基于改进的隐马尔科夫模型的人脸识别方法  

Approach of Face Recognition Based on Improved Hidden Markov Model

作　　者：李忠健[1] 杨宜民[1] 梅平[1] 

机构地区：[1]广东工业大学自动化学院,广东广州510006

出　　处：《工业控制计算机》2014年第11期29-31,共3页Industrial Control Computer

摘　　要：提出一种基于改进的隐马尔科夫模型的人脸识别方法。该方法在待识别头像预处理过程中增加了光照补偿操作,减少光照对人脸识别的影响,提高识别的效率;图像在匹配的过程中,将观察序列平均分成几等分,与训练好的HMM相比较时,采用一边比较,一边筛选的方法,提高了识别的速率。实验结果表明,改进后的方法既可以提高识别效率,同时也能提高识别速率。An improved face recognition based on the hidden Markov model is proposed in this paper.To improve the efficiency of recognition,increase the illumination compensation operations in the pretreatment process of face image,while can reduce the impact of light on the face.To improve the recognition rate,observed sequences of the image are divided into several aliquots.Using the principle of compared and filtered at a moment,when observed sequences are being compared with the trained HMM during the matching process.

关 键 词：隐马尔科夫模型 人脸识别 光照补偿 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]