基于RPCA和卷积神经网络的光照人脸识别  被引量:1

Illumination face recognition based on RPCA and convolutional neural network

在线阅读下载全文

作  者:严培宏 李雷[1] YAN Peihong;LI Lei(School of Science,Nanjing University of Posts and Telecommunications,Nanjing 210023,China)

机构地区:[1]南京邮电大学理学院,江苏南京210023

出  处:《电视技术》2018年第8期1-4,14,共5页Video Engineering

基  金:国家自然科学基金(61070234;61071167;61373137)

摘  要:针对传统的人脸识别中的特征提取会受光照、姿态、遮挡等影响,提出了基于RPCA和卷积神经网络的人脸识别算法。该算法将鲁棒主成分分析(Robust Principal Component Analysis,RPCA)与卷积神经网络相结合,来增强人脸的识别效果。首先对人脸样本YaleB进行划分,并利用鲁棒主成分分析提取出训练样本和测试样本的低秩矩阵和稀疏矩阵,最后利用卷积神经网络(Convolutional Neural Network,CNN)对训练样本的低秩矩阵进行识别分类。该算法在YaleB人脸库中进行测试,实验表明:提出的基于RPCA和卷积神经网络的光照人脸识别算法较RPCA和CNN具有更好的鲁棒性。For the traditional feature extraction in face recognition will be affected by light,pose,occlusion,etc.,a face recogni-tion algoritlim based on RPCA and convolutional neural networlc is proposed.The algoritlim uses low-rank robust princijDal compo-nent analysis(Robust PrincijDal Component Analysis,RPCA)is combined with a convolutional neural network to enhance the rec-ognition of human faces.Firstly!the face sample YaleB is divided!and the low rank matrix and sparse matrix of training samples and test samples are extracted by robust principal component analysis.Finally,the lowrank of the training sample is obtained by using a convolutional neural network(CNN).The matrix identifies and classifies.The algorithm is tested in YaleB face database.Experiments showthat the proposed illumination face recognition algorithm based on RPCA and convolutional neural network has better robustness than RPCAand CNN.

关 键 词:RPCA 卷积神经网络 人脸识别 

分 类 号:TP391.41[自动化与计算机技术—计算机应用技术]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象