Automatic Facial Expression Recognition

Famous psychologist Albert Mehrabian had made a conclusion that the communication is made up of 7% words, 38% tone of voice and 55% facial expression. Facial expression which carries rich information of body behavior is the leading carrier of human emotions. As an essential branch of affective computing, facial expression recognition is the groundwork of the emotion understanding, and is the precondition for computers to understand human's emotion. In this dissertation, face detection, face pre-processing, facial expression features extraction and recognition are talked about. The main contents of this dissertation are as follows:1. The research background of the facial expression recognition is introduced, so is the development. A survey of facial expression feature extraction and classification algorithms are presented.2. The face detection and the expression image preprocess is studied. Geometric pre-processing including rotation, clipping and scaling, and Lighting pre-processing such as histogram equalization, are adopted.3. An effective facial feature extraction algorithm, local binary patterns (LBP), is studied. LBP can ensure extract enough detailed expressional information. Though LBP can describe the local features accurately, may not be good for classifying. A combined method of LBP and LDA is proposed. Experiment results show the new method is better than the LDA both in recognition rate and in time consumption.4. An automatic facial expression recognition system is realized, which can classify the facial expression into six types: happiness, surprise, sadness, fear, anger and disgust in real-time.