Pure Detail Feature Extraction Network for Visible-Infrared Re-Identification

被引:2
作者
Cui, Jiaao [1 ]
Chan, Sixian [1 ,2 ]
Mu, Pan [1 ]
Tang, Tinglong [2 ]
Zhou, Xiaolong [3 ]
机构
[1] Zhejiang Univ Technol, Hangzhou 310023, Peoples R China
[2] China Three Gorges Univ, Coll Comp & Informat, Hubei Key Lab Intelligent Vis Based Monitoring Hyd, Yichang 443002, Peoples R China
[3] Quzhou Univ, Coll Elect & Informat Engn, Quzhou 324000, Peoples R China
基金
中国国家自然科学基金;
关键词
Person re-identification; multimedia; image retrieval; PERSON REIDENTIFICATION; IMAGES;
D O I
10.32604/iasc.2023.039894
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Cross-modality pedestrian re-identification has important applications in the field of surveillance. Due to variations in posture, camera perspective, and camera modality, some salient pedestrian features are difficult to provide effective retrieval cues. Therefore, it becomes a challenge to design an effective strategy to extract more discriminative pedestrian detail. Although many effective methods for detailed feature extraction are proposed, there are still some shortcomings in filtering background and modality noise. To further purify the features, a pure detail feature extraction network (PDFENet) is proposed for VI-ReID. PDFENet includes three modules, adaptive detail mask generation module (ADMG), inter-detail interaction module (IDI) and cross-modality cross-entropy (CMCE). ADMG and IDI use human joints and their semantic associations to suppress background noise in features. CMCE guides the model to ignore modality noise by generating modalityshared feature labels. Specifically, ADMG generates masks for pedestrian details based on pose estimation. Masks are used to suppress background information and enhance pedestrian detail information. Besides, IDI mines the semantic relations among details to further refine the features. Finally, CMCE cross-combines classifiers and features to generate modality-shared feature labels to guide model training. Extensive ablation experiments as well as visualization results have demonstrated the effectiveness of PDFENet in eliminating background and modality noise. In addition, comparison experiments in two publicly available datasets also show the competitiveness of our approach.
引用
收藏
页码:2263 / 2277
页数:15
相关论文
共 46 条
[1]   2D Human Pose Estimation: New Benchmark and State of the Art Analysis [J].
Andriluka, Mykhaylo ;
Pishchulin, Leonid ;
Gehler, Peter ;
Schiele, Bernt .
2014 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2014, :3686-3693
[2]  
[Anonymous], 2006, 2006 IEEE COMP SOC C
[3]   Structure-Aware Positional Transformer for Visible-Infrared Person Re-Identification [J].
Chen, Cuiqun ;
Ye, Mang ;
Qi, Meibin ;
Wu, Jingjing ;
Jiang, Jianguo ;
Lin, Chia-Wen .
IEEE TRANSACTIONS ON IMAGE PROCESSING, 2022, 31 :2352-2364
[4]   Learning a similarity metric discriminatively, with application to face verification [J].
Chopra, S ;
Hadsell, R ;
LeCun, Y .
2005 IEEE COMPUTER SOCIETY CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, VOL 1, PROCEEDINGS, 2005, :539-546
[5]   Healthcare system: Moving forward with artificial intelligence [J].
Dicuonzo, Grazia ;
Donofrio, Francesca ;
Fusco, Antonio ;
Shini, Matilda .
TECHNOVATION, 2023, 120
[6]   MSO: Multi-Feature Space Joint Optimization Network for RGB-Infrared Person Re-Identification [J].
Gao, Yajun ;
Liang, Tengfei ;
Jin, Yi ;
Gu, Xiaoyan ;
Liu, Wu ;
Li, Yidong ;
Lang, Congyan .
PROCEEDINGS OF THE 29TH ACM INTERNATIONAL CONFERENCE ON MULTIMEDIA, MM 2021, 2021, :5257-5265
[7]   Cross-Modality Person Re-Identification via Modality Confusion and Center Aggregation [J].
Hao, Xin ;
Zhao, Sanyuan ;
Ye, Mang ;
Shen, Jianbing .
2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021), 2021, :16383-16392
[8]  
He ZJ, 2022, Arxiv, DOI arXiv:2111.14411
[9]  
Hu J, 2018, PROC CVPR IEEE, P7132, DOI [10.1109/TPAMI.2019.2913372, 10.1109/CVPR.2018.00745]
[10]  
Huang N, 2022, arXiv