[1]陈飞玥,朱玉莲,陈晓红.多层特征融合的PCANet及其在人脸识别中的应用[J].南京师大学报(自然科学版),2021,44(02):127-133.[doi:10.3969/j.issn.1001-4616.2021.02.018]
　Chen Feiyue,Zhu Yulian,Chen Xiaohong.Multi-stage Feature Fusion PCANet and Its Application to Face Recognition[J].Journal of Nanjing Normal University(Natural Science Edition),2021,44(02):127-133.[doi:10.3969/j.issn.1001-4616.2021.02.018]
点击复制

多层特征融合的PCANet及其在人脸识别中的应用()
分享到：

《南京师大学报（自然科学版）》[ISSN:1001-4616/CN:32-1239/N]

卷:
第44卷
期数:
2021年02期
页码:
127-133
栏目:
·计算机科学与技术·
出版日期:
2021-06-30

文章信息/Info

Title:
Multi-stage Feature Fusion PCANet and Its Application to Face Recognition
文章编号:
1001-4616(2021)02-0127-07
作者:
陈飞玥1; 朱玉莲2; 陈晓红3
(1.南京航空航天大学计算机科学与技术学院,江苏 南京 211106)(2.南京航空航天大学公共实验教学部,江苏 南京 211106)(3.南京航空航天大学理学院,江苏 南京 211106)
Author(s):
Chen Feiyue1; Zhu Yulian2; Chen Xiaohong3
(1.College of Computer Science and Technology,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China)(2.Fundamental Experimental Teaching Department,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China)(3.College of Science,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China)
关键词:
人脸识别; 主成分分析网络; 密集网络; 特征融合; 多层特征融合的PCANet
Keywords:
face recognition; PCANet; DenseNet; feature fusion; PCANet_dense
分类号:
TP391
DOI:
10.3969/j.issn.1001-4616.2021.02.018
文献标志码:
A
摘要:
主成分分析网络(PCANet)是一种简单的深度学习网络模型,在图像识别领域具有很强的应用潜力. 本文在PCANet的基础上,通过对PCANet结构进行分析,构造了一种基于多层特征融合的PCANet(PCANet_dense)网络模型. 与单纯地只将前一层网络输出作为后一层网络输入的PCANet不同,PCANet_dense利用了不同层的特征信息. 在2层网络结构中,它首先将原始图像特征和第1层网络的输出进行级联,然后将级联后的结果作为第2层网络的输入. 而在3层网络结构中,它则将第1层和第2层网络的输出级联起来,作为第3层网络的输入. 由于PCANet_dense在训练每一层(除了第1层)时使用了更多信息,因此能够获得比原PCANet更好的效果. 为了验证所提方法的有效性,本文使用CMU PIE数据集构建网络模型,并在ORL、AR和Extended Yale B 3个公开人脸数据集上对所提出方法的性能进行了测试,实验结果表明,本文提出的PCANet_dense获得了比PCANet更好的性能.
Abstract:
Principal component analysis network(PCANet)is a simple deep learning network model but shows strong application potential in many fields,especially in image recognition. In this paper,a new multi-stage feature fusion model based PCANet(PCANet_dense)is constructed by analyzing the structure of PCANet. Different from PCANet which only takes the output of the former layer as the input of the latter layer,PCANet_dense uses the feature information of different layers. In the two-layer network structure,it firstly cascades the original image features with the output of the first layer network,and then takes the fused features as the input of the second layer network. In the three-layer network structure,it cascades the output of the first and the second layer in series as the input of the third layer network. Due to the use of more information in each layer(except the first layer)in training process,the new model gets good performance. In order to verify the effectiveness of the proposed method,several comparative experiments on different face datasets,such as ORL,AR and Extended Yale B,are established. The network model of PCANet and PCANet_dense are well constructed on CMU PIE dataset before all the experiments. The experimental results show that the proposed multi-stage feature fusion PCANet method achieves better performance than PCANet.

参考文献/References:

[1] LIU C,WECHSLER H. Gabor feature based classification using the enhanced fisher linear discriminant model for face recognition[J]. IEEE transactions on image processing,2002,11(4):467-476.
[2]AHONEN T,HADID A,PIETIKAINEN M. Face description with local binary patterns:application to face recognition[J]. IEEE transactions on pattern analysis and machine intelligence,2006,28(12):2037-2041.
[3]LOWE D G. Object recognition from local scale-invariant features[C]//International Conference on Computer Vision(ICCV). Kerkyra,Greece:IEEE,1999:1150-1157.
[4]DALAL N,TRIGGS B. Histograms of oriented gradients for human detection[C]//Computer Vision and Pattern Recognition(CVPR). San Diego,CA,USA:IEEE,2005:886-893.
[5]LEI Z,YI D,LI S Z. Learning stacked image descriptor for face recognition[J]. IEEE transactions on circuits and systems for video technology,2015,26(9):1685-1696.
[6]LECUN Y,BOTTOU L. Gradient-based learning applied to recognition[J]. Proceedings of the IEEE,1998,86(11):2278-2324.
[7]HE K,ZHANG X,REN S,et al. Deep residual learning for image recognition[C]//Computer Vision and Pattern Recognition(CVPR). Las Vegas,NV,USA:IEEE,2016:770-778.
[8]TAIGMAN Y,MING Y,RANZATO M,et al. DeepFace:closing the gap to human-level performance in face verification[C]//Computer Vision and Pattern Recognition(CVPR). Columbus,OH,USA:IEEE,2014:1701-1708.
[9]SCHROFF F,KALENICHENKO D,PHILBIN J. FaceNet:a unified embedding for face recognition and clustering[C]//Computer Vision and Pattern Recognition(CVPR). Boston,MA,USA:IEEE,2015:815-823.
[10]CHAN T,JIA K,GAO S,et al. PCANet:a simple deep learning baseline for image classification[J]. IEEE transactions on image processing,2015,24(12):5017-5032.
[11]XI M,CHEN L,POLAJNAR D,et al. Local binary pattern network:a deep learning approach for face recognition[C]//International Conference on Image Processing(ICIP). Phoenix,AZ,USA:IEEE,2016:3224-3228.
[12]LI S,YU B,WU W,et al. Feature learning based on SAE-PCA network for human gesture recognition in RGBD images[J]. Neurocomputing,2015,151:565-573.
[13]HUANG G,LIU Z,LAURENS V,et al. Densely connected convolutional networks[C]//Computer Vision and Pattern Recognition(CVPR). Honolulu,HI,USA:IEEE,2017:2261-2269.
[14]LEE K,HO J,KRIEGMAN D J. Acquiring linear subspaces for face recognition under variable lighting[J]. IEEE transactions on pattern analysis and machine intelligence,2005,27(5):684-698.

相似文献/References:

[1]朱长水,丁 勇,袁宝华,等.融合LBP和LPQ的人脸识别[J].南京师大学报(自然科学版),2015,38(01):104.
　Zhu Changshui,Ding Yong,Yuan Baohua,et al.Face Recognition Based on Local Binary Patternand Local Phase Quantization[J].Journal of Nanjing Normal University(Natural Science Edition),2015,38(02):104.
[2]沈学华,詹永照,程显毅,等.基于样本融合的核稀疏人脸识别方法[J].南京师大学报(自然科学版),2016,39(04):0.[doi:10.3969/j.issn.1001-4616.2016.04.007]
　Shen Xuehua,Zhan Yongzhao,Cheng Xianyi,et al.A Kernel Sparse Representation Method Based onSamples Fusion for Face Recognition[J].Journal of Nanjing Normal University(Natural Science Edition),2016,39(02):0.[doi:10.3969/j.issn.1001-4616.2016.04.007]
[3]朱 炯.核偏最小二乘回归在面部表情识别中的应用[J].南京师大学报(自然科学版),2018,41(03):14.[doi:10.3969/j.issn.1001-4616.2018.03.003]
　Zhu Jiong.Kernel Partial Least Squares Regression with Applicationsto Facial Expression Recognition[J].Journal of Nanjing Normal University(Natural Science Edition),2018,41(02):14.[doi:10.3969/j.issn.1001-4616.2018.03.003]
[4]顼 聪,朱 毅,陶永鹏.面向方向选择的差值局部方向模式的人脸识别[J].南京师大学报(自然科学版),2020,43(04):113.[doi:10.3969/j.issn.1001-4616.2020.04.016]
　Xu Cong,Zhu Yi,Tao Yongpeng.Face Recognition Based on Direction-SelectedDifference Local Direction Pattern[J].Journal of Nanjing Normal University(Natural Science Edition),2020,43(02):113.[doi:10.3969/j.issn.1001-4616.2020.04.016]

备注/Memo

备注/Memo:
收稿日期:2020-08-08.
基金项目:国家自然科学基金项目(61703206).
通讯作者:朱玉莲,副教授,研究方向:模式识别与人工智能. E-mail:lianyi_1999@nuaa.edu.cn
更新日期/Last Update: 2021-06-30