Facial expression features extraction based on Gabor wavelet transformation

Features extraction is a key step in facial expression recognition system. In order to extract facial expression features that are subject-independent and robust to illumination variety, this paper introduces a facial expression features extraction algorithm. Given a still image containing facial expression information, preprocessors are executed firstly, which include expression sub-regions segmentation, grayscale and scale normalization. Secondly, expression feature vectors of the expression sub-regions are extracted by Gabor wavelet transformation to form elastic graph for expression. Finally, the features of six basic expressions shown by different subjects under different illumination conditions are extracted and compared each other. The experimental results show that expression features can be extracted effectively based on Gabor wavelet transformation, which is insensitive to illumination variety and individual difference.