Facial expression recognition using local Gabor features and adaboost classifiers

To improve the accuracy and efficiency of facial expression classification, a facial expression recognition method using Gabor features and Adaboost classifiers is proposed. Local regions that best represent facial expressions are first segmented and located, and then the Gabor features of the local regions are extracted. Gabor features are selected by using distance discrimination and feature ranking, and local weighted Adaboost classifiers are constructed for facial expression recognition. Experimental results demonstrate that the feature dimensionality of the proposed algorithm is reduced by about 10 times compared to global Gabor features. Meanwhile, the recognition speed and accuracy are better than those of algorithms based on global Gabor features.