改进的隐马尔可夫表情识别模型参数优化算法  

Modified parameter optimization algorithm for facial expression recognition based on hidden Markov model

作　　者：黄小娟[1] 吴荣腾[1] 

机构地区：[1]闽江学院计算机科学系,福建福州350108

出　　处：《河南工程学院学报（自然科学版）》2014年第4期59-64,共6页Journal of Henan University of Engineering：Natural Science Edition

基　　金：大学生创新创业训练计划项目(201310395010);福建省教育厅科技项目(JK2012038)

摘　　要：人脸表情识别是人工智能领域中极富挑战性的课题,针对表情识别中存在的识别率低与计算量大的问题,提出了一种新的改进的隐马尔可夫表情识别模型参数优化的算法.先采用新的初始参数优化模型,然后利用Baum-Welch算法进行重估计,从而建立新的HMM人脸表情模型.实验结果表明,新模型明显提高了人脸表情的识别率并降低了计算量.Facial expression recognition is quite a challenging subject in the field of artificial intelligence. Aiming at the problems of low recognition rate and the large computational problem of face expression recognition,a new modified parameter optimization algorithm is proposed for facial expression recognition based on the hidden Markov model. The method uses the initial parameters to optimize the model,and then uses Baum-Welch algorithm to estimate the parameters again. Hence,the new facial expression model based on HMM is established. The experimental results show that the new model significantly reduces the calculation amount and improve the facial expression recognition rate.

关 键 词：人脸表情识别 隐马尔可夫模型 离散余弦变换 EM算法 前向后向算法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]