| Facial expression contains abundant and exquisite emotional and psychological information,facial expression recognition is mainly involved with two issues,namely how to effectively obtain facial expression feature and how to effectively classify facial expression.The method of classification and recognition for facial expression by extracting facial feature has become an important research topic in the field of pattern recognition,machine vision,robotics,image processing,security,medical,communication and human-computer interaction etc.In this dissertation,the main research contents and innovative work include the following two aspects:(1)After studying the traditional facial expression recognition methods,and analyzing their merits and shortcomings,a new approach for facial expression recognition based on Fuzzy Support Vector Machine(FSVM)and K-Nearest Neighbor(KNN)is presented in this paper.At first,the feature of the static facial expression image is extracted by the Principle Component Analysis(PCA),then,the algorithm calculates the Euclidean distance of the input samples to all the categories,divides the region into different types,and combines with the characteristic of the FSVM and KNN,switch the classification methods to the different types.The result of the experiment show that proposed algorithm can achieve good recognition accuracy,and can simplify the computation complexity.(2)To improve the accuracy recognition,a new approach for facial expression recognition based on Local Binary Pattern(LBP)and Compression Sensing is presented in this paper.The face images or image sequences in the expression database is inspected and located at first,then,the LBP is adopted to extract features for the static facial expression image.Finally,the facial expression is discriminated using the sparse representation classifier.The result of the experiment on JAFFE(Japanese Fe-male Facial Expression)database indicates that the proposed method has good classification performance. |
