Stacked Progressive Auto-Encoders (SPAE) for Face Recognition Across Poses

被引:166
作者
Kan, Meina [1 ]
Shan, Shiguang [1 ]
Chang, Hong [1 ]
Chen, Xilin [1 ]
机构
[1] Chinese Acad Sci, Inst Comp Technol, Key Lab Intelligent Informat Proc, Beijing 100190, Peoples R China
来源
2014 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR) | 2014年
关键词
D O I
10.1109/CVPR.2014.243
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Identifying subjects with variations caused by poses is one of the most challenging tasks in face recognition, since the difference in appearances caused by poses may be even larger than the difference due to identity. Inspired by the observation that pose variations change non-linearly but smoothly, we propose to learn pose-robust features by modeling the complex non-linear transform from the non-frontal face images to frontal ones through a deep network in a progressive way, termed as stacked progressive auto-encoders (SPAE). Specifically, each shallow progressive auto-encoder of the stacked network is designed to map the face images at large poses to a virtual view at smaller ones, and meanwhile keep those images already at smaller poses unchanged. Then, stacking multiple these shallow auto-encoders can convert non-frontal face images to frontal ones progressively, which means the pose variations are narrowed down to zero step by step. As a result, the outputs of the topmost hidden layers of the stacked network contain very small pose variations, which can be used as the pose-robust features for face recognition. An additional attractiveness of the proposed method is that no pose estimation is needed for the test images. The proposed method is evaluated on two datasets with pose variations, i.e., MultiPIE and FERET datasets, and the experimental results demonstrate the superiority of our method to the existing works, especially to those 2D ones.
引用
收藏
页码:1883 / 1890
页数:8
相关论文
共 28 条
[1]  
[Anonymous], IEEE C COMP VIS PATT
[2]  
[Anonymous], 2008, CVPR, DOI DOI 10.1109/CVPR.2008.4587754
[3]  
[Anonymous], 2007, TECHNICAL REPORT
[4]   Energy Normalization for Pose-Invariant Face Recognition Based on MRF Model Image Matching [J].
Arashloo, Shervin Rahimzadeh ;
Kittler, Josef .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2011, 33 (06) :1274-1280
[5]  
Asthana A, 2011, IEEE I CONF COMP VIS, P937, DOI 10.1109/ICCV.2011.6126336
[6]   Eigenfaces vs. Fisherfaces: Recognition using class specific linear projection [J].
Belhumeur, PN ;
Hespanha, JP ;
Kriegman, DJ .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 1997, 19 (07) :711-720
[7]  
Bengio Yoshua, 2012, Neural Networks: Tricks of the Trade. Second Edition: LNCS 7700, P437, DOI 10.1007/978-3-642-35289-8_26
[8]   Learning Deep Architectures for AI [J].
Bengio, Yoshua .
FOUNDATIONS AND TRENDS IN MACHINE LEARNING, 2009, 2 (01) :1-127
[9]  
BEYMER D, 1995, FIFTH INTERNATIONAL CONFERENCE ON COMPUTER VISION, PROCEEDINGS, P500, DOI 10.1109/ICCV.1995.466898
[10]   Face recognition based on fitting a 3D morphable model [J].
Blanz, V ;
Vetter, T .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2003, 25 (09) :1063-1074