Heterogeneous Face Recognition by Margin-Based Cross-Modality Metric Learning

Huo, Jing; Shi, Yinghuan; Yang, Wanqi; Yin, Hujun

doi:10.1109/tcyb.2017.2715660

Cited by 45 publications

(13 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Heterogeneous face recognition [16]- [20] is a wide and important application in heterogeneous data field, which aims to match face images from different modalities or data sources. A margin-based cross-modality metric learning (MCM 2 L) [22] method is presented for heterogeneous face recognition task, which can learn a projective metric and incorporate other features. An asymmetric joint learning (AJL) [8] method is introduced to solve HFR problem, which can transforms the cross-modality differences into the learning process and reduce the diversity for inter-classes.…”

Section: B Heterogeneous Datasetsmentioning

confidence: 99%

Person Re-Identification With Character-Illustration-Style Image and Normal Photo

Jing

Cheng

et al. 2021

IEEE Access

View full text Add to dashboard Cite

Retrieving the given objects hidden amidst the gallery set is important for public safety and decision-making. Heterogeneous pedestrian retrieval (person re-identification) aims to retrieve the same person images from different modality set for identification. To address this problem, we contribute a new character-illustration-style image and normal photo pedestrian re-identification dataset (CINPID), which is collected on campus. The CINPID dataset includes two modalities, i.e., normal photos captured by one camera and character-illustration-style images drawn by the painter. To handle the problem of pedestrian retrieval with character-illustration-style images and normal photos , we propose a semi-coupled mapping and discriminant dictionary learning (SMD 2 L), which can learn a semi-coupled mapping matrix and dictionary pair from heterogeneous samples. With the learned semi-coupled mapping matrix, the differences between heterogeneous data can be reduced to some extent. Experimental results on the new CINPID dataset show that our approach outperforms the compared competing methods.INDEX TERMS Heterogeneous person re-identification, heterogeneous pedestrian dataset, semi-coupled mapping.

show abstract

Section: B Heterogeneous Datasetsmentioning

confidence: 99%

Person Re-Identification With Character-Illustration-Style Image and Normal Photo

Jing

Cheng

et al. 2021

IEEE Access

View full text Add to dashboard Cite

show abstract

“…the 48 composite sketch-photo pairs from E-PRIP and PRIP-VSGC database are randomly selected as the training set, the rest pairs are the testing set. NJU-ID database [Huo et al, 2017] contains 256 persons. For each person, there are one card image with low resolution and one face photo from a high resolution digital camera.…”

Section: Databasesmentioning

confidence: 99%

“…A multi-view discriminant analysis (MvDA) method [Kan et al, 2016] was proposed to exploit both inter-view and intra-view correlations of heterogeneous face images. [Huo et al, 2017] proposed a margin based cross-modality metric learning to address the gap of different modalities. Yet the projection procedure may losses some discriminative information.…”

Section: Introductionmentioning

confidence: 99%

Deep Attribute Guided Representation for Heterogeneous Face Recognition

Liu

Wang

Peng

et al. 2018

Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence

View full text Add to dashboard Cite

Heterogeneous face recognition (HFR) is a challenging problem in face recognition, subject to large texture and spatial structure differences of face images. Different from conventional face recognition in homogeneous environments, there exist many face images taken from different sources (including different sensors or different mechanisms) in reality. Motivated by human cognitive mechanism, we naturally utilize the explicit invariant semantic information (face attributes) to help address the gap of different modalities. Existing related face recognition methods mostly regard attributes as the high level feature integrated with other engineering features enhancing recognition performance, ignoring the inherent relationship between face attributes and identities. In this paper, we propose a novel deep attribute guided representation based heterogeneous face recognition method (DAG-HFR) without labeling attributes manually. Deep convolutional networks are employed to directly map face images in heterogeneous scenarios to a compact common space where distances mean similarities of pairs. An attribute guided triplet loss (AGTL) is designed to train an end-to-end HFR network which could effectively eliminate defects of incorrectly detected attributes. Extensive experiments on multiple heterogeneous scenarios (composite sketches, resident ID cards) demonstrate that the proposed method achieves superior performances compared with state-of-the-art methods.

show abstract

“…Unfortunately, due to the large domain discrepancy, the performance of the recognition network trained on VIS images often degrades dramatically in such a heterogeneous case [3]. Other HFR tasks also include Sketch-Photo [4], Profile-Frontal Photo [5], Thermal-VIS [6], and ID-Camera [7]. In order to bridge the domain discrepancy between heterogeneous data, researchers have put substantial efforts to match cross-domain features.…”

Section: Introductionmentioning

confidence: 99%

DVG-Face: Dual Variational Generation for Heterogeneous Face Recognition

Fu¹,

Wang²,

Hu³

et al. 2020

Preprint

View full text Add to dashboard Cite

Heterogeneous Face Recognition (HFR) refers to matching cross-domain faces, playing a crucial role in public security. Nevertheless, HFR is confronted with the challenges from large domain discrepancy and insufficient heterogeneous data. In this paper, we formulate HFR as a dual generation problem, and tackle it via a novel Dual Variational Generation (DVG-Face) framework. Specifically, a dual variational generator is elaborately designed to learn the joint distribution of paired heterogeneous images. However, the small-scale paired heterogeneous training data may limit the identity diversity of sampling. With this in mind, we propose to integrate abundant identity information of large-scale VIS images into the joint distribution. Furthermore, a pairwise identity preserving loss is imposed on the generated paired heterogeneous images to ensure their identity consistency. As a consequence, massive new diverse paired heterogeneous images with the same identity can be generated from noises. The identity consistency and diversity properties allow us to employ these generated images to train the HFR network via a contrastive learning mechanism, yielding both domain invariant and discriminative embedding features. Concretely, the generated paired heterogeneous images are regarded as positive pairs, and the images obtained from different samplings are considered as negative pairs. Our method achieves superior performances over state-of-the-art methods on seven databases belonging to five HFR tasks, including NIR-VIS, Sketch-Photo, Profile-Frontal Photo, Thermal-VIS, and ID-Camera. The related code will be released at https://github.com/BradyFU.

show abstract

Heterogeneous Face Recognition by Margin-Based Cross-Modality Metric Learning

Cited by 45 publications

References 39 publications

Person Re-Identification With Character-Illustration-Style Image and Normal Photo

Person Re-Identification With Character-Illustration-Style Image and Normal Photo

Deep Attribute Guided Representation for Heterogeneous Face Recognition

DVG-Face: Dual Variational Generation for Heterogeneous Face Recognition

Contact Info

Product

Resources

About