A Facial Expression Recognition Methond Based on Improved HOG Features and Geometric Features

Facial expression recognition plays a very important role in many applications. In order to extract facial expression features better and improve facial expression recognition rate, this paper proposes an expression classification method that combines facial texture features and geometric features. Firstly, an improved HOG operator is proposed to better extract the edge information of the expression texture. Then, using an ensemble of regression trees algorithm to effectively extract facial feature points, and through a large number of experiments, define and select effective expression geometry information. Finally, the expression classification is realized by fusing the improved HOG operator to extract texture feature information and expression geometry information. Experiments show that the proposed algorithm achieves better results than the general texture feature recognition algorithm.