摘要： 介绍了基于计算机视觉的表情识别的定义、应用前景和困难所在；阐述了表情识别的步骤，并比较了与人脸识别的异同；重点按照不同的特征提取和分类器设计方法对表情识别技术进行了综述。介绍了几何特征、统计特征、频率域特征和运动特征的提取方法及线性、神经网络、支持向量机分类器的设计和选择方法，并进行了简单的分析和比较；最后展望了表情识别的发展方向。

关键词: 表情识别；光流；支持向量机

Abstract: This paper gives definition, application future and difficulties of facial expression recognition based on computer vision. It describes the steps of facial expression recognition and compares it with face recognition. Then current recognition technologies are roughly introduced and classified according to different method of feature extraction and classifier design. Four main methods of extracting geometric, statistical, spectral-transform-based and kinetic features are elaborated. So are the design of linear, ANN, SVM classifiers. Finally, based on the analysis and comparison, key factors in facial expression recognition technologies are concluded.

Key words: Facial expression recognition; Optical flow; SVM