Discriminative Shared Gaussian Processes for Multiview and View-Invariant Facial Expression Recognition

被引:170
作者
Eleftheriadis, Stefanos [1 ]
Rudovic, Ognjen [1 ]
Pantic, Maja [1 ,2 ]
机构
[1] Univ London Imperial Coll Sci Technol & Med, Dept Comp, London SW7 2AZ, England
[2] Univ Twente, Fac Elect Engn Math & Comp Sci, NL-7522 NB Enschede, Netherlands
关键词
View-invariant; multi-view learning; facial expression recognition; Gaussian Processes; TEXTURE CLASSIFICATION; SCALE;
D O I
10.1109/TIP.2014.2375634
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Images of facial expressions are often captured from various views as a result of either head movements or variable camera position. Existing methods for multiview and/or view-invariant facial expression recognition typically perform classification of the observed expression using either classifiers learned separately for each view or a single classifier learned for all views. However, these approaches ignore the fact that different views of a facial expression are just different manifestations of the same facial expression. By accounting for this redundancy, we can design more effective classifiers for the target task. To this end, we propose a discriminative shared Gaussian process latent variable model (DS-GPLVM) for multiview and view-invariant classification of facial expressions from multiple views. In this model, we first learn a discriminative manifold shared by multiple views of a facial expression. Subsequently, we perform facial expression classification in the expression manifold. Finally, classification of an observed facial expression is carried out either in the view-invariant manner (using only a single view of the expression) or in the multiview manner (using multiple views of the expression). The proposed model can also be used to perform fusion of different facial features in a principled manner. We validate the proposed DS-GPLVM on both posed and spontaneously displayed facial expressions from three publicly available datasets (MultiPIE, labeled face parts in the wild, and static facial expressions in the wild). We show that this model outperforms the state-of-the-art methods for multiview and view-invariant facial expression classification, and several state-of-the-art methods for multiview learning and feature fusion.
引用
收藏
页码:189 / 204
页数:16
相关论文
共 53 条
[1]   DISCRETE COSINE TRANSFORM [J].
AHMED, N ;
NATARAJAN, T ;
RAO, KR .
IEEE TRANSACTIONS ON COMPUTERS, 1974, C 23 (01) :90-93
[2]  
[Anonymous], 2005, MG STAT PRO
[3]  
[Anonymous], 2011, INT C MACHINE LEARNI
[4]  
[Anonymous], 2003, UNMASKING FACE GUIDE
[5]  
[Anonymous], 2007, Proceedings of the 6th ACM international conference on Image and video retrieval, DOI [10.1145/1282280.1282340, DOI 10.1145/1282280.1282340]
[6]  
[Anonymous], 2006, P 23 INT C MACHINE L, DOI DOI 10.1145/1143844.1143909
[7]  
[Anonymous], 2008, IEEE INT C PATT REC
[8]  
Belhumeur PN, 2011, PROC CVPR IEEE, P545, DOI 10.1109/CVPR.2011.5995602
[9]  
Bertsekas D. P., 2014, Constrained Optimization and Lagrange Multiplier Methods
[10]  
Bishop C, 2006, PATTERN RECOGN