摘要
探讨了利用G abor小波和隐马尔可夫模型(HMM)进行人脸识别的方法.首先对人脸图像进行多分辨率的G abor小波变换;然后在图像上放置一组网格结点,每个结点用该结点处的多尺度G abor幅度特征描述,采用独立元分析法对每个结点进行去相关和降维;最后形成特征结,把每个特征结作为观测向量,对隐马尔可夫模型进行训练,并将优化的模型参数用于人脸识别.ORL人脸库的实验结果表明,该方法识别率高,工程上易于应用.
A new method based on Gabor wavelets transform and hidden Markov model (HMM) for face recognition is proposed. The Gabor wavelet representation of an image is the convolution of the image with a family of multiresolution Gabor kernels. The vectors called nodes, over a dense grid of image points are formed. Each node is labeled with a set of complex Gabor wavelets coefficients. The magnitudes of the coefficients are used for recognition. Feature nodes, as observation vectors of HMM, are derived by using independent component analysis to reduce the dimensionality of nodes. A set of images representing different instances of the same person is used to train each HMM. Each individual in the database is represented by an optimal HMM face model. Experimental results on the ORL face database show that the proposed algorithm provides a high recognition rate with a good perspective.
出处
《控制与决策》
EI
CSCD
北大核心
2005年第9期1073-1076,共4页
Control and Decision
基金
中国科学院科技创新基金项目(1021-07)
关键词
人脸识别
GABOR小波变换
独立元分析
隐马尔可夫模型
Face recognition
Gabor wavelets transform
Independent component analysis
Hidden Markov model(HMM)