2011 14th IEEE International Conference on Computational Science and Engineering (2011)
Dalian, Liaoning China
Aug. 24, 2011 to Aug. 26, 2011
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/CSE.2011.96
Facial expression recognition research is an important research direction of computer vision on human face analysis field. This paper propose a mark scheme which can be compatible with Constrained Local Model (CLM), and then propose a method which combines local binary patterns' features and SVM classifier to recognize specific expressions. Our method first extracts LBP features from training data, then uses these descriptors to train SVM classifier, which can later be used to do classification on new features. Experiment results indicate this method combine the properties of LBP, which can be easy to realize and has good performance of description, and the properties of SVM, which is insensitive to the dimension of sample data, and has strong generalization capabilities.
Expression Recognition, LBP, SVM, Mark Scheme
Z. Jian, Y. Xiaoguang, W. Lirong, X. Jing and W. Jianlei, "Facial Expression Recognition Based on Local Texture Features," 2011 14th IEEE International Conference on Computational Science and Engineering(CSE), Dalian, Liaoning China, 2011, pp. 543-546.