A novel approach of facial expression recognition based on shearlet transform
Yang Lu, Shigang Wang, Wenting Zhao, Yan Fei Zhao, Jian Jun Wei · 2017
A novel facial expression recognition method based on discrete shearlet transform is proposed in this paper, which is a new image multiscale geometric analysis method. In addition to multi-resolution and time-frequency localization owned by wavelet transform, the shearlet transform also has anisotropy and directionality. Firstly, normalization and equalization are applied to all test and training images. Then the facial expression features are extracted based on discrete shearlet transform, and Vector Support Machine is used to classify the seven expressions (happiness, sadness, surprise, disgust, fear, anger and neural) of JAFFE database. Experimental results show that the recognition rate of the proposed method is improved comparing to several known methods.