基于MG-LTP与ELM的微表情识别

特征提取和表情分类是表情识别的关键技术。针对传统方法识别率低的缺点,首先,提出了一种基于平均灰度的局部三值模式(MG-LTP)新算法,用于提取表情特征;其次,使用极限学习机(ELM)作为分类器,用于特征分类;最后,将二者结合用于表情识别,并进一步应用于人脸微表情识别中。在JAFFE数据库及CASME人脸微表情数据库进行试验,与传统方法对比,取得了较好的效果。

Micro-expression Recognition Based on MG-LTP and ELM

Feature extraction and expression classification are the key technologies of expression recognition. Considering of the low recognition rate of traditional methods, a new algorithm called mean gray local ternary patterns(MG-LTP) based on mean gray is firstly proposed in this paper,and MG-LTP is used to extract expression feature. Then, extreme learning machine(ELM) is used as a classifier for feature classification.Finally, the above two methods are combined for expression recognition,and further for facial micro-expression recognition.Experiments are completed on JAFFE database for expression recognition and CASME databases for facial micro-expression recognition.Compared with traditional methods,the method used in this paper achieves better results.