Relational Deep Feature Learning for Heterogeneous Face Recognition

被引:35
作者
Cho, MyeongAh [1 ]
Kim, Taeoh [1 ]
Kim, Ig-Jae [2 ]
Lee, Kyungjae [3 ]
Lee, Sangyoun [1 ]
机构
[1] Yonsei Univ, Sch Elect & Elect Engn, Seoul 03722, South Korea
[2] Korea Inst Sci & Technol, Ctr Imaging Media Res, Seoul 02792, South Korea
[3] Yong Univ, Dept Comp Sci, Yongin 17092, South Korea
基金
新加坡国家研究基金会;
关键词
Heterogeneous face recognition; relation embedding; graph structured module; face recognition; DISCRIMINANT-ANALYSIS; REPRESENTATION; NETWORKS;
D O I
10.1109/TIFS.2020.3013186
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
Heterogeneous Face Recognition (HFR) is a task that matches faces across two different domains such as visible light (VIS), near-infrared (NIR), or the sketch domain. Due to the lack of databases, HFR methods usually exploit the pre-trained features on a large-scale visual database that contain general facial information. However, these pre-trained features cause performance degradation due to the texture discrepancy with the visual domain. With this motivation, we propose a graph-structured module called Relational Graph Module (RGM) that extracts global relational information in addition to general facial features. Because each identity's relational information between intra-facial parts is similar in any modality, the modeling relationship between features can help cross-domain matching. Through the RGM, relation propagation diminishes texture dependency without losing its advantages from the pre-trained features. Furthermore, the RGM captures global facial geometrics from locally correlated convolutional features to identify long-range relationships. In addition, we propose a Node Attention Unit (NAU) that performs node-wise recalibration to concentrate on the more informative nodes arising from relation-based propagation. Furthermore, we suggest a novel conditional-margin loss function (C-softmax) for the efficient projection learning of the embedding vector in HFR. The proposed method outperforms other state-of-the-art methods on five HFR databases. Furthermore, we demonstrate performance improvement on three backbones because our module can be plugged into any pre-trained face recognition backbone to overcome the limitations of a small HFR database.
引用
收藏
页码:376 / 388
页数:13
相关论文
共 68 条
[1]  
[Anonymous], 2016, LECT NOTES COMPUT SC, DOI DOI 10.1007/978-3-319-46487-9_6
[2]  
[Anonymous], 2018, 32 AAAI C ART INT
[3]  
[Anonymous], 2016, PROC CVPR IEEE, DOI DOI 10.1109/CVPR.2016.90
[4]  
[Anonymous], 2016, PROC CVPR IEEE, DOI DOI 10.1109/CVPR.2016.527
[5]  
[Anonymous], 2016, LECT NOTES COMPUT SC, DOI DOI 10.1007/978-3-319-49409-8_40
[6]  
[Anonymous], 2018, PROC CVPR IEEE, DOI DOI 10.1109/CVPR.2018.00813
[7]  
[Anonymous], 2015, IEEE I CONF COMP VIS, DOI DOI 10.1109/ICCV.2015.170
[8]  
[Anonymous], 2016, LECT NOTES COMPUT SC, DOI DOI 10.1007/978-3-319-46478-7_31
[9]  
[Anonymous], 2016, IMAGE VISION COMPUT, DOI DOI 10.1016/J.IMAVIS.2016.09.001
[10]  
[Anonymous], 2018, INT CONF BIOMETR, DOI DOI 10.1109/ICB2018.2018.00035