基于ICA与HMM的表情识别  

Facial Expression Recognition Based on Independent Component Analysis and Hidden Markov Model

作　　者：周书仁[1,2] 梁昔明[1] 朱灿[1,2] 杨秋芬[1] 

机构地区：[1]中南大学信息科学与工程学院,长沙410083 [2]长沙理工大学计算机与通信工程学院,长沙410076

出　　处：《中国图象图形学报》2008年第12期2321-2328,共8页Journal of Image and Graphics

基　　金：国家重点基础研究发展规划(973)项目(2002CB312200);湖南省自然科学基金项目(06JJ50109)

摘　　要：独立分量分析(independent component analysis,ICA)是一种盲源分离的有效方法,为了进一步有效提取表情图像中隐藏的信息和提高表情识别率,可将它应用于人脸表情识别。由于脸部表情为人类情感、认知过程的研究提供了极为重要的测量依据,因此表情特征的提取和特征序列所代表的表情状态是表情识别过程中的重要步骤。为了更好地进行表情和情感的分类,提出了一种ICA结合隐马尔可夫模型(HMM)识别表情的情感分类系统,该系统首先利用ICA算法进行表情特征提取,为了加快特征提取的速度,这里采用了FastICA算法;然后通过7个训练好的HMM进行表情识别。实验结果显示,该系统使人脸表情识别的整体效果有了提高,取得了令人满意的效果,可以用来识别人脸表情。As an effective approach of blind source separation （BSS）, independent component analysis （ICA） is a recently developed method in facial expression recognition field, which is used to effectively extract the hidden information of expression images and can improve the rate of expression recognition. Facial expression provides a crucial measure for studies of human emotion, cognitive processes, and social interaction. The key focuses of facial expression recognition are the extraction of expression features and the expression states using features. This paper proposes an expression recognition system based on ICA and hidden markov model （HMM）. The system includes two parts： First, it is applied to extraction of expression features using ICA algorithm. In this process it adopts FastlCA algorithm in order to increase the speed of feature extraction and its function is prior to primary component analysis （PCA）. Second, it is applied to recognizing facial expression using seven HMMs its time efficiency is prior to support vector machine （SVM）. Experimental results show that the system increases the whole effectiveness and accuracy of facial expression recognition, and prove that the algorithm is efficient and feasible.

关 键 词：盲源分离 独立分量分析 表情识别 隐马尔可夫模型 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]