Abstract
Person re-identification (re-ID) is commonly investigated as a ranking problem. However, the performance of existing re-ID models drops dramatically, when they encounter extreme positive-negative class imbalance (e.g., very small ratio of positive and negative samples) during training. To alleviate this problem, this article designs a rank-in-rank loss to optimize the distribution of feature embeddings. Specifically, we propose a Differentiable Retrieval-Sort Loss (DRSL) to optimize the re-ID model by ranking each positive sample ahead of the negative samples according to the distance and sorting the positive samples according to the angle (e.g., similarity score). The key idea of the proposed DRSL lies in minimizing the distance between samples of the same category along with the angle between them. Considering that the ranking and sorting operations are non-differentiable and non-convex, the DRSL also performs the optimization of automatic derivation and backpropagation. In addition, the analysis of the proposed DRSL is provided to illustrate that the DRSL not only maintains the inter-class distance distribution but also preserves the intra-class similarity structure in terms of angle constraints. Extensive experimental results indicate that the proposed DRSL can improve the performance of the state-of-the-art re-ID models, thus demonstrating its effectiveness and superiority in the re-ID task.
- [1] . 2015. An improved deep learning architecture for person re-identification. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 3908–3916.
DOI: Google ScholarCross Ref
- [2] . 2020. Deep-person: Learning discriminative deep features for person re-identification. Pattern Recognition 98 (2020), 107036.
DOI: Google ScholarCross Ref
- [3] . 2019. Boosting search performance using query variations. ACM Transactions on Information Systems 37, 4 (2019), 1–25.
DOI: Google ScholarCross Ref
- [4] . 2020. Smooth-ap: Smoothing the path towards large-scale image retrieval. In Proceedings of the European Conference on Computer Vision. 677–694.
DOI: Google ScholarCross Ref
- [5] . 2006. Learning to rank with nonsmooth cost functions. In Proceedings of the Advances in Neural Information Processing Systems. 193–200.Google Scholar
- [6] . 2019. Deep metric learning to rank. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 1861–1870.
DOI: Google ScholarCross Ref
- [7] . 2018. Multi-level factorisation net for person re-identification. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2109–2118.
DOI: Google ScholarCross Ref
- [8] . 2007. Large margin optimization of ranking measures. In Proceedings of the NIPS Workshop: Machine Learning for Web Search.Google Scholar
- [9] . 2021. Person re-identification via attention pyramid. IEEE Transactions on Image Processing 30 (2021), 7663–7676.
DOI: Google ScholarCross Ref
- [10] . 2020. Deep credible metric learning for unsupervised domain adaptation person re-identification. In Proceedings of the European Conference on Computer Vision. 643–659.
DOI: Google ScholarCross Ref
- [11] . 2018. Deep transfer learning for person re-identification. In Proceedings of the 2018 IEEE 4th International Conference on Multimedia Big Data. IEEE, 1–5.Google Scholar
Cross Ref
- [12] . 2019. Towards accurate one-stage object detection with ap-loss. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 5119–5127.
DOI: Google ScholarCross Ref
- [13] Kean Chen, Weiyao Lin, Jianguo Li, John See, Ji Wang, and Junni Zou. 2021. AP-loss for accurate one-stage object detection. IEEE Transactions on Pattern Analysis and Machine Intelligence 43, 11 (2021), 3782–3798.
DOI: Google ScholarCross Ref
- [14] . 2017. Beyond triplet loss: A deep quadruplet network for person re-identification. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 403–412.
DOI: Google ScholarCross Ref
- [15] . 2021. Mobile-former: Bridging mobilenet and transformer. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 5270–5279.Google Scholar
- [16] . 2009. Imagenet: A large-scale hierarchical image database. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 248–255.
DOI: Google ScholarCross Ref
- [17] . 2018. Image-image domain adaptation with preserved self-similarity and domain-dissimilarity for person re-identification. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 994–1003.
DOI: Google ScholarCross Ref
- [18] . 2019. Horizontal pyramid matching for person re-identification. In Proceedings of the AAAI Conference on Artificial Intelligence. 8295–8302.
DOI: Google ScholarCross Ref
- [19] . 2007. Evaluating appearance models for recognition, reacquisition, and tracking. In Proceedings of the IEEE International Workshop on Performance Evaluation for Tracking and Surveillance. Citeseer, 1–7.Google Scholar
- [20] . 2019. Beyond human parts: Dual part-aligned representations for person re-identification. In Proceedings of the IEEE/CVF International Conference on Computer Vision. 3642–3651.
DOI: Google ScholarCross Ref
- [21] . 2018. Efficient and deep person re-identification using multi-level similarity. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2335–2344.
DOI: Google ScholarCross Ref
- [22] . 2018. Hashing as tie-aware learning to rank. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 4023–4032.
DOI: Google ScholarCross Ref
- [23] . 2017. In defense of the triplet loss for person re-identification. arXiv:1703.07737. Retrieved from https://arxiv.org/abs/1703.07737.Google Scholar
- [24] . 2021. Combined depth space based architecture search for person re-identification. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 6729–6738.
DOI: Google ScholarCross Ref
- [25] . 2014. A maximal figure-of-merit learning approach to maximizing mean average precision with deep neural network based classifiers. In Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing. IEEE, 4503–4507.
DOI: Google ScholarCross Ref
- [26] . 2014. Deepreid: Deep filter pairing neural network for person re-identification. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 152–159.
DOI: Google ScholarCross Ref
- [27] . 2018. Harmonious attention network for person re-identification. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2285–2294.
DOI: Google ScholarCross Ref
- [28] . 2020. Scalable person re-identification by harmonious attention. International Journal of Computer Vision 128, 6 (2020), 1635–1653.
DOI: Google ScholarCross Ref
- [29] . 2019. Improving person re-identification by attribute and identity learning. Pattern Recognition 95 (2019), 151–161.
DOI: Google ScholarCross Ref
- [30] . 2022. Generative metric learning for adversarially robust open-world person re-identification. ACM Transactions on Multimedia Computing, Communications, and Applications 1, 1 (2022), 1–20.
DOI: Google ScholarCross Ref
- [31] . 2017. Sphereface: Deep hypersphere embedding for face recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 212–220.
DOI: Google ScholarCross Ref
- [32] . 2019. A strong baseline and batch normalization neck for deep person re-identification. IEEE Transactions on Multimedia 22, 10 (2019), 2597–2609.
DOI: Google ScholarCross Ref
- [33] . 2017. No fuss distance metric learning using proxies. In Proceedings of the IEEE International Conference on Computer Vision. 360–368.
DOI: Google ScholarCross Ref
- [34] . 2016. Deep metric learning via lifted structured feature embedding. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 4004–4012.
DOI: Google ScholarCross Ref
- [35] . 2018. Localization recall precision (LRP): A new performance metric for object detection. In Proceedings of the European Conference on Computer Vision. 504–519.
DOI: Google ScholarCross Ref
- [36] . 2020. A ranking-based, balanced loss function unifying classification and localisation in object detection. In Proceedings of the Advances in Neural Information Processing Systems. 15534–15545.Google Scholar
- [37] . 2021. Rank & sort loss for object detection and instance segmentation. In Proceedings of the IEEE International Conference on Computer Vision. 3009–3018.
DOI: Google ScholarCross Ref
- [38] Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, Alban Desmaison, Andreas Kopf, Edward Yang, Zachary DeVito, Martin Raison, Alykhan Tejani, Sasank Chilamkurthy, Benoit Steiner, Lu Fang, Junjie Bai, and Soumith Chintala. 2019. Pytorch: An imperative style, high-performance deep learning library. In Proceedings of the Advances in Neural Information Processing Systems. 8024–8035.Google Scholar
- [39] . 2010. A general approximation framework for direct optimization of information retrieval measures. Information Retrieval 13, 4 (2010), 375–397.
DOI: Google ScholarCross Ref
- [40] . 2019. Auto-reid: Searching for a part-aware convnet for person re-identification. In Proceedings of the IEEE International Conference on Computer Vision. 3750–3759.
DOI: Google ScholarCross Ref
- [41] . 2021. Resilience in multi-robot multi-target tracking with unknown number of targets through reconfiguration. IEEE Transactions on Control of Network Systems 8, 2 (2021), 609–620.
DOI: Google ScholarCross Ref
- [42] . 2019. Learning with average precision: Training image retrieval with a listwise loss. In Proceedings of the IEEE/CVF International Conference on Computer Vision. 5107–5116.
DOI: Google ScholarCross Ref
- [43] . 2018. Features for multi-target multi-camera tracking and re-identification. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 6036–6046.
DOI: Google ScholarCross Ref
- [44] . 2020. Optimizing rank-based metrics with blackbox differentiation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 7620–7630.
DOI: Google ScholarCross Ref
- [45] . 2018. A pose-sensitive embedding for person re-identification with expanded cross neighborhood re-ranking. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 420–429.
DOI: Google ScholarCross Ref
- [46] . 2015. Facenet: A unified embedding for face recognition and clustering. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 815–823.
DOI: Google ScholarCross Ref
- [47] . 2016. Improved deep metric learning with multi-class n-pair loss objective. In Proceedings of the Advances in Neural Information Processing Systems. 1857–1865. Retrieved from https://proceedings.neurips.cc/paper/2016/hash/6b180037abbebea991d8b1232f8a8ca9-Abstract.html.Google Scholar
- [48] . 2018. Part-aligned bilinear representations for person re-identification. In Proceedings of the European Conference on Computer Vision. 402–419.
DOI: Google ScholarCross Ref
- [49] . 2018. Beyond part models: Person retrieval with refined part pooling (and a strong convolutional baseline). In Proceedings of the European Conference on Computer Vision. 480–496.
DOI: Google ScholarCross Ref
- [50] . 2021. Incomplete descriptor mining with elastic loss for person re-identification. IEEE Transactions on Circuits and Systems for Video Technology 14, 8 (2021), 1–12.
DOI: Google ScholarCross Ref
- [51] . 2016. Learning deep embeddings with histogram loss. In Proceedings of the Advances in Neural Information Processing Systems, Vol. 29. 4170–4178. http://papers.nips.cc/paper/6464-learning-deep-embeddings-with-histogram-loss.Google Scholar
- [52] . 2016. Joint learning of single-image and cross-image representations for person re-identification. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 1288–1296.
DOI: Google ScholarCross Ref
- [53] . 2018. Learning discriminative features with multiple granularities for person re-identification. In Proceedings of the 26th ACM International Conference on Multimedia. 274–282.
DOI: Google ScholarCross Ref
- [54] . 2012. Movie2comics: Towards a lively video content presentation. IEEE Transactions on Multimedia 14, 3–2 (2012), 858–870.
DOI: Google ScholarCross Ref
- [55] . 2012. Multimodal graph-based reranking for web image search. IEEE Transactions on Image Processing 21, 11 (2012), 4649–4661.
DOI: Google ScholarCross Ref
- [56] . 2019. Ranked list loss for deep metric learning. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 5207–5216.
DOI: Google ScholarCross Ref
- [57] . 2020. Re-identification= retrieval+ verification: Back to essence and forward with a new metric. arXiv:2011.11506. Retrieved from https://arxiv.org/abs/2011.11506.Google Scholar
- [58] . 2018. Person transfer gan to bridge domain gap for person re-identification. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 79–88.
DOI: Google ScholarCross Ref
- [59] . 2016. A discriminative feature learning approach for deep face recognition. In Proceedings of the European Conference on Computer Vision. Springer, 499–515.
DOI: Google ScholarCross Ref
- [60] . 2019. Cross-entropy adversarial view adaptation for person re-identification. IEEE Transactions on Circuits and Systems for Video Technology 30, 7 (2019), 2081–2092.
DOI: Google ScholarCross Ref
- [61] . 2020. Deep coattention-based comparator for relative representation learning in person re-identification. IEEE Transactions on Neural Networks and Learning Systems 32, 2 (2020), 722–735.
DOI: Google ScholarCross Ref
- [62] . 2019. Few-shot deep adversarial learning for video-based person re-identification. IEEE Transactions on Image Processing 29 (2019), 1233–1245.
DOI: Google ScholarCross Ref
- [63] . 2021. Unsupervised video person re-identification via noise and hard frame aware clustering. In Proceedings of the 2021 IEEE International Conference on Multimedia and Expo. IEEE, 1–6.
DOI: Google ScholarCross Ref
- [64] . 2021. Rethinking data collection for person re-identification: Active redundancy reduction. Pattern Recognition 113 (2021), 107827.
DOI: Google ScholarCross Ref
- [65] . 2019. Attention driven person re-identification. Pattern Recognition 86 (2019), 143–155.
DOI: Google ScholarCross Ref
- [66] . 2020. Learning to match on graph for fashion compatibility modeling. In Proceedings of the AAAI Conference on Artificial Intelligence. 287–294.
DOI: Google ScholarCross Ref
- [67] . 2017. Enhancing person re-identification in a self-trained subspace. ACM Transactions on Multimedia Computing, Communications, and Applications 13, 3 (2017), 1–23.
DOI: Google ScholarCross Ref
- [68] . 2017. Person re-identification with metric learning using privileged information. IEEE Transactions on Image Processing 27, 2 (2017), 791–805.
DOI: Google ScholarCross Ref
- [69] . 2021. Equidistant distribution loss for person re-identification. Neurocomputing 455 (2021), 255–264.
DOI: Google ScholarCross Ref
- [70] Mang Ye, Jianbing Shen, Gaojie Lin, Tao Xiang, Ling Shao, and Steven C. H. Hoi. 2022. Deep Learning for Person Re-identification: A Survey and Outlook. IEEE Transactions on Pattern Analysis and Machine Intelligence 44, 6 (2022), 2872–2893.
DOI: Google ScholarCross Ref
- [71] . 2020. Resource scheduling for distributed multi-target tracking in netted colocated MIMO radar systems. IEEE Transactions on Signal Processing 68 (2020), 1602–1617.
DOI: Google ScholarCross Ref
- [72] . 2018. Hard-aware point-to-set deep metric for person re-identification. In Proceedings of the European Conference on Computer Vision. 188–204.
DOI: Google ScholarCross Ref
- [73] . 2020. In defense of the triplet loss again: Learning robust person re-identification with fast approximated triplet loss and label distillation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops. 354–355.
DOI: Google ScholarCross Ref
- [74] . 2019. In defense of the classification loss for person re-identification. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops. 1526–1535.
DOI: Google ScholarCross Ref
- [75] . 2019. Densely semantically aligned person re-identification. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 667–676.
DOI: Google ScholarCross Ref
- [76] . 2020. Deep fusion feature representation learning with hard mining center-triplet loss for person re-identification. IEEE Transactions on Multimedia 22, 12 (2020), 3180–3195.
DOI: Google ScholarCross Ref
- [77] . 2017. Spindle net: Person re-identification with human body region guided feature decomposition and fusion. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 1077–1085.
DOI: Google ScholarCross Ref
- [78] . 2015. Scalable person re-identification: A benchmark. In Proceedings of the IEEE International Conference on Computer Vision. 1116–1124.
DOI: Google ScholarCross Ref
- [79] . 2017. Person re-identification in the wild. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 1367–1376.Google Scholar
Cross Ref
- [80] . 2017. A discriminatively learned cnn embedding for person reidentification. ACM Transactions on Multimedia Computing, Communications, and Applications 14, 1 (2017), 1–20.
DOI: Google ScholarCross Ref
- [81] . 2017. Unlabeled samples generated by gan improve the person re-identification baseline in vitro. In Proceedings of the IEEE International Conference on Computer Vision. 3754–3762.
DOI: Google ScholarCross Ref
- [82] . 2017. Re-ranking person re-identification with k-reciprocal encoding. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 3652–3661.
DOI: Google ScholarCross Ref
- [83] . 2021. Learning generalisable omni-scale representations for person re-identification. IEEE Transactions on Pattern Analysis and Machine Intelligence (2021), 1–14.
DOI: Google ScholarCross Ref
Index Terms
Rank-in-Rank Loss for Person Re-identification
Recommendations
RA Loss: Relation-Aware Loss for Robust Person Re-identification
Computer Vision – ACCV 2022AbstractPrevious relation-based losses in person re-identification (ReID) typically comprise two sequential steps: they firstly sample both positive pair and negative pair and then deploy constraints to simultaneously improve intra-identity compactness ...
Support Neighbor Loss for Person Re-Identification
MM '18: Proceedings of the 26th ACM international conference on MultimediaPerson re-identification (re-ID) has recently been tremendously boosted due to the advancement of deep convolutional neural networks (CNN). The majority of deep re-ID methods focus on designing new CNN architectures, while less attention is paid on ...
TriEP: Expansion-Pool TriHard Loss for Person Re-Identification
AbstractPerson re-identification aims to identify the same person across different cameras, which is widely applied in the intelligent monitoring field. The research of TriHard loss has been verified to improve performance efficiently in the person re-...






Comments