Facial Expression Recognition Based On Facial Image

Facial expression recognition is one of the most challenging problems in the fields of biometric identification, image processing, machine vision, movement tracking, pattern recognition, physiology and psychology, and it has become a hot research topic in the filed of pattern recognition and artificial intelligence in recent years. 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. In recent years, with the effects of some research institutes and universities all over the world, face expression recognition technology develops rapidly. Meanwhile, some new problems occurred. Until now, many difficult problems are unsolved, and there is still no system which can be used in practice.Facial expression recognition consists of such modules as face detection, feature extraction, feature selection and expression classification. In this paper, feature extraction, feature selection and expression classification are studied. Several improved algorithms and methods for these tasks are developed. The performances of our methods are illustrated by simulation experiment results. The major contributions of this paper are as follows.1. An active appearance model (AAM) which based on statistical theory is implemented for facial feature points locating.2. According to the already existing geometric features of facial expression, a feature table that covers almost all geometric features appeared in the literatures is summed up.3. Attribution reduction in rough set theory is adopted as the method for facial feature selection. Attribution reduction based on rough set theory is a feature selection method that doesn't rely on priori knowledge, and has been applied successfully in many fields. It is applied for facial geometric feature selection. It is proved to be an effective approach that can reduce the dimensions of features.4. A novel approach based on Rough Set theory based attribution reduction and Support Vector Machine (RS-SVM) is proposed. SVM is an efficient classifier which is developed in the 90's of the last century. It has already been applied in a wide range of applications. SVM is evolved from Statistical Learning Theory (SLT) and is proved to be a good learning machine theorically. It has the most generalization ability based on small training set. In this paper RS-SVM processes data with attribution reduction firstly, and then classifies data with SVM. The experiment results show that our approach reduces the computing cost of SVM greatly with a little loss of classification ability.5. Facial Expression Recognition System (FERS) is developed. It consists of such modules as face detection, feature extraction and expression classification. Users can choose each method freely, and can add new methods into the system also.