>> 自然科学版期刊 >> 2021 >> 2021年05期 >> 正文
基于LBP和ELM的人脸识别算法研究与实现
时间: 2021-09-10 次数:

王红星, 胡永阳, 邓超.基于LBPELM的人脸识别算法研究与实现[J].河南理工大学学报(自然科学版),2021,40(5):139-145.

WANG H X, HU Y Y, DENG C.Research and implementation of face recognition algorithm based on LBP and ELM[J].Journal of Henan Polytechnic University(Natural Science) ,2021,40(5):139-145.

基于LBPELM的人脸识别算法研究与实现

王红星1, 胡永阳2, 邓超1

1.河南理工大学 物理与电子信息工程学院,河南 焦作 454000;2.河南理工大学 机械与动力工程学院,河南 焦作 454000

摘要:针对传统的局部二值模式(LBP)在人脸图像特征提取时容易受到灰度和噪声影响的问题,在传统LBP基础上提出一种改进的LBP算法。该算法计算邻域各像素与中心像素差值的平方和C,若C在限定范围内,选取中心像素值为阈值计算LBP值,充分考虑中心像素值与邻域像素值的作用,更准确描述局部图像特征;反之,若C不在限定范围内,则选择邻域像素和中心像素的中值作为阈值进行计算,降低噪声点的影响。利用主成分分析法(PCA)降低LBP 提取的人脸图像特征维数。为了解决普通极限学习机(ELM)的不足,引入加权共轭核极限学习机(WCGKELM)进行人脸图像特征的分类。经实验验证,该算法能有效提高人脸识别率。

关键词:加权共轭核极限学习机;局部二值模式;主成分分析;特征提取;人脸识别

doi:10.16186/j.cnki.1673-9787.2020040103

基金项目:河南省高等学校重点科研项目(15A510024);河南省科技攻关重点项目(182102311067

收稿日期:2020/04/28

修回日期:2020/06/15

出版日期:2021/09/15

Research and implementation of face recognition algorithm based on LBP and ELM

WANG Hongxing1, HU Yongyang2, DENG Chao1

1.School of Physics and Electronic Information Engineering Henan Polytechnic University Jiaozuo  454000 Henan China;2.School of Mechanical and Power Engineering Henan Polytechnic University Jiaozuo  454000 Henan China

Abstract:Aiming at the problem that traditional local binary pattern LBP is easy to be affected by gray and noise in face feature extraction an improved LBP algorithm was proposed based on traditional LBP. This meth-od calculated the sum of the squares of the difference between each pixel in the neighborhood and the center pixel C.lf the C was within a limited range the central pixel was selected as the threshold for calculation and the roles of the central pixel value and the domain pixel value was fully considered to describe the local features more accurately. Otherwise the median value of the neighborhood pixel and the center pixel was selected as the threshold value for comparison to reduce the influence of noise points. Principal component analysis PCA was used to reduce the face feature dimension extracted by LBP. In order solve to the shortcomings of the ordinary extreme learning machine ELM a weighted conjugate kernel extreme learning machineWCCKELM was introduced to classify the face features. Experimental results showed that the algorithm can effectively improve the face recognition rate.

Key words:weighted conjugate kernel extreme learning machine;local binary pattern;principal component analysis;face recognition feature extraction;

 基于LBP和ELM的人脸识别算法研究与实现_王红星.pdf

最近更新