Curvelet entropy for facial expression recognition
PCM'10 Proceedings of the Advances in multimedia information processing, and 11th Pacific Rim conference on Multimedia: Part II
| Hi-index | 0.00 |
Facial expression recognition is necessary for designing any human-machine interfaces. A novel facial expression recognition method based on the Wavelet energy feature and neural network ensemble classifier is proposed in this paper. And six basic expressions – anger, disgust, surprise, happiness, fear and sadness are analyzed. Firstly, wavelet transform is used for static facial expression images and the wavelet energy is extracted from various sub-areas as facial expression features; Secondly, the neural network ensemble based on Bagging algorithm is used to offer the classifier trainings on facial expression recognition. Experiments results demonstrate an expression classification accuracy of 75.9% on the CMU-PITTSBURGH AU-Coded Face Expression Image Database, which conduct classification more accurately than other single neural network.