Person Re-Identification using Centroid and Quadruple Loss based Deep Learning Model
Keywords:
Loss function, Deep learning, Centroid Loss, person re-identification, Quadruple lossAbstract
Person re-identification (Re-ID) stands as a crucial objective within expansive video surveillance landscapes, concentrating on the recognition of individuals across diverse camera sources. Lately, the utilization of deep learning networks in conjunction with the triplet loss has emerged as a prevalent framework for advancing person Re-ID endeavours. In this paper we have discussed various loss functions of Deep Learning and their application in person-re-identification. Utilizing the mean centroid representation offers heightened resilience to outliers and guarantees a more dependable set of features. This arises from the fact that every class is symbolized by a sole embedding, denoted as the class centroid. This approach not only substantially curtails retrieval time and storage needs but also enhances stability. This paper proposes a novel Centroid Quadruple Loss based approach. Through comprehensive experiments, it becomes evident that the proposed approach yields substantial improvements in results compared to the centroid triplet loss based approach as well as other recent state-of-the-art person re-identification methods.
Downloads
References
A. Peña et al., “Human-Centric Multimodal Machine Learning: Recent Advances and Testbed on AI-Based Recruitment,” SN Comput. Sci., vol. 4, no. 5, 2023, doi: 10.1007/s42979-023-01733-0.
K. Singh, M. K. Ahirwal, and M. Pandey, “Mental Health Monitoring Using Deep Learning Technique for Early-Stage Depression Detection,” SN Comput. Sci., vol. 4, no. 6, p. 701, 2023, doi: 10.1007/s42979-023-02113-4.
Y. Salini and J. HariKiran, “ViT: Quantifying Chest X-Ray Images Using Vision Transformer & XAI Technique,” SN Comput. Sci., vol. 4, no. 6, p. 754, 2023, doi: 10.1007/s42979-023-02204-2.
L. Bazzani, M. Cristani, and V. Murino, “Symmetry-Driven Accumulation of Local Features for Human Characterization and Re-identification.”
D. Yi, Z. Lei, S. Liao, and S. Z. Li, “Deep metric learning for person re-identification,” in 2014 22nd International Conference on Pattern Recognition, 2014, pp. 34–39.
W. Li, R. Zhao, T. Xiao, and X. Wang, “DeepReID: Deep Filter Pairing Neural Network for Person Re-identification,” in 2014 IEEE Conference on Computer Vision and Pattern Recognition, Jun. 2014, pp. 152–159, doi: 10.1109/CVPR.2014.27.
M. Wieczorek, B. Rychalska, and J. Dąbrowski, On the Unreasonable Effectiveness of Centroids in Image Retrieval, vol. 13111 LNCS, no. 1. Association for Computing Machinery, 2021.
W. Chen, X. Chen, J. Zhang, and K. Huang, “Beyond triplet loss: A deep quadruplet network for person re-identification,” Proc. - 30th IEEE Conf. Comput. Vis. Pattern Recognition, CVPR 2017, vol. 2017-January, pp. 1320–1329, 2017, doi: 10.1109/CVPR.2017.145.
A. Hermans, L. Beyer, and B. Leibe, “In Defense of the Triplet Loss for Person Re-Identification,” 2017, [Online]. Available: http://arxiv.org/abs/1703.07737.
V. Yathish, “Loss Functions and Their Use In Neural Networks,” Towards Data Science, Aug. 04, 2022. https://towardsdatascience.com/loss-functions-and-their-use-in-neural-networks-a470e703f1e9 (accessed Apr. 12, 2023).
K. Mahendru, “Understanding Loss Functions to Maximize Machine Learning Model Performance (Updated 2023),” Aug. 14, 2019. https://www.analyticsvidhya.com/blog/2019/08/detailed-guide-7-loss-functions-machine-learning-python-code/ (accessed May 25, 2023).
Y. Sarıgöz, “Triplet Loss — Advanced Intro,” Towards Data Science, Mar. 25, 2022. https://towardsdatascience.com/triplet-loss-advanced-intro-49a07b7d8905.
Z. Hu, H. Wu, S. Liao, H. Hu, S. Liu, and B. Li, “Person Re-Identification with Hybrid Loss and Hard Triplets Mining,” in 2018 IEEE Fourth International Conference on Multimedia Big Data (BigMM), 2018, pp. 1–5, doi: 10.1109/BigMM.2018.8499463.
N. K. Singh, M. Khare, and H. B. Jethva, “A comprehensive survey on person re-identification approaches: various aspects,” Multimed. Tools Appl., vol. 81, no. 11, pp. 15747–15791, 2022, doi: 10.1007/s11042-022-12585-w.
G. Wang, Y. Yuan, X. Chen, J. Li, and X. Zhou, “Learning discriminative features with multiple granularities for person re-identification,” MM 2018 - Proc. 2018 ACM Multimed. Conf., pp. 274–282, 2018, doi: 10.1145/3240508.3240552.
K. Janocha and W. M. Czarnecki, “On loss functions for deep neural networks in classification,” Schedae Informaticae, vol. 25, pp. 49–59, 2016, doi: 10.4467/20838476SI.16.004.6185.
H. Luo et al., “A Strong Baseline and Batch Normalization Neck for Deep Person Re-Identification,” IEEE Trans. Multimed., vol. 22, no. 10, pp. 2597–2609, 2020, doi: 10.1109/TMM.2019.2958756.
H. Gu, J. Li, G. Fu, M. Yue, and J. Zhu, “Loss function search for person re-identification,” Pattern Recognit., vol. 124, p. 108432, 2022, doi: https://doi.org/10.1016/j.patcog.2021.108432.
H. Gu, J. Li, G. Fu, C. Wong, X. Chen, and J. Zhu, “AutoLoss-GMS: Searching Generalized Margin-based Softmax Loss Function for Person Re-identification,” Proc. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit., vol. 2022-June, pp. 4734–4743, 2022, doi: 10.1109/CVPR52688.2022.00470.
Y. Li, G. Yin, C. Liu, X. Yang, and Z. Wang, “Triplet online instance matching loss for person re-identification,” Neurocomputing, vol. 433, pp. 10–18, 2021, doi: https://doi.org/10.1016/j.neucom.2020.12.018.
Z. Yang, J. Liu, T. Liu, Y. Zhu, L. Wang, and D. Tao, “Equidistant distribution loss for person re-identification,” Neurocomputing, vol. 455, pp. 255–264, 2021, doi: https://doi.org/10.1016/j.neucom.2021.04.070.
H. Luo, Y. Gu, X. Liao, S. Lai, and W. Jiang, “Bag of tricks and a strong baseline for deep person re-identification,” IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit. Work., vol. 2019-June, pp. 1487–1495, 2019, doi: 10.1109/CVPRW.2019.00190.
L. He, X. Liao, W. Liu, X. Liu, P. Cheng, and T. Mei, “FastReID: A Pytorch Toolbox for General Instance Re-identification,” 2020, [Online]. Available: http://arxiv.org/abs/2006.02631.
C. Yan et al., “Beyond Triplet Loss: Person Re-Identification With Fine-Grained Difference-Aware Pairwise Loss,” IEEE Trans. Multimed., vol. 24, pp. 1665–1677, 2022, doi: 10.1109/TMM.2021.3069562.
M. L. Ha and V. Blanz, “Deep Ranking with Adaptive Margin Triplet Loss,” 2021, [Online]. Available: http://arxiv.org/abs/2107.06187.
S. Dräger and J. Dunkelau, “Evaluating the Impact of Loss Function Variation in Deep Learning for Classification,” 2022, [Online]. Available: http://arxiv.org/abs/2210.16003.
S. Hu, K. Wang, J. Cheng, H. Tan, and J. Pang, “Triplet Ratio Loss for Robust Person Re-identification,” in Pattern Recognition and Computer Vision, 2022, pp. 42–54.
Y. Sun et al., “Circle loss: A unified perspective of pair similarity optimization,” Proc. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit., pp. 6397–6406, 2020, doi: 10.1109/CVPR42600.2020.00643.
D. Aganian, M. Eisenbach, J. Wagner, D. Seichter, and H. M. Gross, “Revisiting Loss Functions for Person Re-identification,” Lect. Notes Comput. Sci. (including Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinformatics), vol. 12895 LNCS, pp. 30–42, 2021, doi: 10.1007/978-3-030-86383-8_3.
Z. Yang, J. Liu, T. Liu, L. Wang, and S. Zhao, “Circle-Based Ratio Loss for Person Reidentification,” Complexity, vol. 2020, 2020, doi: 10.1155/2020/9860562.
J. SU, X. He, L. Qing, Y. Yu, S. Xu, and Y. Peng, “A New Discriminative Feature Learning for Person Re-Identification Using Additive Angular Margin Softmax Loss,” in 2019 UK/ China Emerging Technologies (UCET), 2019, pp. 1–4, doi: 10.1109/UCET.2019.8881838.
L. Zheng, L. Shen, L. Tian, S. Wang, J. Wang, and Q. Tian, “Scalable person re-identification: A benchmark,” in Proceedings of the IEEE international conference on computer vision, 2015, pp. 1116–1124.
Z. Zheng, L. Zheng, and Y. Yang, “Unlabeled samples generated by gan improve the person re-identification baseline in vitro,” in Proceedings of the IEEE International Conference on Computer Vision, 2017, pp. 3754–3762.
G. Wang, J. Lai, P. Huang, and X. Xie, “Spatial-Temporal Person Re-Identification,” 2018.
Downloads
Published
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
All papers should be submitted electronically. All submitted manuscripts must be original work that is not under submission at another journal or under consideration for publication in another form, such as a monograph or chapter of a book. Authors of submitted papers are obligated not to submit their paper for publication elsewhere until an editorial decision is rendered on their submission. Further, authors of accepted papers are prohibited from publishing the results in other publications that appear before the paper is published in the Journal unless they receive approval for doing so from the Editor-In-Chief.
IJISAE open access articles are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. This license lets the audience to give appropriate credit, provide a link to the license, and indicate if changes were made and if they remix, transform, or build upon the material, they must distribute contributions under the same license as the original.
