摘要
为获得人脸表情变化过程的动态信息,提高表情特征点在动态序列中跟踪的准确性,提出将主动表观模型与高斯金字塔Lucas-Kanada光流法相结合的表情识别方法。构造中性表情的基本主动表观模型,得到中性表情帧中的表情特征点,在动态序列中用高斯金字塔Lucas-Kanada光流法跟踪这些表情特征点,获得表情动态变化信息作为人脸表情特征,用支持向量机方法进行表情分类。在Cohn-Kanade+人脸表情数据库进行实验,实验结果表明,该方法能有效提高人脸表情识别的准确率。
To obtain the dynamic information changes and improve the accuracy of face feature point tracking in the dynamic sequence,an expression recognition method based on active appearance model and Gaussian pyramid Lucas-Kanada optical flow method was proposed.The basic shape model of the neutral expression was constructed by using the active appearance model to obtain the feature points in the initial neutral expression frame.In the dynamic sequence,the initial expression frame feature points were tracked using Gaussian pyramid Lucas-Kanada optical flow method,and the facial expression dynamic change information was obtained.Support vector machine was used to classify facial expressions.Results of experiments on the Cohn-Kanade+face expression database show that this method can effectively improve the accuracy of facial expression recognition.
出处
《计算机工程与设计》
北大核心
2017年第6期1642-1646,1656,共6页
Computer Engineering and Design
基金
辽宁省高等学校杰出青年学者成长计划基金项目(LJQ2013013)
关键词
主动表观模型
光流法
表情识别
支持向量机
动态序列
active appearance model
optical flow
expression recognition
support vector machine
dynamic sequence
