Expression recognition algorithm based on the relative relationship of the facial landmarks

摘要

In order to improve the facial expression recognition accuracy in complex contexts and decrease the identification time, this paper presents an efficient expression recognition algorithm based on feature points of the facial landmarks. In practice, facial expressions mainly focus on a few parts of muscle activity, which provides valuable reference to infer human emotions and intentions. Facial expression is a powerful nonverbal way for human to transmit information and reveal emotion. In this paper, we focus on geometric positions of key parts of the face. Firstly, the face area is detected in a photo or video, then the key parts of the face is extracted and the position correction is performed. A set of key points is located using the relative position of the face. This process can not only effectively avoid the impact of the environment and the light on the sample, but also greatly improve the recognition of facial expressions. With the development of human-computer interaction, facial expression recognition has become a hot topic in the field of pattern recognition. After years of development, facial expression recognition has achieved some success such as HOG. In this paper, HOG feature extraction method of facial expression is used as a contrast. Experimental results show that the proposed algorithm can extract key information and achieves higher recognition accuracy.