm of neural network in order to reduce the dimension of the expression features,which also decrease the workload for calculation and increase the recognition rate.In this method,the expression features which with high dimension are extracted by improve VLBP algorithm,and then the features extracted are selected by contribution analysis algorithm of neutral network. Furthermore,the effectiveness of the selected facial expression features is analyzed by cluster analysis.Thereby the selected features are confirmed.Experiment shows that this method can reduce the dimension of the facial expression features and improve the expression recognition rate more.(4) A prototype system of facial expression recognition based on image sequence is designed and implemented by using the object-oriented methods.It can be used to prove the effectiveness of above algorithms.