NIR-to-VIS Face Recognition via Embedding Relations and Coordinates of the Pairwise Features

Cho, MyeongAh; Chung, Tae‐Young; Kim, Taeoh; Lee, Sangyoun

doi:10.1109/icb45273.2019.8987306

Cited by 10 publications

(15 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We compared the our method with other deep learning methods, including TRIVET [19], IDR [20], ADFL [4], CDL [21], WCNN [7], RM [13], and RGM [15]. In Table 2, our PRAM Models CASIA NIR-VIS 2.0 [1] Rank-1 Acc.…”

Section: Comparison With Deep Learning Methodsmentioning

confidence: 99%

A NIR-to-VIS face recognition via part adaptive and relation attention module

Xu¹,

Cho²,

Lee³

2021

Preprint

Self Cite

View full text Add to dashboard Cite

In the face recognition application scenario, we need to process facial images captured in various conditions, such as at night by near-infrared (NIR) surveillance cameras. The illumination difference between NIR and visible-light (VIS) causes a domain gap between facial images, and the variations in pose and emotion also make facial matching more difficult. Heterogeneous face recognition (HFR) has difficulties in domain discrepancy, and many studies have focused on extracting domain-invariant features, such as facial part relational information. However, when pose variation occurs, the facial component position changes, and a different part relation is extracted. In this paper, we propose a part relation attention module that crops facial parts obtained through a semantic mask and performs relational modeling using each of these representative features. Furthermore, we suggest component adaptive triplet loss function using adaptive weights for each part to reduce the intra-class identity regardless of the domain as well as pose. Finally, our method exhibits a performance improvement in the CASIA NIR-VIS 2.0 [1] and achieves superior result in the BUAA-VisNir [2] with large pose and emotion variations.

show abstract

Section: Comparison With Deep Learning Methodsmentioning

confidence: 99%

A NIR-to-VIS face recognition via part adaptive and relation attention module

Xu¹,

Cho²,

Lee³

2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…We compared the our method with other deep learning methods, including TRIVET [19], IDR [20], ADFL [4], CDL [21], WCNN [7], RM [13], and RGM [16]. In Table 2, our PRAM performed better than the RM, which pairwise concatenated the feature vector with the addition of conditional triplet loss (L C ).…”

Section: Comparison With Deep Learning Methodsmentioning

confidence: 99%

“…In Table 3, our approach exhibits the best performance Models CASIA NIR-VIS 2.0 [1] Rank-1 Acc. (%) VR@FAR=0.1%(%) TRIVET [19] 95.7 78 IDR [20] 97.33 95.73 ADFL [4] 98.15 97.21 CDL [21] 98.62 98.32 WCNN [7] 98.7 98.4 RM [13] 94.73 94.31 RGM [16] 97 TRIVET [19] 93.9 80.9 IDR [20] 94.3 84.7 ADFL [4] 95.2 95.3 CDL [21] 96.9 95.9 WCNN [7] 97.4 96 RGM [16] 97. on the BUAA-VisNir database with a large variance in emotion and pose. Compared to the WCNN and ADFL, our method is slightly lower in the CASIA NIR-VIS 2.0 database but still demonstrate competitive performance, and higher performance with 2.04% and 4.24% in Buaa database.…”

Section: Comparison With Deep Learning Methodsmentioning

confidence: 99%

“…Chowdhury et al [11] used a bilinear CNN [12] that multiplied the convolutional layer output feature maps from two-stream CNNs. Cho et al [13] concatenated the feature vectors of the last feature map pair-wisely to extract the relations between two different facial parts. They both treated each vector of the feature map as representing a certain part of a facial component, such as the lips or nose.…”

Section: Part Relation Attention Module (Pram)mentioning

confidence: 99%

See 1 more Smart Citation

A Heterogeneous Face Recognition Via Part Adaptive And Relation Attention Module

Cho

Lee

2021

2021 IEEE International Conference on Image Processing (ICIP)

Self Cite

View full text Add to dashboard Cite

In the face recognition application scenario, we need to process facial images captured in various conditions, such as at night by near-infrared (NIR) surveillance cameras. The illumination difference between NIR and visible-light (VIS) images causes a domain gap, and the variations in pose and emotion also make facial matching more difficult. Since heterogeneous face recognition (HFR) has difficulties in domain discrepancy, many studies have focused on extracting domaininvariant features, such as facial part relational information. However, when pose variation occurs, the facial component position changes and a different part relation is extracted. In this paper, we propose a part relation attention module that crops facial parts obtained through a semantic mask and performs relational modeling using each of these representative features. Furthermore, we suggest component adaptive triplet loss using adaptive weights for each part to reduce the intraclass distance regardless of the domain as well as pose. Finally, our method exhibits a performance improvement in the CASIA NIR-VIS 2.0 [1] and achieves superior results in the BUAA-VisNir [2] with large pose and emotion variations.

show abstract

“…The feature match- ing is conducted on the latent subspace. Cho et al [3] presented a post-processing relation module to capture relations and coordinates of the pairwise feature to reduce the domain discrepancy. In addition, a triplet loss with a conditional margin is introduced to reduce intra-class variation in training.…”

Section: Related Workmentioning

confidence: 99%

DBLFace: Domain-Based Labels for NIR-VIS Heterogeneous Face Recognition

Kakadiaris

2020

Preprint

View full text Add to dashboard Cite

Deep learning-based domain-invariant feature learning methods are advancing in near-infrared and visible (NIR-VIS) heterogeneous face recognition. However, these methods are prone to overfitting due to the large intra-class variation and the lack of NIR images for training. In this paper, we introduce Domain-Based Label Face (DBLFace), a learning approach based on the assumption that a subject is not represented by a single label but by a set of labels. Each label represents images of a specific domain. In particular, a set of two labels per subject, one for the NIR images and one for the VIS images, are used for training a NIR-VIS face recognition model. The classification of images into different domains reduces the intra-class variation and lessens the negative impact of data imbalance in training. To train a network with sets of labels, we introduce a domain-based angular margin loss and a maximum angular loss to maintain the inter-class discrepancy and to enforce the close relationship of labels in a set. Quantitative experiments confirm that DBLFace significantly improves the rank-1 identification rate by 6.7% on the EDGE20 dataset and achieves state-of-the-art performance on the CASIA NIR-VIS 2.0 dataset.

show abstract

NIR-to-VIS Face Recognition via Embedding Relations and Coordinates of the Pairwise Features

Cited by 10 publications

References 15 publications

A NIR-to-VIS face recognition via part adaptive and relation attention module

A NIR-to-VIS face recognition via part adaptive and relation attention module

A Heterogeneous Face Recognition Via Part Adaptive And Relation Attention Module

DBLFace: Domain-Based Labels for NIR-VIS Heterogeneous Face Recognition

Contact Info

Product

Resources

About