一种边缘二进制码的表情识别方法

针对如何从图像中提取有效的表情特征来提高表情识别率的问题,提出了一种基于边缘二进制码的表情特征提取方法,用于表情识别.该方法首先对图像进行边缘检测,然后对边缘的局部结构进行二进制码描述作为表情特征,最后利用支持向量机进行表情分类.在JAFFE人脸表情数据库上,分别用该方法和传统的方法进行试验,结果表明,该方法可显著提高表情识别率.

Facial expression recognition based on the binary code of edges

Facial expression recognition mainly involves face locating, feature extraction, and expression classification. However, extracting effective features of facial expressions from original face images is a vital step for improving facial expression recognition. In this paper, a binary code of edge method is proposed for extracting the expression feature, which is used for facial expression recognition. Firstly, the face image is processed by an edge detection operator. Then, the proposed binary code method is applied to describe the edge binary image as the expression feature. Finally, the SVM classifier is used for facial expression recognition. Experiments are performed on JAFFE facial expression database, and the results show that the proposed method achieves a higher recognition rate than the traditional method.