Person re-identification accuracy improvement by training a CNN with the new large joint dataset and re-rank

• Autorzy: Bohush Rykhard , Ihnatsyeva Sviatlana , Ablameyko Sergey

Identyfikatory

DOI

10.22630/MGV.2022.31.1.5

• Języki publikacji: EN

Abstrakty

EN

The paper is aimed to improve person re-identification accuracy in distributed video surveillance systems based on constructing a large joint image dataset of people for training convolutional neural networks (CNN). For this aim, an analysis of existing datasets is provided. Then, a new large joint dataset for person re-identification task is constructed that includes the existing public datasets CUHK02, CUHK03, Market, Duke, MSMT17 and PolReID. Testing for re-identification is performed for such frequently cited CNNs as ResNet-50, DenseNet121 and PCB. Re-identification accuracy is evaluated by using the main metrics Rank, mAP and mINP. The use of the new large joint dataset makes it possible to improve Rank1 mAP, mINP on all test sets. Re-ranking is used to further increase the re-identification accuracy. Presented results confirm the effectiveness of the proposed approach.

Słowa kluczowe

EN

convolution neural network; PolReID; re-identification; large-scale dataset; re-rank

• Wydawca: Institute of Information Technology of the Warsaw University of Life Sciences – SGGW
• Czasopismo: Machine Graphics and Vision
• Rocznik: 2022
• Tom: Vol. 31, No. 1/4
• Strony: 93--109
• Opis fizyczny: Bibliogr. 45 poz., il., tab., wykr.

Twórcy

autor

Bohush Rykhard

• Polotsk State University, Novopolotsk, Belarus

• https://orcid.org/0000-0002-6609-5810

autor

Ihnatsyeva Sviatlana

• Polotsk State University, Novopolotsk, Belarus

• https://orcid.org/0000-0002-9780-5731

autor

Ablameyko Sergey

• Belarusian State University, Minsk, Belarus
• United Institute for Informatics Problems of NAS of Belarus, Minsk, Belarus

• https://orcid.org/0000-0001-9404-1206

Bibliografia

• [1] A. Bochkovskiy, C. Y. Wang, and H. Y. M. Liao. YOLOv4: Optimal speed and accuracy of object detection. arXiv, 2020. arXiv:2004.10934. doi:10.48550/arXiv.2004.10934.
• [2] S. Bąk and P. Carr. One-shot metric learning for person re-identification. In Proc. 2017 IEEE Conf. Computer Vision and Pattern Recognition (CVPR 2017), pages 1571-1580, Honolulu, HI, USA, 21-26 Jul 2017. doi:10.1109/CVPR.2017.171.
• [3] Y. Dai, J. Liu, Y. Sun, et al. IDM: An intermediate domain module for domain adaptive person re-ID. In Proc. 2021 IEEE/CVF Conf. Computer Vision (ICCV 2021), pages 11844-11854, Montreal, QC, Canada, 10-17 Oct. doi:10.1109/ICCV48922.2021.01165.
• [4] Y. Dai, Y. Sun, J. Liu, et al. Bridging the source-to-target gap for cross-domain person re-identification with intermediate domains. ArXiv, 2022. arXiv:2203.01682v1. doi:10.48550/arXiv.2203.01682.
• [5] Z. Ding, C. Ding, Z. Shao, and D. Tao. Semantically self-aligned network for text-to-image part-aware person re-identification. arXiv, 2021. arXiv:2107.12666v2. doi:10.48550/arXiv.2107.12666.
• [6] D. Fu, D. Chen, J. Bao, et al. Unsupervised pre-training for person re-identification. In Proc. 2021 IEEE/CVF Conf. Computer Vision and Pattern Recognition (CVPR 2021), pages 14745-14754, Nashville, TN, USA, 20-25 Jun 2021. doi:10.1109/CVPR46437.2021.01451.
• [7] S. Gong and T. Xiang. Person re-identification. In Visual Analysis of Behaviour: From Pixels to Semantics, pages 301-313, London, 2011. Springer. doi:10.1007/978-0-85729-670-2_14.
• [8] D. Gray, S. Brennan, and H. Tao. Evaluating appearance models for recognition, reacquisition, and tracking. In Proc. 10th IEEE Int. Workshop on Performance Evaluation of Tracking and Surveillance (PETS 2007), Sep 2007. https://www.researchgate.net/publication/228345677_Evaluating_appearance_models_for_recognition_reacquisition_and_tracking.
• [9] K. He, X. Zhang, Sh. Ren, and J. Sun. Deep residual learning for image recognition. 2016 IEEE Conf. Computer Vision and Pattern Recognition (CVPR), pages 770-778, 2015. doi:10.1109/cvpr.2016.90.
• [10] M. Hirzer, C. Beleznai, P. M. Roth, and H. Bischof. Person re-identification by descriptive and discriminative classification. In Proc. Scandinavian Conf. Image Analysis (SCIA 2011), volume 6688 of Lecture Notes in Computer Science, pages 91- 102, Ystad, Sweden, 23-25 May 2011. doi:10.1007/978-3-642-21227-7_9.
• [11] G. Huang, Zh. Liu, and K. Q. Weinberger. Densely connected convolutional networks. 2017 IEEE Conf. Computer Vision and Pattern Recognition (CVPR), pages 2261- 2269, 2017. doi:10.1109/CVPR.2017.243.
• [12] S. Ihnatsyeva, R. Bohush, and Ablameyko. Joint dataset for CNN-based person re-identification. In Proc. 15th Int. Conf. Pattern Recognition and Information Processing (PRIP 2021), pages 33- 37, Minsk, Belarus, 21-24 Sep 2021. United Institute of Informatics Problems, NAS Belarus, Minsk. https://elib.psu.by/handle/123456789/28586.
• [13] SvetlanaIgn (S. Ihnatsyeva). PolReID. GitHub, Sep 2022. https://github.com/SvetlanaIgn/PolReID. [Accessed 1 Dec 2022].
• [14] X. Jin, C. Lan, W. Zeng, et al. Style normalization and restitution for generalizable person re-identification. In Proc. 2020 IEEE/CVF Conf. Computer Vision and Pattern Recognition (CVPR 2020), pages 3140- 3149, Seattle, WA, USA, 13-19 Jun 2020. doi:10.1109/cvpr42600.2020.00321.
• [15] X. Jing, X. Zhu, F. Wu, et al. Super-resolution person re-identification with semi-coupled low-rank discriminant dictionary learning. IEEE Transactions on Image Processing, 26(3):1363-1378, 2015. doi:10.1109/TIP.2017.2651364.
• [16] R. Layne, T. M. Hospedales, and S. Gong. Investigating open-world person re-identification using a drone. In Agapito L. et al., editors, Computer Vision - Proc. European Conf. Computer Vision Workshops (ECCVW 2014), volume 8927, Part III of Lecture Notes in Computer Science, pages 225-240, Zurich, Switzerland, 6-7 Sep 2014. doi:10.1007/978-3-319-16199-0_16.
• [17] S. Li, T. Xiao, H. Li, et al. Person search with natural language description. In Proc. 2017 IEEE Conf. Computer Vision and Pattern Recognition (CVPR 2017), pages 5187- 5196, Honolulu, HI, USA, 21-26 Jul 2017. doi:10.1109/CVPR.2017.551.
• [18] W. Li and X. Wang. Locally aligned feature transforms across views. In Proc. 2013 IEEE Conf. Computer Vision and Pattern Recognition (CVPR 2013), pages 3594-3601, Portland, OR, USA, 23-28 Jun 2013. doi:10.1109/CVPR.2013.461.
• [19] W. Li, R. Zhao, and X. Wang. Human reidentification with transferred metric learning. In K. M. Lee et al., editors, Computer Vision - Proc. 11th Asian Conf. Computer Vision (ACCV 2012), volume 7724 of Lecture Notes in Computer Science, pages 31- 44, Daejeon, Republic of Korea, 5-9 Nov 2012. doi:10.1007/978-3-030-58555-6_14.
• [20] W. Li, R. Zhao, T. Xiao, and X. Wang. DeepReID: Deep filter pairing neural network for person re-identification. In Proc. 2014 IEEE Conf. Computer Vision and Pattern Recognition (CVPR 2014), pages 152-159, Columbus, OH, USA, 23-28 Jun 2014. doi:10.1109/CVPR.2014.27.
• [21] X. Li, W. Zheng, X. Wang, et al. Multi-scale learning for low-resolution person re-identification. In Proc. 2015 IEEE Int. Conf. Computer Vision (ICCV 2015), pages 3765- 3773, Santiago, Chile, 7-13 Dec 2015. doi:10.1109/ICCV.2015.429.
• [22] C. Luo, C. Song, and Z. Zhang. Generalizing person re-identification by camera-aware invariance learning and cross-domain mixup. In A. Vedaldi et al., editors, Computer Vision - Proc. European Conf. Computer Vision (ECCV 2020), volume 12360 of Lecture Notes in Computer Science, pages 224-241, Glasgow, United Kingdom, 23-28 Aug 2020. doi:10.1007/978-3-030-58555-6_14.
• [23] T. D. Nguyen, H. G. Hong, K. W. Kim, and K. R. Park. Person recognition system based on a combination of body images from visible light and thermal cameras. Sensors, 17(3):605, 2017. doi:10.3390/s17030605.
• [24] L. Pang, Y. Wang, Y. Song, et al. Cross-domain adversarial feature learning for sketch re-identification. In Proc. 26th ACM Int. Conf. Multimedia (MM ’18), pages 609-617, Seoul, Republic of Korea, 22-26 Oct 2018. doi:10.1145/3240508.3240606.
• [25] E. Ristani, F. Solera, R. S. Zou, et al. Performance measures and a data set for multi-target, multicamera tracking. In G. Hua et al., editors, Computer Vision - Proc. European Conf. Computer Vision Workshops (ECCVW 2020), volume 9914 of Lecture Notes in Computer Science, pages 17-35, Amsterdam, The Netherlands, 8-16 Oct 2016. doi:10.1007/978-3-319-48881-3_2.
• [26] Y. Shiping, S. Ihnatsyeva, R. Bohush, C. Chen, and S. Ablameyko. Estimation CNN-based person re-identification accuracy in video using different datasets. In C.-H. Chen et al., editors, Applied Mathematics, Modeling and Computer Simulation, volume 30 of Advances in Transdisciplinary Engineering, pages 978- 985. IOS Press, 2022. doi:10.3233/ATDE221122.
• [27] G. Song, B. Leng, Y. Liu, et al. Region-based quality estimation network for large-scale person re-identification. Proc. AAAI Conf. Artificial Intelligence, 32(1):7347- 7354, 2018. doi:10.1609/aaai.v32i1.12305.
• [28] Y. Sun, L. Zheng, Y. Yang, et al. Beyond part models: Person retrieval with refined part pooling. In V. Ferrari et al., editors, Computer Vision - Proc. European Conf. Computer Vision (ECCV 2017), volume 11208, Part IV of Lecture Notes in Computer Science, pages 501- 518, Munich, Germany, 8-14 Sep 2018. doi:10.1007/978-3-030-01225-0_30.
• [29] Tianxiaomo. Pytorch-YOLOv4. GitHub, 2020. https://github.com/Tianxiaomo/pytorch-YOLOv4. [Accessed 1 Dec 2022].
• [30] Y. Wang, X. Liang, and S. Liao. Cloning outfits from real-world images to 3D characters for generalizable person re-identification. In Proc. 2022 IEEE/CVF Conf. Computer Vision and Pattern Recognition (CVPR 2022), pages 4890-4899, New Orleans, LA, USA, 18-24 Jun 2022. doi:10.1109/CVPR52688.2022.00485.
• [31] Y. Wang, S. Liao, and L. Shao. Surpassing real-world source training data: Random 3d characters for generalizable person re-identification. In Proc. 28th ACM Int. Conf. Multimedia (MM ’20), Seattle, WA, USA, 12-16 Oct 2020. doi:10.1145/3394171.3413815.
• [32] L. Wei, S. Zhang, W. Gao, and Tian Q. Person transfer GAN to bridge domain gap for person re-identification. In Proc. 2018 IEEE/CVF Conf. Computer Vision and Pattern Recognition (CVPR 2018), pages 79-88, Salt Lake City, UT, USA, 18-23 Jun 2018. doi:10.1109/CVPR.2018.00016.
• [33] A. Wu, W. Zheng, H. Yu, et al. RGB-infrared cross-modality person re-identification. In Proc. 2017 IEEE Int. Conf. Computer Vision (ICCV 2017), pages 5390- 5399, Venice, Italy, 22-29 Oct 2017. doi:10.1109/ICCV.2017.575.
• [34] D. Wu, S.-J. Zheng, X.-P. Zhang, et al. Deep learning-based methods for person re-identification: A comprehensive review. Neurocomputing, 337: 354-371, 2019. doi:10.1016/j.neucom.2019.01.079.
• [35] T. Xiao, S. Li, B. Wang, et al. Joint detection and identification feature learning for person search. In Proc. 2017 IEEE Conf. Computer Vision and Pattern Recognition (CVPR 2017), pages 3376-3385, Honolulu, HI, USA, 21-26 Jul 2017. doi:10.1109/CVPR.2017.360.
• [36] T. Xiao, S. Li, B. Wang, L. Lin, and X. Wang. End-to-end deep learning for person search. Xiaogang Wang: personal web page, 2016. http://www.ee.cuhk.edu.hk/~xgwang/PS/paper.pdf.
• [37] M. Ye, J. Shen, G. Lin, et al. Deep learning for person re-identification: A survey and out-look. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(6): 2872-2893, 2020. doi:10.1109/TPAMI.2021.3054775.
• [38] T. Zhang, L. Xie, L. Wei, et al. UnrealPerson: An adaptive pipeline towards costless person re-identification. In Proc. 2021 IEEE/CVF Conf. Computer Vision and Pattern Recognition (CVPR 2021), pages 11501-11510, Nashville, TN, USA, 20-25 Jun 2021. doi:10.1109/CVPR46437.2021.01134.
• [39] F. Zhao, S. Liao, G. Xie, et al. Unsupervised domain adaptation with noise resistible mutual training for person re-identification. In A. Vedaldi et al., editors, Computer Vision - Proc. European Conf. Computer Vision (ECCV 2020), volume 12356 of Lecture Notes in Computer Science, pages 526-544, Glasgow, United Kingdom, 23-28 Aug 2020. doi:10.1007/978-3-030-58621-8_31.
• [40] L. Zheng, Z. Bie, Y. Sun, et al. MARS: A video benchmark for large-scale person re-identification. In B. Leibe et al., editors, Computer Vision - Proc. European Conf. Computer Vision (ECCV 2016), volume 9910 of Lecture Notes in Computer Science, pages 868-884, Amsterdam, The Netherlands, 11-14 Oct 2016. doi:10.1007/978-3-319-46466-4_52.
• [41] L. Zheng, L. Shen, L. Tian, et al. Scalable person re-identification: A benchmark. In Proc. 2015 IEEE Int. Conf. Computer Vision (ICCV 2015), pages 1116-1124, Santiago, Chile, 7-13 Dec 2015. doi:10.1109/ICCV.2015.133.
• [42] L. Zheng, H. Zhang, S. Sun, et al. Person re-identification in the wild. In Proc. 2017 IEEE Conf. Computer Vision and Pattern Recognition (CVPR 2017), pages 3346-3355, Honolulu, HI, USA, 21-26 Jul 2017. doi:10.1109/CVPR.2017.357.
• [43] Z. Zheng, X. Yang, Z. Yu, et al. Joint discriminative and generative learning for person re-identification. In Proc. 2019 IEEE Conf. Computer Vision and Pattern Recognition (CVPR 2019), pages 2133- 2142, Long Beach, CA, USA, 15-20 Jun 2019. doi:10.1109/CVPR.2019.00224.
• [44] layumi (Z. Zheng). Person_reID_baseline_pytorch. GitHub. https://github.com/layumi/Person_reID_baseline_pytorch. [Accessed 1 Dec 2022].
• [45] Z. Zhong, L. Zheng, D. Cao, and S. Li. Re-ranking person re-identification with k-reciprocal encoding. In Proc. 2017 IEEE Conf. Computer Vision and Pattern Recognition (CVPR 2017), pages 3652-3661, Honolulu, HI, USA, 21-26 Jul 2017. doi:10.1109/CVPR.2017.389.

Uwagi

Opracowanie rekordu ze środków MEiN, umowa nr SONP/SP/546092/2022 w ramach programu "Społeczna odpowiedzialność nauki" - moduł: Popularyzacja nauki i promocja sportu (2022-2023).