全文.doc-湖南文理学院学报编辑部.doc

doi: 10.3969/j.issn.1672–6146.2016.03.005 基于视频图像的面部表情识别研究综述( 梅英1, 2, 谭冠政2, 刘振焘3 (1. 湖南文理学院 电气与信息工程学院, 湖南 常德, 415000; 2. 中南大学 信息科学与工程学院, 湖南 长沙, 410083; 3. 中国地质大学 自动化学院, 湖北 武汉, 430074) 摘要: 面部表情识别是机器理解人类情感的前提, 是改善人机交互关系的关键。首先, 按照视频图像中面部表情识别的流程, 综述了表情识别的3个阶段: 人脸检测、表情特征提取、表情分类。重点介绍了表情特征提取和表情分类中所采用算法的原理、优缺点及应用场合, 并给出了部分算法的识别率对比结果。其次, 对人机交互中的微表情识别及表情识别的鲁棒性研究也做了介绍。最后总结了面部表情识别研究中存在的问题及难点, 探讨了该领域值得进一步研究的问题。 关键词: 表情识别; 特征提取; 表情分类; 微表情; 鲁棒性 中图分类号: TP 391 文章编号: 1672–6146(2016)03–0019–07 Review on facial expression recognition based on video image Mei Ying1, 2, Tan Guanzheng2, Liu Zhentao3 (1. College of Electrical and Information Engineering, Hunan University of Arts and Science, Changde 415000, China; 2. School of Information Science and Engineering, Central South University, Changsha 410083, China) Abstract: Facial expression recognition is the precondition for the computer understanding human emotion, and it is the key to improve human-computer interaction. Firstly, a survey facial expression recognition in video is provided according to the three steps: face detection, facial feature extraction, expression classification. Algorithms applied in feature extraction and expression classification are mainly summarized, including their principles, advantages and disadvantages, applications, as well as the accuracy comparisons of some algorithms. Secondly, some issues about micro-expression recognition and robustness research are described in the human-computer interaction. Finally, the existing problems and the difficulties of facial expression recognition, as well as the issues worthy of further study are concluded. Key words: Facial expression recognition; feature extraction; expression classification; micro-expression; robustness 面部表情识别是近些年逐步兴起的情感计算的重要组成部分。美国麻省理工学院的Picard教授[1]曾在《Affective Computing》中指出“情感是未来计算机能够有效工作的必要条件之一, 希望通过赋予计算机识别用户情感的能力, 以便更好地服务人类。”人类的情感常常通过面部表情、语音、姿态等来表达, 但是, 它们所传递的信息量有差别。美国心理学家Mehrabian[2]提出