Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
80 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
7 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Mixed Pseudo Labels for Semi-Supervised Object Detection (2312.07006v1)

Published 12 Dec 2023 in cs.CV and cs.AI

Abstract: While the pseudo-label method has demonstrated considerable success in semi-supervised object detection tasks, this paper uncovers notable limitations within this approach. Specifically, the pseudo-label method tends to amplify the inherent strengths of the detector while accentuating its weaknesses, which is manifested in the missed detection of pseudo-labels, particularly for small and tail category objects. To overcome these challenges, this paper proposes Mixed Pseudo Labels (MixPL), consisting of Mixup and Mosaic for pseudo-labeled data, to mitigate the negative impact of missed detections and balance the model's learning across different object scales. Additionally, the model's detection performance on tail categories is improved by resampling labeled data with relevant instances. Notably, MixPL consistently improves the performance of various detectors and obtains new state-of-the-art results with Faster R-CNN, FCOS, and DINO on COCO-Standard and COCO-Full benchmarks. Furthermore, MixPL also exhibits good scalability on large models, improving DINO Swin-L by 2.5% mAP and achieving nontrivial new records (60.2% mAP) on the COCO val2017 benchmark without extra annotations.

PDF HTML Abstract
Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Bookmark Chat Bubble Oval Streamline Icon: https://streamlinehq.com Chat (Pro)
Definition Search Book Streamline Icon: https://streamlinehq.com
References (52)
  1. Mixmatch: A holistic approach to semi-supervised learning. In NeurIPS, 2019.
  2. Yolov4: Optimal speed and accuracy of object detection. CoRR, abs/2004.10934, 2020.
  3. Cascade r-cnn: high quality object detection and instance segmentation. IEEE TPAMI, 2019.
  4. End-to-end object detection with transformers. In ECCV, 2020.
  5. Label matching semi-supervised object detection. In CVPR, 2022a.
  6. Dense learning based semi-supervised object detection. In CVPR, 2022b.
  7. Hybrid task cascade for instance segmentation. In CVPR, 2019a.
  8. MMDetection: Open mmlab detection toolbox and benchmark. arXiv preprint arXiv:1906.07155, 2019b.
  9. Improved regularization of convolutional neural networks with cutout. arXiv preprint arXiv:1708.04552, 2017.
  10. The pascal visual object classes (voc) challenge. IJCV, 2010.
  11. Tood: Task-aligned one-stage object detection. In 2021 IEEE/CVF International Conference on Computer Vision (ICCV), pages 3490–3499. IEEE Computer Society, 2021.
  12. YOLOX: exceeding YOLO series in 2021. CoRR, abs/2107.08430, 2021.
  13. NAS-FPN: Learning scalable feature pyramid architecture for object detection. In CVPR, 2019.
  14. Consistency-based semi-supervised learning for object detection. In NeurIPS, 2019.
  15. Interpolation-based semi-supervised learning for object detection. In CVPR, 2021.
  16. Temporal ensembling for semi-supervised learning. arXiv preprint arXiv:1610.02242, 2016.
  17. Dong-Hyun Lee et al. Pseudo-label: The simple and efficient semi-supervised learning method for deep neural networks. In Workshop on challenges in representation learning, ICML, 2013.
  18. Gradient harmonized single-stage detector. In AAAI, 2019.
  19. Pseco: Pseudo labeling and consistency training for semi-supervised object detection. In ECCV. Springer, 2022.
  20. Feature pyramid networks for object detection. In ICCV, 2017a.
  21. Microsoft coco: Common objects in context. In European conference on computer vision, pages 740–755. Springer, 2014.
  22. Focal loss for dense object detection. In ICCV, 2017b.
  23. Dab-detr: Dynamic anchor boxes are better queries for detr. arXiv preprint arXiv:2201.12329, 2022a.
  24. Unbiased teacher for semi-supervised object detection. arXiv preprint arXiv:2102.09480, 2021a.
  25. Unbiased teacher v2: Semi-supervised object detection for anchor-free and anchor-based detectors. In CVPR, 2022b.
  26. Swin transformer: Hierarchical vision transformer using shifted windows. In ICCV, 2021b.
  27. Decoupled weight decay regularization. In ICLR. OpenReview.net, 2019.
  28. RTMDet: An empirical study of designing real-time object detectors. CoRR, abs/2212.07784, 2022.
  29. Faster r-cnn: Towards real-time object detection with region proposal networks. In NeurIPS, 2015.
  30. Grad-cam: Visual explanations from deep networks via gradient-based localization. In Proceedings of the IEEE international conference on computer vision, pages 618–626, 2017.
  31. Fixmatch: Simplifying semi-supervised learning with consistency and confidence. In NeurIPS, 2020a.
  32. A simple semi-supervised learning framework for object detection. arXiv preprint arXiv:2005.04757, 2020b.
  33. Sparse r-cnn: End-to-end object detection with learnable proposals. In CVPR, 2021.
  34. Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results. NeurIPS, 2017.
  35. Fcos: Fully convolutional one-stage object detection. In CVPR, 2019.
  36. Freematch: Self-adaptive thresholding for semi-supervised learning. arXiv preprint arXiv:2205.07246, 2022.
  37. Detectron2. https://github.com/facebookresearch/detectron2, 2019.
  38. Unsupervised data augmentation for consistency training. In NeurIPS, 2020a.
  39. Self-training with noisy student improves imagenet classification. In CVPR, 2020b.
  40. End-to-end semi-supervised object detection with soft teacher. In CVPR, 2021.
  41. Flexmatch: Boosting semi-supervised learning with curriculum pseudo labeling. In NeurIPS, pages 18408–18419, 2021.
  42. Semi-supervised object detection with adaptive class-rebalancing self-training. In AAAI, 2022a.
  43. mixup: Beyond empirical risk minimization. arXiv preprint arXiv:1710.09412, 2017.
  44. Dino: Detr with improved denoising anchor boxes for end-to-end object detection. In ICLR, 2022b.
  45. Semi-detr: Semi-supervised object detection with detection transformers. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 23809–23818, 2023.
  46. Bridging the gap between anchor-based and anchor-free detection via adaptive training sample selection. In CVPR, 2020.
  47. Simmatch: Semi-supervised learning with similarity matching. In CVPR, 2022.
  48. Random erasing data augmentation. In AAAI, 2020.
  49. Dense teacher: Dense pseudo-labels for semi-supervised object detection. In ECCV. Springer, 2022.
  50. Instant-teaching: An end-to-end semi-supervised object detection framework. In CVPR, 2021.
  51. Objects as points. arXiv preprint arXiv:1904.07850, 2019.
  52. Deformable detr: Deformable transformers for end-to-end object detection. arXiv preprint arXiv:2010.04159, 2020.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (5)
  1. Zeming Chen (18 papers)
  2. Wenwei Zhang (77 papers)
  3. Xinjiang Wang (32 papers)
  4. Kai Chen (512 papers)
  5. Zhi Wang (261 papers)
Citations (4)
View on Semantic Scholar