Progressive Discriminative Feature Learning for Visible-Infrared Person Re-Identification

被引:0
|
作者
Zhou, Feng [1 ]
Cheng, Zhuxuan [2 ]
Yang, Haitao [1 ]
Song, Yifeng [2 ]
Fu, Shengpeng [2 ]
机构
[1] Hunan Police Acad, Dept Criminal Ivest, Changsha 410138, Peoples R China
[2] Chinese Acad Sci, Shenyang Inst Automat, Shenyang 110016, Peoples R China
基金
中国国家自然科学基金;
关键词
visible-infrared person re-identification; cross-modality; deep learning; instance normalization;
D O I
10.3390/electronics13142825
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
The visible-infrared person re-identification (VI-ReID) task aims to retrieve the same pedestrian between visible and infrared images. VI-ReID is a challenging task due to the huge modality discrepancy and complex intra-modality variations. Existing works mainly complete the modality alignment at one stage. However, aligning modalities at different stages has positive effects on the intra-class and inter-class distances of cross-modality features, which are often ignored. Moreover, discriminative features with identity information may be corrupted in the processing of modality alignment, further degrading the performance of person re-identification. In this paper, we propose a progressive discriminative feature learning (PDFL) network that adopts different alignment strategies at different stages to alleviate the discrepancy and learn discriminative features progressively. Specifically, we first design an adaptive cross fusion module (ACFM) to learn the identity-relevant features via modality alignment with channel-level attention. For well preserving identity information, we propose a dual-attention-guided instance normalization module (DINM), which can well guide instance normalization to align two modalities into a unified feature space through channel and spatial information embedding. Finally, we generate multiple part features of a person to mine subtle differences. Multi-loss optimization is imposed during the training process for more effective learning supervision. Extensive experiments on the public datasets of SYSU-MM01 and RegDB validate that our proposed method performs favorably against most state-of-the-art methods.
引用
收藏
页数:15
相关论文
共 50 条
  • [41] Dual-Semantic Consistency Learning for Visible-Infrared Person Re-Identification
    Zhang, Yiyuan
    Kang, Yuhao
    Zhao, Sanyuan
    Shen, Jianbing
    IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY, 2023, 18 : 1554 - 1565
  • [42] Learning Modality-Specific Representations for Visible-Infrared Person Re-Identification
    Feng, Zhanxiang
    Lai, Jianhuang
    Xie, Xiaohua
    IEEE TRANSACTIONS ON IMAGE PROCESSING, 2020, 29 : 579 - 590
  • [43] Modality-agnostic learning for robust visible-infrared person re-identification
    Gong, Shengrong
    Li, Shuomin
    Xie, Gengsheng
    Yao, Yufeng
    Zhong, Shan
    SIGNAL IMAGE AND VIDEO PROCESSING, 2025, 19 (03)
  • [44] Cross-modality consistency learning for visible-infrared person re-identification
    Shao, Jie
    Tang, Lei
    JOURNAL OF ELECTRONIC IMAGING, 2022, 31 (06)
  • [45] Multi-Stage Auxiliary Learning for Visible-Infrared Person Re-Identification
    Zhang, Huadong
    Cheng, Shuli
    Du, Anyu
    IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, 2024, 34 (11) : 12032 - 12047
  • [46] Attention-enhanced feature mapping network for visible-infrared person re-identification
    Liu, Shuaiyi
    Han, Ke
    MACHINE VISION AND APPLICATIONS, 2025, 36 (02)
  • [47] Visible-infrared person re-identification model based on feature consistency and modal indistinguishability
    Sun, Jia
    Li, Yanfeng
    Chen, Houjin
    Peng, Yahui
    Zhu, Jinlei
    MACHINE VISION AND APPLICATIONS, 2023, 34 (01)
  • [48] TWO-PHASE FEATURE FUSION NETWORK FOR VISIBLE-INFRARED PERSON RE-IDENTIFICATION
    Cheng, Yunzhou
    Xiao, Guoqiang
    Tang, Xiaoqin
    Ma, Wenzhuo
    Gou, Xinye
    2021 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP), 2021, : 1149 - 1153
  • [49] FMCNet: Feature-Level Modality Compensation for Visible-Infrared Person Re-Identification
    Zhang, Qiang
    Lai, Changzhou
    Liu, Jianan
    Huang, Nianchang
    Han, Jungong
    2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2022, : 7339 - 7348
  • [50] Modality Unifying Network for Visible-Infrared Person Re-Identification
    Yu, Hao
    Cheng, Xu
    Peng, Wei
    Liu, Weihao
    Zhao, Guoying
    2023 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2023), 2023, : 11151 - 11161