您的位置：山东大学 -> 科技期刊社 -> 《山东大学学报（工学版）》

山东大学学报 (工学版) ›› 2026, Vol. 56 ›› Issue (4): 1-9.doi: 10.6040/j.issn.1672-3961.0.2025.218

• 机器学习与数据挖掘 •    

面向跨分辨率身份匹配的行人重识别

刘志刚1,2,冯涵霖1,周元核1,宿健珩1,张岩1,2   

  1. 1.东北石油大学计算机与信息技术学院, 黑龙江 大庆 163318;2.黑龙江省石油大数据与智能分析重点实验室(东北石油大学), 黑龙江 大庆 163318
  • 发布日期:2026-08-12
  • 作者简介:刘志刚(1979— ),男,吉林蛟河人,教授,硕士生导师,博士,主要研究方向为深度学习理论和计算机视觉应用. E-mail:ZhigangLiu@nepu.edu.cn
  • 基金资助:
    国家自然科学基金资助项目(42002138);黑龙江省科技创新基地资助项目(JD24A009);黑龙江省博士后科研启动基金资助项目(216230006);黑龙江省高等教育教学改革研究重点资助项目(SJGZB2024173)

Person re-identification for cross-resolution identity matching

Liu Zhigang1,2, Feng Hanlin1, Zhou Yuanhe1, Su Jianheng1, Zhang Yan1,2   

  1. Liu Zhigang1, 2, Feng Hanlin1, Zhou Yuanhe1, Su Jianheng1, Zhang Yan1, 2(1. School of Computer and Information Technology, Northeast Petroleum University, Daqing 163318, Heilongjiang, China;
    2. Heilongjiang Petroleum Big Data and Intelligent Analysis Key Laboratory, Northeast Petroleum University, Daqing 163318, Heilongjiang, China
  • Published:2026-08-12

摘要： 针对真实场景中摄像机成像条件和环境条件不同,行人识别会产生分辨率不匹配的问题,提出一种基于小波超分辨率和双域特征融合的行人重识别网络。引入小波分解,将图像特征分为结构特征和细节特征,通过小波卷积扩大感受野,增强结构特征提取能力;在小波域,针对高低频子带特性设计差异化损失函数,引导模型生成更具判别性的超分辨率图像;通过注意力机制强化超分辨率图像结构特征,与小波域结构特征进行门控融合,构建包含丰富结构信息的联合特征。仿真结果表明,所提方法在识别性能上优于主流方法,在最具挑战性的CAVIAR数据集上,行人图像第一次匹配正确的概率达67.8%,改善了跨分辨率场景下的行人重识别问题。

关键词: 行人重识别, 跨分辨率, 超分辨率重建, 小波卷积, 注意力机制

Abstract: Aiming at the problem of resolution mismatch in person recognition caused by variations in camera imaging conditions and environmental factors in real-world scenarios, a person re-identification network integrating wavelet-based super-resolution and dual-domain feature fusion was proposed. Wavelet decomposition was introduced to separate image features into structural and detail components, and wavelet convolution was employed to enlarge the receptive field and strengthen structural feature extraction. In the wavelet domain, differentiated loss functions were designed for high- and low-frequency subbands to guide the reconstruction of more discriminative super-resolution images. An attention mechanism was used to enhance the structural features of super-resolution images, and the enhanced features were gate-fused with structural features in the wavelet domain to construct a joint representation containing richer structural cues. Simulation results showed that the proposed method outperformed the mainstream methods in recognition performance. On the challenging CAVIAR dataset, the probability of the first correct matching of person images reached 67.8%, which effectively improved person re-identification under cross-resolution conditions.

Key words: person re-identification, cross-resolution, super-resolution reconstruction, wavelet convolution, attention mechanism

中图分类号: 

  • TP391
[1] 闫铭, 李雷孝, 林浩, 等. 少样本行人重识别研究综述[J]. 计算机工程与应用, 2025, 61(17): 62-88. Yan Ming, Li Leixiao, Lin Hao, et al. Survey of research on few-shot person re-identification[J]. Computer Engineering and Applications, 2025, 61(17): 62-88.
[2] 傅桂霞, 邹国锋, 毛帅, 等. 融合Gabor特征与卷积特征的小样本行人重识别[J]. 山东大学学报(工学版), 2021, 51(3): 22-29. Fu Guixia, Zou Guofeng, Mao Shuai, et al. Small sample person re-identification combining Gabor features and convolution features[J]. Journal of Shandong University(Engineering Science), 2021, 51(3): 22-29.
[3] 王路遥, 王凤随, 闫涛, 等. 结合多尺度特征与混淆学习的跨模态行人重识别[J]. 智能系统学报, 2024,19(4): 898-908. Wang Luyao, Wang Fengsui, Yan Tao, et al. Cross-modal person re-identification combining multi-scale features and confusion learning[J]. CAAI Transactions on Intelligent Systems, 2024, 19(4): 898-908.
[4] 张继, 金翠, 王洪元, 等. 基于奇异值分解行人对齐网络的行人重识别[J]. 山东大学学报(工学版), 2019, 49(5): 91-97. Zhang Ji, Jin Cui, Wang Hongyuan, et al. Pedestrian recognition based on singular value decomposition pedestrian alignment network[J]. Journal of Shandong University(Engineering Science), 2019, 49(5): 91-97.
[5] 朱沛伍, 高树辉. 低高频多尺度融合的跨模态行人重识别研究[J]. 重庆邮电大学学报(自然科学版), 2024, 36(6): 1183-1193. Zhu Peiwu, Gao Shuhui. Research on cross-modal pedestrian re-identification based on low-high frequency multi-scale fusion[J]. Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition), 2024, 36(6): 1183-1193.
[6] Jiao J N, Zheng W S, Wu A C, et al. Deep low-resolution person re-identification[C] //Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence. New Orleans, USA: AAAI, 2018: 5123-5132.
[7 ] Wang Z, Ye M, Yang F, et al. Cascaded SR-GAN for scale-adaptive low resolution person re-identification[C] //Proceedings of the 27th International Joint Conference on Artificial Intelligence. Stockholm, Sweden: ACM, 2018: 3891-3897.
[8] Cheng Z Y, Dong Q, Gong S G, et al. Inter-task association critic for cross-resolution person re-identification[C] //2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR). Seattle, USA: IEEE, 2020: 2602-2612.
[9] Zhang G Q, Ge Y, Dong Z C, et al. Deep high-resolution representation learning for cross-resolution person re-identifica-tion[J]. IEEE Transactions on Image Processing, 2021, 30: 8913-8925.
[10] Li Y J, Chen Y C, Lin Y Y, et al. Recover and identify: a generative dual model for cross-resolution person re-identification[C] //2019 IEEE/CVF Interna-tional Conference on Computer Vision(ICCV). Seoul: IEEE, 2019: 8089-8098.
[11] Zhang W C, Xiong S H, He X H, et al. Multi deep invariant feature learning for cross-resolution person re-identification[J]. Information Processing & Manage-ment, 2024, 61(4): 103764.
[12] Huang Y K, Zha Z J, Fu X Y, et al. Real-world person re-identification via degradation invariance learning[C] //2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR). Seattle, USA: IEEE, 2020: 14072-14082.
[13] Huang Y K, Fu X Y, Li L, et al. Learning degradation-invariant representation for robust real-world person re-identification[J]. International Journal of Computer Vision, 2022, 130(11): 2770-2796.
[14] Yu Y, She K, Liu J H, et al. A super-resolution network for medical imaging via transformation analysis of wavelet multi-resolution[J]. Neural Networks, 2023, 166: 162-173.
[15] 黄裕青, 李华锋, 原铭, 等. 基于卷积神经网络梯度和纹理补偿的单幅图像超分辨率重建[J]. 数据采集与处理, 2023, 38(5): 1112-1124. Huang Yuqing, Li Huafeng, Yuan Ming, et al. Super-resolution reconstruction of single image based on convolutional neural network gradient and texture compensation[J]. Journal of Data Acquisition and Processing, 2023, 38(5): 1112-1124.
[16] Guo T T, Mousavi H S, Vu T H, et al. Deep wavelet prediction for image super-resolution[C] //2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops(CVPR). Honolulu, USA: IEEE, 2017: 1100-1109.
[17] Xu Y F, Zhou Y, Ma H B, et al. Wavelet-based dual discriminator GAN for image super-resolution[J]. Knowledge-Based Systems, 2025, 317: 113383.
[18] Chouchane A, Bessaoudi M, Boutellaa E, et al. A new multidimensional discriminant representation for robust person re-identification[J]. Pattern Analysis and Appli-cations, 2023, 26(3): 1191-1204.
[19] Finder S E, Amoyal R, Treister E, et al. Wavelet convolutions for large receptive fields[C] //Computer Vision-ECCV 2024. Milan, Italy: Springer, 2024: 363-380.
[20] Yao Y M, Jiang X Y, Fujita H, et al. A sparse graph wavelet convolution neural network for video-based person re-identification[J]. Pattern Recognition, 2022, 129: 108708.
[21] Hou Q B, Zhou D Q, Feng J S. Coordinate attention for efficient mobile network design[C] //2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR). Nashville, USA: IEEE, 2021: 13708-13717.
[22] Wu X J, Jiang X, Dong L G. Gated weighted normative feature fusion for multispectral object detection[J]. The Visual Computer, 2024, 40(9): 6409-6419.
[23] Cheng D S, Cristani M, Stoppa M, et al. Custom pictorial structures for re-identification[C] //Proceedings of the British Machine Vision Conference 2011. Dundee, UK: BMVA, 2011: 68.
[24] Zheng L, Shen L Y, Tian L, et al. Scalable person re-identification: a benchmark[C] //2015 IEEE International Conference on Computer Vision(ICCV). Santiago, Chile: IEEE, 2016: 1116-1124.
[25] Zheng Z D, Zheng L, Yang Y. Unlabeled samples generated by GAN improve the person re-identification baseline in vitro[C] //2017 IEEE International Conference on Computer Vision(ICCV). Venice, Italy: IEEE, 2017: 3774-3782.
[26] Zhu F Z, Li D L, Sun C, et al. Cross-resolution person re-identification based on double-layer graph convolution network[J]. Engineering Applications of Artificial Intelligence, 2025, 147: 110376.
[27] Han K, Huang Y, Wang L, et al. Self-supervised recovery and guide for low-resolution person re-identification[J]. IEEE Transactions on Information Forensics and Security, 2024, 19: 6252-6263.
[28] Ouyang R Y, Wang Z W. Low-resolution guided dual-branch fusion network for cross-resolution person re-identification[C] //2025 IEEE International Conference on Pattern Recognition, Machine Vision and Artificial Intelligence(PRMVAI). Loudi, China: IEEE, 2025: 11108330.
[29] Wu L Y, Liu L Q, Wang Y, et al. Learning resolution-adaptive representations for cross-resolution person re-identification[J]. IEEE Transactions on Image Pro- cessing, 2023, 32: 4800-4811.
[1] 任红伟,孟菲,王继凯,田威杨,魏明召,程之恒,杜聪,吴建清. 基于深度学习的路基病害智能检测方法[J]. 山东大学学报 (工学版), 2026, 56(3): 93-105.
[2] 王倩,张瑞敏,李明津,孟宪静,耿蕾蕾. 基于频域图卷积网络的时空序列预测[J]. 山东大学学报 (工学版), 2026, 56(3): 84-92.
[3] 郑哲明, 孔玲玲, 何印. 基于双图结构的时空图卷积网络短期风电功率预测模型[J]. 山东大学学报 (工学版), 2026, 56(2): 130-138.
[4] 刘飞宇,张静,王亦楠. 基于渐近式特征融合的轻量化SAR舰船检测算法[J]. 山东大学学报 (工学版), 2026, 56(2): 52-59.
[5] 赵峰,刘瑞,王英,陈小强,葛磊蛟,马爱平. 多尺度融合与动态自校正旋转的吊弦检测算法[J]. 山东大学学报 (工学版), 2026, 56(2): 1-10.
[6] 王禹鸥,苑迎春,何振学,何晨. 融合多特征和多头自注意力机制的高校学业命名实体识别[J]. 山东大学学报 (工学版), 2025, 55(6): 35-44.
[7] 周群颖,隋家成,张继,王洪元. 基于自监督卷积和无参数注意力机制的工业品表面缺陷检测[J]. 山东大学学报 (工学版), 2025, 55(4): 40-47.
[8] 李丰,文益民. 融合多尺度视觉和文本语义特征的图像描述生成算法[J]. 山东大学学报 (工学版), 2025, 55(3): 80-87.
[9] 王禹鸥,苑迎春,何振学,王克俭. 改进RoBERTa、多实例学习和双重注意力机制的关系抽取方法[J]. 山东大学学报 (工学版), 2025, 55(2): 78-87.
[10] 邹正标,刘毅志,廖祝华,赵肄江. 动态交通流量预测的时空注意力图卷积网络[J]. 山东大学学报 (工学版), 2024, 54(5): 50-61.
[11] 李家春,李博文,常建波. 一种高效且轻量的RGB单帧人脸反欺诈模型[J]. 山东大学学报 (工学版), 2023, 53(6): 1-7.
[12] 王碧瑶,韩毅,崔航滨,刘毅超,任铭然,高维勇,陈姝廷,刘嘉巍,崔洋. 基于图像的道路语义分割检测方法[J]. 山东大学学报 (工学版), 2023, 53(5): 37-47.
[13] 宋佳芮,陈艳平,王凯,黄瑞章,秦永彬. 基于Affix-Attention的命名实体识别语义补充方法[J]. 山东大学学报 (工学版), 2023, 53(2): 70-76.
[14] 刘方旭,王建,魏本征. 基于多空间注意力的小儿肺炎辅助诊断算法[J]. 山东大学学报 (工学版), 2023, 53(2): 135-142.
[15] 武新章,梁祥宇,朱虹谕,张冬冬. 基于CEEMDAN-GRA-PCC-ATCN的短期风电功率预测[J]. 山东大学学报 (工学版), 2022, 52(6): 146-156.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
版权所有 © 2018 《山东大学学报(工学版)》编辑部　
地址: 济南市山大南路27号（250100）　 电话: 0531-88366735　 E-mail: xbgxb@sdu.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发