Face Recognition Algorithm using Vector Quantization Codebook Space Information Processing
Face Recognition Algorithm using Vector Quantization Codebook Space Information Processing
复制标题
基于矢量量化码本空间信息处理的人脸识别算法
DOI：
10.1080/10798587.2004.10642870
复制
发表时间：
2004
期刊：
影响因子：
--
通讯作者：
Feifei Lee
中科院分区：
文献类型：
--
作者：
T. Ohmi;K. Kotani;Qiu Chen;Feifei Lee
摘要提出了一种新的信息处理算法-矢量量化（VQ）码书空间信息处理。在此基础上，我们提出了一种非常简单但可靠性很高的人脸识别方法--VQ直方图法。用理论方法构造了由33个低频模式组成的矢量量化通用码本。通过对人脸图像进行矢量量化处理得到的码矢量参考（或匹配）计数直方图，作为一种非常有效的个人特征值。通过对人脸图像进行适当的低通滤波和矢量量化处理，可以提取出对人脸识别有用的特征。实验结果表明，识别率为95.6%的40人的400张图像的公开可用的AT&T数据库包含变化的照明，姿势和表情。通过结合多个低通滤波程序，识别率提高到97%或更高.在考虑了矢量量化直方图法的本质后，还对矢量量化直方图法进行了改进。
Abstract We propose a novel information-processing algorithm called Vector Quantization (VQ) codebook space information processing. Based on this algorithm, we have developed a very simple yet highly reliable face recognition method called VQ histogram method. General codebook consisting of 33 low-frequency patterns for VQ processing is created by theoretical method. codevector referred (or matched) count histogram, which is obtained by VQ processing of facial image, is utilized as a very effective personal feature value. By applying appropriate low pass filtering to facial image and VQ processing, useful features for face recognition can be extracted. Experimental results show recognition rate of 95.6 %for 40 persons’ 400 images of publicly available AT&T database containing variations in lighting, posing, and expressions. By combining multiple low pass filtering procedures, recognition rate increases up to 97 % or higher. After considering the essence of the VQ histogram method, moreover, we also have d...