Locality-constrained linear coding based bi-layer model for multi-view facial expression recognition.pdfVIP

Neurocomputing 239 (2017) 143–152 Contents lists available at ScienceDirect Neurocomputing journal homepage: /locate/neucom Locality-constrained linear coding based bi-layer model for multi-view facial expression recognition Jianlong Wu a,b, Zhouchen Lin a,b,?, Wenming Zheng c, Hongbin Zha a,b a Key Laboratory of Machine Perception (MOE), School of EECS, Peking University, Beijing, 100871, P.R. China b Cooperative Medianet Innovation Center, Shanghai Jiao Tong University, Shanghai, 200240, P.R. China c Key Laboratory of Child Development and Learning Science (MOE), Research Center for Learning Science, Southeast University, Jiangsu, 210096, P.R. China article info Article history: Received 7 May 2016 Revised 2 October 2016 Accepted 3 February 2017 Available online 9 February 2017 Communicated by Prof. Zidong Wang Keywords: Multi-view facial expression recognition Locality-constrained linear coding based bi-layer model Bag-of-features abstract Multi-view facial expression recognition is a challenging and active research area in computer vision. In this paper, we propose a simple yet effective method, called the locality-constrained linear coding based bi-layer (LLCBL) model, to learn discriminative representation for multi-view facial expression recognition. To address the issue of large pose variations, locality-constrained linear coding is adopted to construct an overall bag-of-features model, which is then used to extract overall features as well as estimate poses in the ?rst layer. In the second layer, we establish one speci?c view-dependent model for each view, respectively. After the pose information of the facial image is known, we use the corresponding view-dependent model in the second layer to further extract features. By combining all the features in these two layers, we obtain a uni?ed representation of the image. To evaluate the proposed approach, we conduct extensive experiments on both BU-3DFE and Multi-PIE databases. Experimental results show that our