Facial Expression Recognition Using Texture and Edge Descriptors

AbstractFacial expression recognition is one of the most important computer vision issues that has many applications. One of them is the Human computer interaction. In this paper, a method for facial expression recognition using texture and edge descriptors is proposed. Facial expression recognition generally consists of three steps: preprocessing, feature extraction and classification. In this paper, histogram Equalization has been used in the proposed method for pre-process the input images in which the face is present. In this paper, the focus is on the feature extraction and a combination of LDP1 and HOG2 descriptors has been used to improve the existing methods. After feature extraction, the support vector machine was used to classification the facial expression recognition. This article uses the JAFFE database. The database contains 213 images of seven facial expressions (happy, sad, angry, fear, disgust, surprised and natural) taken from 10 Japanese female models. The results showed that the proposed method with 99.04% accuracy in the facial recognition test had a better performance than the methods of previous researchers.