ALGORITHMIC AND SOFTWARE FOR PERSON RE-IDENTIFICATION BASED ON NEURAL NETWORK FEATURES OF FACE AND FIGURE IMAGES
Article Sidebar
Main Article Content
Abstract
This paper considers the key challenges in implementing open person re-identification systems based on neural network technologies. The influence of hyperparameters during convolutional neural network training on the learning dynamics and accuracy of the person re-identification algorithm is investigated. The experimental selection of neural network hyperparameters for person re-identification consists of two stages. The first stage requires determining the most effective learning rate and image batch size. The second stage involves determining the number of training epochs, taking into account changes in batch size and speed during neural network training. A series of experiments were conducted using ResNet-50 and DenseNet-121 on the PolReID1077, Market-1501, DukeMTMC-ReID, and MSMT17 datasets. Hyperparameters such as batch size, learning rate, and number of neural network training epochs were determined. The experimental results confirmed an increase in re-identification accuracy. In addition, the training time of neural networks using the proposed hyperparameter adjustment method allows for a reduction in training time compared to using the training method on the base model. An algorithm for person re-identification is presented. It uses a descriptor to describe human features based on neural network features of their face and figure. One global and two local neural network descriptors are used to describe a person's facial features. The human figure image is described by a vector of 1536 elements obtained using DenseNet-121. The proposed approach ensures high re-identification accuracy when facial identification is possible and enables re-identification in the case of hidden faces and body parts. Experimental results are presented. A software implementation of a prototype for re-identification in an open world is described.
Article Details
This work is licensed under a Creative Commons Attribution 4.0 International License.
References
Deep Learning for Person Re-Identification: A Survey and Outlook / M. Ye, J. Shen, G. Lin et al. // 2021 IEEE Transactions on Pattern Analysis and Machine Intelligence. – 2021. – Vol. 44, iss. 6 – P. 2872–2893. – DOI: 10.1109/TPAMI.2021.3054775.
Brkljac B., Brkljac M. Person detection and re-identification in open-world settings of retail stores and public spaces // ArXiv: 2505.00772. – 2025. – DOI: 10.48550/arXiv.2505.00772.
Li X., Wu A., Zheng W. Adversarial Open-World Person Re-Identification // 15th European Conference, Munich, Germany, September 8–14, 2018, Proceedings, Part II. – 2018. – P. 287–303. – DOI: 10.1007/978-3-030-01216-8_18.
Illumination Unification for Person Re-Identification / G. Zhang, Z. Luo, Y. Chen et al. // IEEE Transactions on Circuits and Systems for Video Technology. – 2022. – Vol. 32, iss. 10. – P. 6766–6777. – DOI: 10.1109/TCSVT.2022.3169422.
Learning Clothing and Pose Invariant 3D Shape Representation for Long-Term Person Re-Identification / F. Liu, M. Kim, Z. Gu et al. // 2023 IEEE/CVF International Conference on Computer Vision (ICCV). – 2023. – P. 19560–19569. – DOI: 10.1109/ICCV51070.2023.01797.
Person Re-identification in Video Surveillance Systems Using Deep Learning: Analysis of the Existing Methods / H. Chen, S. A. Ihnatsyeva, R. P. Bohush et al. // Automation and Remote Control. – 2023. – Vol. 84, iss. 5. – P. 497–528. – DOI: 10.1134/S0005117923050041.
Style Normalization and Restitution for Generalizable Person Re-Identification / X. Jin, C. Lan, W. Zeng et al. // 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). – 2020. – P. 3140–3149. – DOI: 10.1109/CVPR42600.2020.00321.
Generalizable Person Re-Identification by Domain-Invariant Mapping Network / J. Song, Y. Yang, Y. Song et al. // 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). – 2019. – P. 719–728. – DOI: 10.1109/CVPR.2019.00081.
DeVries T., Taylor G. W. Improved regularization of convolutional neural networks with CutOut // ArXiv. – 2017. – DOI: 10.48550/arXiv.1708.04552.
Dropout: a simple way to prevent neural networks from overfitting / N. Srivastava, G.E. Hinton, A. Krizhevsky et al. // Journal of Machine Learning Research. – 2014. – Vol. 15, iss. 1. – P. 1929–1958. – DOI: 10.5555/2627435.2670313.
Choice of activation function in convolution neural network for person re-identification in video surveillance systems / H. Chen, S. Ihnatsyeva, R. Bohush et al. // Programming and computer software. – 2022. – № 5. – P. 312–321. – DOI: 10.1134/S0361768822050036.
Random Erasing Data Augmentation / Z. Zhong, L. Zheng, G. Kang et al. // Proceedings of the AAAI Conference on Artificial Intelligence. – 2020. – Vol. 34, iss. 7. – P. 13001–13008. – DOI: 10.1609/aaai.v34i07.7000.
CutMix: Regularization Strategy to Train Strong Classifiers With Localizable Features / S. Yun, D. Han, S. Oh Kang et al. // 2019 IEEE/CVF International Conference on Computer Vision (ICCV). – 2019. – P. 6022–6031. – DOI: 10.1109/ICCV.2019.00612.
Scalable Person Reidentification: A Benchmark / L. Zheng, L. Shen, L. Tian et al. // 2015 IEEE International Conference on Computer Vision (ICCV). – 2015. – P. 1116–1124. – DOI: 10.1109/ICCV.2015.133.
Performance Measures and a Data Set for Multi-target, Multicamera Tracking / E. Ristani, F. Solera, R. S. Zou et al. // ArXiv. – DOI: 10.48550/arXiv.1609.01775.
Person Transfer GAN to Bridge Domain Gap for Person Re-identification / L. Wei, S. Zhang, W. Gao et al. // 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. – 2018. – P. 79–88. – DOI: 10.1109/CVPR.2018.00016.
Li W., Wang X. Locally Aligned Feature Transforms across Views // 2013 IEEE Conference on Computer Vision and Pattern Recognition, Portland, OR, USA. – 2013. – P. 3594–3601. – DOI: 10.1109/CVPR.2013.461.
DeepReID: Deep Filter Pairing Neural Network for Person Re-identification / W. Li, R. Zhao, T. Xiao et al. // 2014 IEEE Conference on Computer Vision and Pattern Recognition, Columbus, OH, USA. – 2014. – P. 152–159. – DOI: 10.1109/CVPR.2014.27.
Data augmentation and fine tuning of convolution neural network during training for person re-identification in video surveillance systems / S. Ye, R. Bohush, S. Ihnatsyeva et al. // Optical memory and Neural Network. – 2023. – № 4. – P. 233–246. – DOI: 10.3103/S1060992X23040124.
Huang, G. Densely Connected Convolutional Networks / G. Huang, Z. Liu, K. Q. Weinberger et al. // 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). – 2017. – P. 2261–2269. – DOI: 10.1109/CVPR.2017.243.
Deep Residual Learning for Image Recognition / K. He, X. Zhang, S. Ren et al. // 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). – 2016. – P. 770–778. – DOI: 10.1109/CVPR.2016.90.
Miao, J. Pose-Guided Feature Alignment for Occluded Person Re-Identification / J. Miao, Y. Wu, P. Liu et al. // 2019 IEEE/CVF International Conference on Computer Vision (ICCV). – 2019. – P. 542–551. – DOI: 10.1109/ICCV.2019.00063.
Person re-identification in video surveillance systems by feature replacement of occluded parts of human figures // S. Ye, R. Bohush, S. A. Ihnatsyeva et al. // Pattern Analysis and Applications. – 2025. – Vol. 28. – Art. ID 102. – DOI: 10.1007/s10044-025-01482-1.
Most read articles by the same author(s)
- D. GLUKHOV, R. BOHUSH, T. GLUKHOVA, SPARE MATRIX REPRESENTATION USING ASSOCIATIVE CONTAINERS C ++ STL LIBRARIES, Vestnik of Polotsk State University. Part C. Fundamental Sciences: No. 12 (2020)
- R. BOHUSH, Y. ADAMOVSKIY, V. CHERTKOV, DATA COMPOSITION AND REPRESENTATION FOR COGNITIVE COMMUNICATION SYSTEM MODEL BASED ON LTE, Vestnik of Polotsk State University. Part C. Fundamental Sciences: No. 12 (2021)
- R. BOHUSH, I. ZAKHARAVA, Y. PASTUKHOV, D. PASTUKHOV, N. NAUMOVICH, SIMULATION OF EARTH REMOTE SENSING DATA COMPRESSION BASED ON BLOCK ADAPTIVE QUANTIZATION, Vestnik of Polotsk State University. Part C. Fundamental Sciences: No. 4 (2019)
- A. HODOSEVICH, R. BOHUSH, CLIENT-SERVER SYSTEM FOR PARKING MANAGEMENT BASED ON VIDEO SURVEILLANCE DATA ANALYSIS, Vestnik of Polotsk State University. Part C. Fundamental Sciences: No. 4 (2022)
- S. IHNATSYEVA, ORGANIZATION PRINCIPLES AND APPROACHES ANALYSIS TO IMPROVING THE PERSON RE-IDENTIFICATION ACCURACY IN DISTRIBUTED VIDEO SURVEILLANCE SYSTEMS, Vestnik of Polotsk State University. Part C. Fundamental Sciences: No. 4 (2022)
- Y. ADAMOVSKIY, R. BOHUSH, V. CHERTKOV, N. NAUMOVICH, I. STEGKO, USER ACTIVITY MODELING BASED ON MARKOV CHAIN FOR RADIO ENVIRONMENT MAP IN COGNITIVE RADIO NETWORKS, Vestnik of Polotsk State University. Part C. Fundamental Sciences: No. 11 (2022)
- R. BOHUSH, I. ZAKHARAVA, CLIENT-SERVER SYSTEM FOR PEOPLE IDENTIFICATION AND TRACKING IN INDOOR BASED ON VIDEO DATA ANALYSIS, Vestnik of Polotsk State University. Part C. Fundamental Sciences: No. 12 (2020)
- R. BOHUSH, S. ABLAMEYKO, I. ZAHKARAVA, CALCULATION AND ANALYSIS OF MOVING OBJECT FEATURES FOR TRACKING IN VIDEO SEQUENCES, Vestnik of Polotsk State University. Part C. Fundamental Sciences: No. 4 (2021)
- R. BOHUSH, I. ZAHAROVA, V. CHERTKOV, N. NAUMOVICH, SPACECRAFT SAR DATA PROCESSING ALGORITHMS MODELING FOR TRANSMITTING AND RADAR IMAGE FORMATION, Vestnik of Polotsk State University. Part C. Fundamental Sciences: No. 12 (2018)
- N. LUPENKO, R. BOHUSH, H. CHEN, ANALYSIS OF METHODS FOR DISTANCE ESTIMATION TO AN OBJECT FROM A SINGLE VIDEO CAMERA IMAGE USING NEURAL NETWORKS, Vestnik of Polotsk State University. Part C. Fundamental Sciences: No. 2 (2024)