Toward a Practical Face Recognition System: Robust Alignment and Illumination by Sparse Representation

被引:440
作者
Wagner, Andrew [1 ]
Wright, John [2 ]
Ganesh, Arvind [1 ]
Zhou, Zihan [1 ]
Mobahi, Hossein [3 ]
Ma, Yi [1 ,4 ]
机构
[1] Univ Illinois, Dept Elect & Comp Engn, Coordinated Sci Lab, Urbana, IL 61801 USA
[2] Columbia Univ, Dept Elect Engn, New York, NY 10027 USA
[3] Univ Illinois, Coordinated Sci Lab, Dept Comp Sci, Urbana, IL 61801 USA
[4] Microsoft Res Asia, Visual Comp Grp, Beijing, Peoples R China
关键词
Face recognition; face alignment; illumination variation; occlusion and corruption; sparse representation; error correction; validation and outlier rejection; STRONG UNIQUENESS; MODELS; IMAGE; CONVERGENCE; REFLECTANCE; OBJECTS; 3D;
D O I
10.1109/TPAMI.2011.112
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Many classic and contemporary face recognition algorithms work well on public data sets, but degrade sharply when they are used in a real recognition system. This is mostly due to the difficulty of simultaneously handling variations in illumination, image misalignment, and occlusion in the test image. We consider a scenario where the training images are well controlled and test images are only loosely controlled. We propose a conceptually simple face recognition system that achieves a high degree of robustness and stability to illumination variation, image misalignment, and partial occlusion. The system uses tools from sparse representation to align a test face image to a set of frontal training images. The region of attraction of our alignment algorithm is computed empirically for public face data sets such as Multi-PIE. We demonstrate how to capture a set of training images with enough illumination variation that they span test images taken under uncontrolled illumination. In order to evaluate how our algorithms work under practical testing conditions, we have implemented a complete face recognition system, including a projector-based training acquisition system. Our system can efficiently and effectively recognize faces under a variety of realistic conditions, using only frontal images under the proposed illuminations as training.
引用
收藏
页码:372 / 386
页数:15
相关论文
共 45 条
[1]   Face description with local binary patterns:: Application to face recognition [J].
Ahonen, Timo ;
Hadid, Abdenour ;
Pietikainen, Matti .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2006, 28 (12) :2037-2041
[2]  
Amberg B., 2007, PROC IEEE INT C COMP, P1
[3]  
[Anonymous], 1994, IEEE C COMP VIS PATT
[4]  
[Anonymous], P IEEE C AUT FAC GES
[5]  
[Anonymous], P IEEE C COMP VIS PA
[6]  
[Anonymous], 2008, P NEUR INF PROC SYST
[7]  
[Anonymous], P INT C IM PROC
[8]  
[Anonymous], 1981, IJCAI 81 7 INT JOINT
[9]  
[Anonymous], P BRIT MACH VIS C
[10]  
[Anonymous], PROXIMAL METHOD COMP