摘要
对贝叶斯人脸识别公式进行了简化,在此基础上设计了基于加权小波子带图像的贝叶斯人脸识别算法.首先对人脸图像进行小波分解,把分解得到的低频子图与类内均值做差作为类内差异图像进行贝叶斯测试,选择相似度最高的N幅图像作为候选图像,然后对候选图像再次利用高频子图与对应频段的类内均值做差作为模式矢量并行进行贝叶斯测试,通过加权排序得到最后结果.实验结果验证了该方法的有效性.
In this paper,we propose an improved Bayesian face recognition measure,and then a novel Bayesian approach to face recognition based on wavelet transform is proposed.Images are decomposed by applying wavelet transform,firstly the system uses low frequency sub-band and the average value of within class to select the N top-ranked candidate images by using Bayesian,secondly Bayesian recognition is parallel processed using these difference images of high frequency sub-band and average value of within class of all the candidate images.The face recognition results were gained through weigh-adding arraying.Experimental results prove the effectiveness of the proposed method.
出处
《北京交通大学学报》
CAS
CSCD
北大核心
2012年第5期124-128,共5页
JOURNAL OF BEIJING JIAOTONG UNIVERSITY
基金
河南省科技攻关计划项目资助(122102210505
12A520026)
河南省基础与前沿技术研究计划项目资助(122300410111)
关键词
人脸识别
小波变换
贝叶斯方法
face recognition
wavelet transform
Bayesian theory
