基于MSVR和Arousal-Valence情感模型的表情识别研究  

Facial expression recognition method based on MSVR and Arousal-Valence emotion model

作　　者：杨勇[1,2] 黄文波[1] 金裕成 顾西存 

机构地区：[1]重庆邮电大学计算智能重庆市重点实验室,重庆400065 [2]韩国仁荷大学情报通信工学部 [3]重庆邮电大学图形图像与多媒体实验室,重庆400065

出　　处：《重庆邮电大学学报（自然科学版）》2016年第6期836-843,共8页Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)

基　　金：韩国科学与信息科技未来规划部2013年ICT研发项目(10039149);重庆市自然科学基金项目(CSTC;2007BB2445);2015年重庆市研究生科研创新项目(CYS15174)~~

摘　　要：通常的表情识别方法是对基本情绪进行表情分类,然而基本情绪对情感的表达能力有限。为了丰富情感的表达,研究采用Arousal-Valence情感模型,从心理学的角度对Arousal-Valence模型中Arousal维度和Valence维度之间的相关性进行了分析,并用统计学方法对AVEC2013,NVIE和Recola 3个数据集进行研究,实验结果表明它们之间具有正相关关系。为了利用Arousal-Valence之间的相关性,采用多输出支持向量回归(multiple dimensional output support vector regression,MSVR)算法作为表情的训练和预测算法,并结合特征融合和决策融合提出了一种基于MSVR的两层融合表情识别方法。实验结果表明提出的表情识别方法比传统的方法能取得更好的识别效果。The most commonly used facial emotion recognition method is classitying basic emotions. However,the basic emotion theory has a limited leval of ability to express emotion. To enrich emotion expression,the arousal-valence continuous emotion space model is adopted in this paper. Firstly,the correlation between arousal and valence is discussed from the perspective of psychology and researched based on the statistics. The experimental results on AVEC2013,NVIE and Recola datasets indicate the correlation is positive. Then,in order to use the correlation between arousal and valence,MSVR（ multiple dimensional output support vector regression） is adopted to train and predict facial emotion,and a new facial emotion recognition method based on MSVR and two-level fusion is proposed,which combines feature fusion and decision fusion. The contrast experimental results show that the proposed method can get better recognition result than the traditional methods.

关 键 词：表情识别 Arousal-Valence情感维度 相关性 多输出支持向量回归(MSVR) 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]