Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
169 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
45 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Pedestrian Attribute Recognition as Label-balanced Multi-label Learning (2405.04858v1)

Published 8 May 2024 in cs.CV

Abstract: Rooting in the scarcity of most attributes, realistic pedestrian attribute datasets exhibit unduly skewed data distribution, from which two types of model failures are delivered: (1) label imbalance: model predictions lean greatly towards the side of majority labels; (2) semantics imbalance: model is easily overfitted on the under-represented attributes due to their insufficient semantic diversity. To render perfect label balancing, we propose a novel framework that successfully decouples label-balanced data re-sampling from the curse of attributes co-occurrence, i.e., we equalize the sampling prior of an attribute while not biasing that of the co-occurred others. To diversify the attributes semantics and mitigate the feature noise, we propose a Bayesian feature augmentation method to introduce true in-distribution novelty. Handling both imbalances jointly, our work achieves best accuracy on various popular benchmarks, and importantly, with minimal computational budget.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (55)
  1. Learning transferable pedestrian representation from multimodal information supervision. arXiv preprint arXiv:2304.05554, 2023.
  2. A novel self-boosting dual-branch model for pedestrian attribute recognition. Signal Processing: Image Communication, 115:116961, 2023.
  3. Smote: synthetic minority over-sampling technique. Journal of artificial intelligence research, 16:321–357, 2002.
  4. Improving energy-based out-of-distribution detection by sparsity regularization. In Pacific-Asia Conference on Knowledge Discovery and Data Mining, pp.  539–551. Springer, 2022.
  5. Upar challenge: Pedestrian attribute recognition and attribute-based person retrieval–dataset, design, and results. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pp.  166–175, 2023.
  6. Autoaugment: Learning augmentation strategies from data. In 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2019.
  7. Pedestrian attribute recognition at far distance. ACM, 2014.
  8. Dataset augmentation in feature space. arXiv preprint arXiv:1702.05538, 2017a.
  9. Improved regularization of convolutional neural networks with cutout. arXiv preprint arXiv:1708.04552, 2017b.
  10. Generative adversarial models for people attribute recognition in surveillance. In 2017 14th IEEE international conference on advanced video and signal based surveillance (AVSS), pp.  1–6. IEEE, 2017.
  11. Correlation graph convolutional network for pedestrian attribute recognition. IEEE Transactions on Multimedia, PP(99):1–1, 2020.
  12. Parformer: Transformer-based multi-task network for pedestrian attribute recognition. IEEE Transactions on Circuits and Systems for Video Technology, 2023.
  13. Dropout as a bayesian approximation: Representing model uncertainty in deep learning. In international conference on machine learning, pp. 1050–1059. PMLR, 2016.
  14. Concrete dropout. Advances in neural information processing systems, 30, 2017.
  15. Long-tailed multi-label visual recognition by collaborative training on uniform and re-balanced samplings. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.  15089–15098, 2021.
  16. Visual attention consistency under image transforms for multi-label image classification. In 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2020.
  17. Finding structure with randomness: Probabilistic algorithms for constructing approximate matrix decompositions. SIAM review, 53(2):217–288, 2011.
  18. Spatial and semantic consistency regularizations for pedestrian attribute recognition. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pp.  962–971, October 2021a.
  19. Rethinking of pedestrian attribute recognition: Realistic datasets with efficient method. arXiv, 2020.
  20. Rethinking of pedestrian attribute recognition: A reliable evaluation under zero-shot pedestrian identity setting. arXiv preprint arXiv:2107.03576, 2021b.
  21. Learning disentangled attribute representations for robust pedestrian attribute recognition. pp.  1069–1077. AAAI Press, 2022.
  22. Decoupling representation and classifier for long-tailed recognition. arXiv preprint arXiv:1910.09217, 2019.
  23. Auto-encoding variational bayes. arXiv preprint arXiv:1312.6114, 2013.
  24. Multi-attribute learning for pedestrian attribute recognition in surveillance scenarios. In 2015 3rd IAPR Asian Conference on Pattern Recognition (ACPR), pp.  111–115. IEEE, 2015.
  25. A richly annotated dataset for pedestrian attribute recognition. arXiv preprint arXiv:1603.07054, 2016.
  26. Learning deep context-aware features over body and latent parts for person re-identification. IEEE, 2017.
  27. Metasaug: Meta semantic augmentation for long-tailed visual recognition. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.  5212–5221, 2021.
  28. Label2label: A language modeling framework for multi-attribute learning. In Computer Vision–ECCV 2022: 17th European Conference, Tel Aviv, Israel, October 23–27, 2022, Proceedings, Part XII, pp.  562–579. Springer, 2022.
  29. Localization guided learning for pedestrian attribute recognition. arXiv preprint arXiv:1808.09102, 2018.
  30. Large-margin softmax loss for convolutional neural networks. arXiv preprint arXiv:1612.02295, 2016.
  31. Hydraplus-net: Attentive deep features for pedestrian analysis. In Proceedings of the IEEE international conference on computer vision, pp.  350–359, 2017.
  32. A convnet for the 2020s. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.  11976–11986, 2022.
  33. Orientation-aware pedestrian attribute recognition based on graph convolution network. IEEE Transactions on Multimedia, 2023.
  34. Out of distribution data detection using dropout bayesian neural networks. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 36, pp.  7877–7885, 2022.
  35. Upar: Unified pedestrian attribute recognition and person retrieval. ArXiv, abs/2209.02522, 2022.
  36. Relation-aware pedestrian attribute recognition with graph convolutional networks. In Proceedings of the AAAI conference on artificial intelligence, volume 34, pp.  12055–12062, 2020a.
  37. Relation-aware pedestrian attribute recognition with graph convolutional networks. Proceedings of the AAAI Conference on Artificial Intelligence, 34(7):12055–12062, 2020b.
  38. Drformer: Learning dual relations using transformer for pedestrian attribute recognition. Neurocomputing, 497:159–169, 2022.
  39. Rethinking feature distribution for loss functions in image classification. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.  9117–9126, 2018.
  40. Discovering visual concept structure with sparse and incomplete tags. Artificial Intelligence, 250:16–36, 2017.
  41. Pedestrian attribute recognition: A survey. Pattern Recognition, 121:108220, 2022.
  42. Implicit semantic data augmentation for deep networks. Advances in Neural Information Processing Systems, 32, 2019.
  43. Exploring attribute localization and correlation for pedestrian attribute recognition. Neurocomputing, 531:140–150, 2023.
  44. Inter-attribute awareness for pedestrian attribute recognition. Pattern Recognition, 131:108865, 2022.
  45. Adaptive class-balanced loss based on re-weighting. In 2022 6th Asian Conference on Artificial Intelligence Technology (ACAIT), pp.  1–8. IEEE, 2022.
  46. Understanding deep learning (still) requires rethinking generalization. Communications of the ACM, 64(3):107–115, 2021a.
  47. Weakly supervised object localization and detection: A survey. IEEE transactions on pattern analysis and machine intelligence, 44(9):5866–5885, 2021b.
  48. mixup: Beyond empirical risk minimization. arXiv preprint arXiv:1710.09412, 2017.
  49. Distribution alignment: A unified framework for long-tail visual recognition (supplementary material). 2021c.
  50. Bag of tricks for long-tailed visual recognition with deep convolutional neural networks. In Proceedings of the AAAI conference on artificial intelligence, volume 35, pp.  3447–3455, 2021d.
  51. Deep long-tailed learning: A survey. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023.
  52. Diverse features discovery transformer for pedestrian attribute recognition. Engineering Applications of Artificial Intelligence, 119:105708, 2023.
  53. Person re-identification meets image search. arXiv preprint arXiv:1502.02171, 2015.
  54. Zhou, Y. Rethinking reconstruction autoencoder-based out-of-distribution detection. Proceedings of the IEEE conference on computer vision and pattern recognition, 2022.
  55. A solution to co-occurrence bias: Attributes disentanglement via mutual information minimization for pedestrian attribute recognition. arXiv preprint arXiv:2307.15252, 2023.
Citations (1)
View on Semantic Scholar

Summary

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

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

Tweets