Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
110 tokens/sec
GPT-4o
56 tokens/sec
Gemini 2.5 Pro Pro
44 tokens/sec
o3 Pro
6 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Source-Guided Similarity Preservation for Online Person Re-Identification (2402.15206v1)

Published 23 Feb 2024 in cs.CV

Abstract: Online Unsupervised Domain Adaptation (OUDA) for person Re-Identification (Re-ID) is the task of continuously adapting a model trained on a well-annotated source domain dataset to a target domain observed as a data stream. In OUDA, person Re-ID models face two main challenges: catastrophic forgetting and domain shift. In this work, we propose a new Source-guided Similarity Preservation (S2P) framework to alleviate these two problems. Our framework is based on the extraction of a support set composed of source images that maximizes the similarity with the target data. This support set is used to identify feature similarities that must be preserved during the learning process. S2P can incorporate multiple existing UDA methods to mitigate catastrophic forgetting. Our experiments show that S2P outperforms previous state-of-the-art methods on multiple real-to-real and synthetic-to-real challenging OUDA benchmarks.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (51)
  1. Adapting to continuously shifting domains. In ICLR, 2018.
  2. Incremental learning in semantic segmentation from image labels. In CVPR, 2022.
  3. Unsupervised lifelong person re-identification via contrastive rehearsal. arXiv preprint arXiv:2203.06468, 2022.
  4. Instance-guided context rendering for cross-domain person re-identification. In ICCV, 2019.
  5. IDM: an intermediate domain module for domain adaptive person re-id. In ICCV, 2021.
  6. CANU-ReID: a conditional adversarial network for unsupervised person re-identification. In ICPR, 2021.
  7. ImageNet: A large-scale hierarchical image database. In CVPR, 2009.
  8. Image-image domain adaptation with preserved self-similarity and domain-dissimilarity for person re-identification. In CVPR, 2018.
  9. Unsupervised person re-identification: Clustering and fine-tuning. ACM TOMM, 2018.
  10. Complementary pseudo labels for unsupervised domain adaptation on person re-identification. IEEE TIP, 2021.
  11. Online continual learning under extreme memory constraints. In ECCV, 2020.
  12. Self-similarity grouping: A simple unsupervised cross domain adaptation approach for person re-identification. In ICCV, 2019.
  13. Mutual mean-teaching: Pseudo label refinery for unsupervised domain adaptation on person re-identification. In ICLR, 2020.
  14. Self-paced contrastive learning with hybrid memory for domain adaptive object re-id. In NeurIPS, 2020.
  15. Structured domain adaptation with online relation regularization for unsupervised person re-id. IEEE TNNLS, 2022.
  16. Connecting the dots with landmarks: Discriminatively learning domain-invariant features for unsupervised domain adaptation. In ICML, 2013.
  17. A kernel method for the two-sample-problem. NeurIPS, 2006.
  18. Deep residual learning for image recognition. In CVPR, 2016.
  19. Distilling the knowledge in a neural network. In NeurIPSW, 2014.
  20. Lifelong unsupervised domain adaptive person re-identification with coordinated anti-forgetting and adaptation. In CVPR, 2022.
  21. Compacting, picking and growing for unforgetting continual learning. In NeurIPS, 2019.
  22. Understanding catastrophic forgetting and remembering in continual learning with optimal relevance mapping. In NeurIPSW, 2021.
  23. Adam: A method for stochastic optimization. In ICLR, 2015.
  24. Temporal ensembling for semi-supervised learning. In ICLR, 2017.
  25. DeepReID: Deep filter pairing neural network for person re-identification. In CVPR, 2014.
  26. Generative moment matching networks. In ICML, 2015.
  27. Learning without forgetting. In ECCV, 2016.
  28. A bottom-up clustering approach to unsupervised person re-identification. In AAAI, 2019.
  29. Adaptive transfer network for cross-domain person re-identification. In CVPR, 2019.
  30. Bag of tricks and a strong baseline for deep person re-identification. In CVPRW, 2019.
  31. Lifelong person re-identification via adaptive knowledge accumulation. In CVPR, 2021.
  32. Online unsupervised domain adaptation for person re-identification. In CVPRW, 2022.
  33. Unsupervised domain adaptive re-identification: Theory and practice. Pattern Recognition, 2020.
  34. Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results. In NeurIPS, 2017.
  35. Extending conditional convolution structures for enhancing multitasking continual learning. In APSIPA ASC, 2020.
  36. Similarity-preserving knowledge distillation. In ICCV, 2019.
  37. Surpassing real-world source training data: Random 3D characters for generalizable person re-identification. In ACM MM, 2020.
  38. Person transfer GAN to bridge domain gap for person re-identification. In CVPR, 2018.
  39. Camera-conditioned stable feature generation for isolated camera supervised person re-identification. In CVPR, 2022.
  40. Memory replay GANs: learning to generate images from new categories without forgetting. In NeurIPS, 2018.
  41. Lifelong teacher-student network learning. IEEE TPAMI, 2021.
  42. Dynamic graph co-matching for unsupervised video-based person re-identification. IEEE TIP, 2019.
  43. Deep learning for person re-identification: A survey and outlook. IEEE TPAMI, 2022.
  44. Lifelong learning with dynamically expandable networks. In ICLR, 2018.
  45. Paying more attention to attention: Improving the performance of convolutional neural networks via attention transfer. ICLR, 2016.
  46. Lifelong GAN: continual learning for conditional image generation. In ICCV, 2019.
  47. Scalable person re-identification: A benchmark. In ICCV, 2015.
  48. Re-ranking person re-identification with k-reciprocal encoding. In CVPR, 2017.
  49. Generalizing a person retrieval model hetero- and homogeneously. In ECCV, 2018.
  50. Lifelong object detection. arXiv preprint arXiv:2009.01129, 2020.
  51. Unpaired image-to-image translation using cycle-consistent adversarial networks. In ICCV, 2017.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (4)
  1. Hamza Rami (3 papers)
  2. Jhony H. Giraldo (30 papers)
  3. Nicolas Winckler (4 papers)
  4. Stéphane Lathuilière (79 papers)
Citations (3)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now