[1]苏士美,王燕,王明霞.基于加权小波分解的人脸识别算法研究[J].郑州大学学报(工学版),2014,35(01):5-9.[doi:10.3969/j.issn.1671-6833.2014.01.002]
　SU Shimei,WANG Yan,WANG Mingxia.Face Recognition Research Based on Weighted Wavelet Decomposition[J].Journal of Zhengzhou University (Engineering Science),2014,35(01):5-9.[doi:10.3969/j.issn.1671-6833.2014.01.002]
点击复制

基于加权小波分解的人脸识别算法研究()
分享到：

《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]

卷:
35
期数:
2014年01期
页码:
5-9
栏目:
出版日期:
2014-02-28

文章信息/Info

Title:
Face Recognition Research Based on Weighted Wavelet Decomposition
作者:
苏士美; 王燕; 王明霞
郑州大学电气工程学院,河南郑州,450001
Author(s):
SU Shimei;WANG Yan;WANG Mingxia
School of Electrical Engineering, Zhengzhou University, Zhengzhou 450001 , China
关键词:
小波分解; PCA; 人脸识别; 离散小波变换
Keywords:
wavelet decomposition;  PCA;  face recognilion;  diserete wavelet transform
DOI:
10.3969/j.issn.1671-6833.2014.01.002
文献标志码:
A
摘要:
小波变换能有效地将图像分解成高频和低频信息.现有的人脸识别算法多数都是基于小波分解后的低频信息,没有充分利用高频信息.PCA是人脸识别中被广泛使用的一种算法,它具有实现简单、正面图像识别率高等优点,但PCA算法计算量大,且易受光照、表情变化等因素的影响.基于加权小波分解和PCA算法提出一种新的人脸识别算法,对小波二级分解后的低、高频子分量进行加权融合,以便充分利用人脸的细节信息,并分别给PCA前三个最大主分量赋予一个新权值,来弥补传统PCA算法对光照、表情变化敏感的缺点.实验结果表明提出的人脸识别算法在识别率和训练时间方面都得到了明显的改进.
Abstract:
Wavelet transform ean elfeetively deeompose an image into high-requeney and low-frequeney infor.mation. Existing faee recognition algorithms mostly based on the low-frequeney information do not take full useof the high-requeney information. PCA has been widely used in faee recognition algorithms beeause of its sim.ple realization and high reeognition rate for frontal faces. But PCA is computationally intensive, and it’s vul.nerable to the changes of illumination and facial expression. This paper presents a new faee recognition methoduniting weighted wavelet decomposition with the PCA algorithm. This method can make up for the disadvan.tage of the traditional PCA algorithm which is sensitive to the changes of illumination and facial expression , forthe faee details are fully utilized through the weighted fusion of the seeondary decomposition of wavelet low fre.queney components and setting the new weighted values to the first three maximum prineipal vectors of PCA.E’xperimental results show that the proposed method has a marked improvement in recognition rate and trainingtime.

相似文献/References:

[1]董芳芳,尚志刚,刘新玉,等.解码动物转向行为的ICA-小波特征提取方法[J].郑州大学学报(工学版),2017,38(03):39.[doi:10.13705/j.issn.1671-6833.2016.06.006]
　Dong Fangfang Shang Zhigang Liu Xinyu Wanhong.ICA-wavelet Feature Extraction Method for Decoding of Pigeon Turning[J].Journal of Zhengzhou University (Engineering Science),2017,38(01):39.[doi:10.13705/j.issn.1671-6833.2016.06.006]

更新日期/Last Update: