Person Re-Identification System Using Facial Recognition: A Literature Survey

Abstract

Person re-identification (ReID) is a critical task in surveillance, security, and access control systems, aiming to consistently recognize individuals across time and varying environments. While deep learning and face-based methods dominate recent advancements, they often face limitations related to computational complexity, dependence on high-quality data, and vulnerability to occlusions, lighting variations, and appearance changes. This literature survey critically evaluates key ReID methodologies, including deep learning-based, face-centric, hybrid, etc. The review highlights the potential of combining landmark-based modeling with real-time vectorization to overcome the scalability and robustness issues of conventional techniques. Future directions include improving landmark detection accuracy, optimizing anthropometric measurement extraction, and validating these methods in diverse real-world scenarios. This survey positions anthropometry as a promising foundation for next-generation, sustainable person re-identification systems.

Introduction

Person re-identification (ReID) involves recognizing individuals across different times and locations despite challenges like occlusion, pose, and lighting changes. It is crucial for surveillance, safety, and access control. Most current ReID methods rely on deep learning—especially convolutional neural networks (CNNs), attention mechanisms, and face recognition—but face challenges including high computational costs, dependence on large labeled datasets, sensitivity to environmental variations, and privacy concerns.

The literature covers several key approaches:

• Deep Learning-Based Methods: Techniques focus on image enhancement, multi-modal fusion (combining RGB, infrared, thermal), and sophisticated CNN architectures to improve feature extraction and robustness. Limitations include computational demand, need for extensive data, and struggles with occlusion and pose variations. Improvements suggested include lightweight models, unsupervised learning, and better domain adaptation.

• Face-Centric and Multi-Biometric Systems: These use facial recognition and biometric fusion for identification but face issues with low-resolution images, lighting, pose changes, and privacy. Advanced models attempt to preserve identity consistency and augment facial data.

• Hybrid and Multi-Stream Systems: Combining features from multiple body parts, gait, and anthropometric data enhances long-term and context-aware identification but increases system complexity and computational costs.

• Anthropometry-Based ReID: Using human body measurements (e.g., body proportions, gait) offers advantages such as interpretability, environmental robustness, and privacy. Although research is limited, anthropometry is promising for sustainable and low-cost ReID solutions.

• Forensic and Descriptive Techniques: These focus on anatomical studies (like ear shape) and appearance descriptors, highlighting their limits under varying conditions and proposing integration with gait or anthropometry for improved reliability.

Conclusion

The review reveals that while deep learning and face-based identification dominate current ReID literature, they often face challenges in scalability, privacy, and robustness to real-world variation. Hybrid models improve accuracy by combining multiple feature streams, yet they introduce computational overhead and label noise. Studies exploring anthropometry and soft-biometrics represent a promising direction for long-term, low-cost, and privacy-respecting person re-identification. Anthropometric modelling, particularly when combined with vector graphics and landmark detection, offers a stable and interpretable alternative to conventional methods. Future research should focus on refining anthropometric measurements, improving landmark detection, and validating these systems across diverse populations and conditions. Integrating these with real-time vectorization and motion-aware analytics can form the backbone of sustainable ReID systems

References

[1] Ackermann, A., Wischniewski, S., Bonin, D., Gaida, A.-K., & Jaitner, T. (2025). Validation of a 3D whole-body scanning system to collect anthropometric data from a working-age population for ergonomic design. International Journal of Industrial Ergonomics, 105, 103698. [2] Fugiwara, F. V. G., Vassallo, F. N. S., Moritsugui, D. S., Mazzilli, L. E. N., & Melani, R. F. H. (2025). Forensic facial reconstruction: an anthropometric study of the ear. Australian Journal of Forensic Sciences, 57(1), 142–153. https://doi.org/10.1080/00450618.2024.2306679 [3] Hashemifar, S., Marefat, A., Hassannataj Joloudari, J., & Hassanpour, H. (2024). Enhancing Face Recognition with Latent Space Data Augmentation and Facial Posture Reconstruction. Journal of Computer Vision and Artificial Intelligence, 15(2), 112–128. [4] Jiang, L., Wu, A., Ouyang, W., Xu, J., & Zheng, W.-S. (2023). Low-Rank Multi-Scale Multi-Modal Fusion for Person Re-Identification. Pattern Recognition, 144, 109020. [5] Jiang, L., Wu, A., Xu, J., & Zheng, W.-S. (2023). DSAF: Dual Space Alignment Framework for Visible-Infrared Person Re-Identification. IEEE Transactions on Image Processing, 32, 1234–1248. [6] Ming, Z., Zhu, M., Wang, X., Zhu, J., Cheng, J., Gao, C., Yang, Y., & Wei, X. (2022). Deep learning-based person re-identification methods: A survey and outlook of recent works. Image and Vision Computing, 119, 104394. [7] Nambiar, A., Bernardino, A., Nascimento, J. C., & Fred, A. (2017). Context-aware person re-identification in the wild via fusion of gait and anthropometric features.Proceedings of the IEEE International Conference on Computer Vision Theory and Applications (VISAPP), Porto, Portugal. [8] Papantoniou, F. P., Lattas, A., Moschoglou, S., Deng, J., Kainz, B., & Zafeiriou, S. (2024). Arc2Face: A foundation model for ID-consistent human faces.arXiv preprint arXiv:2403.11641. [9] Park, J., Kino, S., & Hsiao-Wecksler, E. T. (2023). A comparative anthropometric analysis of U.S. female firefighters versus the general female population.Applied Ergonomics, 110, 103664. [10] Pham, V.-H., Tran, D.-P., & Hoang, V.-D. (2020). Personal identification based on deep learning technique using facial images for intelligent surveillance systems. International Journal of Machine Learning and Computing, 9(4), 465–470. https://doi.org/10.18178/ijmlc.2020.9.4.827 [11] Rodríguez-Quiñonez, J. C., Trujillo-Hernández, G., Flores-Fuentes, W., Castro-Toscano, M. J., Sergiyenko, O., Hernández-Balbuena, D., & Murrieta-Rico, F. N. (2025). Anthropometric Vision System for Measuring the Windlass Mechanism During the Gait Cycle. Biomedical Signal Processing and Control, 103, 107405. [12] Santana, O. J., Lorenzo-Navarro, J., Freire-Obregón, D., Hernández-Sosa, D., & Castrillón-Santana, M. (2024). Applying deep learning image enhancement methods to improve person re-identification.Neurocomputing, 598, 128011. [13] Sarfraz, M. S., Schumann, A., Eberle, A., & Stiefelhagen, R. (2018). Pose-invariant embedding for deep person re-identification.IEEE Transactions on Image Processing, 27(1), 303–312. [14] Satta, R. (2013). Appearance descriptors for person re-identification: A comprehensive review.arXiv preprint arXiv:1307.5748. [15] Sengupta, A., Anshul, A., & Singh, A. K. (2024). Hardware security against IP piracy using secure fingerprint encrypted fused amino-acid biometric with facial anthropometric signature.Microprocessors and Microsystems, 112, 105131. [16] Teoh, K. H., Ismail, R. C., Naziri, S. Z. M., Hussin, R., Isa, M. N. M., & Basir, M. S. S. M. (2021). Face Recognition and Identification using Deep Learning Approach.Journal of Physics: Conference Series, 1755(1), 012006. https://doi.org/10.1088/1742-6596/1755/1/012006 [17] Thanh, D. T., Lee, Y., & Kang, B. (2024). Enhancing long-term person re-identification using global, local body part, and head streams. Neurocomputing, 580, 127480. [18] Wu, L., Shen, C., & van den Hengel, A. (2016). PersonNet: Person Re-Identification with Deep Convolutional Neural Networks. arXiv preprint arXiv:1601.07255. [19] Younis, A. N. (2021). Personal identification system based on multi-biometric depending on cuckoo search algorithm. Journal of Physics: Conference Series, 1879(1), 022080. [20] Zhao, L., Xie, F., Xing, J., Wu, X., Yan, J., & Tian, Q. (2017). Learning discriminative deep features for person re-identification.IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(3), 365–379.

Copyright

Copyright © 2025 Harshit S, Cyrill Victor, Pravin M, Ashwin Kumar. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.