Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
139 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
46 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

RankED: Addressing Imbalance and Uncertainty in Edge Detection Using Ranking-based Losses (2403.01795v2)

Published 4 Mar 2024 in cs.CV

Abstract: Detecting edges in images suffers from the problems of (P1) heavy imbalance between positive and negative classes as well as (P2) label uncertainty owing to disagreement between different annotators. Existing solutions address P1 using class-balanced cross-entropy loss and dice loss and P2 by only predicting edges agreed upon by most annotators. In this paper, we propose RankED, a unified ranking-based approach that addresses both the imbalance problem (P1) and the uncertainty problem (P2). RankED tackles these two problems with two components: One component which ranks positive pixels over negative pixels, and the second which promotes high confidence edge pixels to have more label certainty. We show that RankED outperforms previous studies and sets a new state-of-the-art on NYUD-v2, BSDS500 and Multi-cue datasets. Code is available at https://ranked-cvpr24.github.io.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (55)
  1. Contour detection and hierarchical image segmentation. IEEE Trans. Pattern Anal. Mach. Intell., 33(5):898–916, 2010.
  2. Deepedge: A multi-scale bifurcated deep network for top-down contour detection. In IEEE Conf. Comput. Vis. Pattern Recog., pages 4380–4389, 2015a.
  3. High-for-low and low-for-high: Efficient boundary detection from deep object features and its applications to high-level vision. In Int. Conf. Comput. Vis., pages 504–512, 2015b.
  4. John Canny. A computational approach to edge detection. IEEE Trans. Pattern Anal. Mach. Intell., 8(6):679–698, 1986.
  5. Towards accurate one-stage object detection with ap-loss. In IEEE Conf. Comput. Vis. Pattern Recog., pages 5119–5127, 2019.
  6. AP-loss for accurate one-stage object detection. IEEE Trans. Pattern Anal. Mach. Intell., 43(11):3782–3798, 2020.
  7. MMSegmentation Contributors. MMSegmentation: Openmmlab semantic segmentation toolbox and benchmark. https://github.com/open-mmlab/mmsegmentation, 2020.
  8. Imagenet: A large-scale hierarchical image database. In IEEE Conf. Comput. Vis. Pattern Recog., pages 248–255. IEEE, 2009.
  9. Deep structural contour detection. In Proceedings of the 28th ACM international conference on multimedia, pages 304–312, 2020.
  10. Learning to predict crisp boundaries. In Eur. Conf. Comput. Vis., pages 562–578, 2018.
  11. Learning to decode contextual information for efficient contour detection. In Proceedings of the 29th ACM International Conference on Multimedia, pages 4435–4443, 2021.
  12. Learning to refine object boundaries. Neurocomputing, 557:126742, 2023.
  13. Structured forests for fast edge detection. In Int. Conf. Comput. Vis., 2013.
  14. Fast edge detection using structured forests. IEEE Trans. Pattern Anal. Mach. Intell., 37(8):1558–1570, 2014.
  15. Refined edge detection with cascaded and high-resolution convolutional network. Pattern Recognition, 138:109361, 2023.
  16. The pascal visual object classes (voc) challenge. Int. J. Comput. Vis., 88:303–338, 2010.
  17. Perceptual organization and recognition of indoor scenes from rgb-d images. In IEEE Conf. Comput. Vis. Pattern Recog., pages 564–571, 2013.
  18. Oriented edge forests for boundary detection. In IEEE Conf. Comput. Vis. Pattern Recog., pages 1732–1740, 2015.
  19. Bayesod: A bayesian approach for uncertainty estimation in deep object detectors. In 2020 IEEE International Conference on Robotics and Automation (ICRA), pages 87–93. IEEE, 2020.
  20. Bi-directional cascade network for perceptual edge detection. In IEEE Conf. Comput. Vis. Pattern Recog., pages 3828–3837, 2019.
  21. Gradient-based uncertainty for monocular depth estimation. In Eur. Conf. Comput. Vis., pages 613–630, 2022.
  22. Efficient uncertainty estimation for semantic segmentation in videos. In Eur. Conf. Comput. Vis., pages 520–535, 2018.
  23. Object contour and edge detection with refinecontournet. In Computer Analysis of Images and Patterns: 18th International Conference, pages 246–258, 2019.
  24. What uncertainties do we need in bayesian deep learning for computer vision? Adv. Neural Inform. Process. Syst., 30, 2017.
  25. Josef Kittler. On the accuracy of the sobel edge detector. Image and Vision Computing, 1(1):37–42, 1983.
  26. Iasonas Kokkinos. Pushing the boundaries of boundary detection using deep learning. arXiv preprint arXiv:1511.07386, 2015.
  27. Uncertainty estimation in one-stage object detection. In 2019 ieee intelligent transportation systems conference (itsc), pages 53–60. IEEE, 2019.
  28. Learning relaxed deep supervision for better edge detection. In IEEE Conf. Comput. Vis. Pattern Recog., pages 231–240, 2016.
  29. Richer convolutional features for edge detection. In IEEE Conf. Comput. Vis. Pattern Recog., pages 3000–3009, 2017.
  30. Swin transformer: Hierarchical vision transformer using shifted windows. In Int. Conf. Comput. Vis., pages 10012–10022, 2021.
  31. Convolutional oriented boundaries. In Eur. Conf. Comput. Vis., pages 580–596, 2016.
  32. Theory of edge detection. Proceedings of the Royal Society of London. Series B. Biological Sciences, 207(1167):187–217, 1980.
  33. A systematic comparison between visual cues for boundary detection. Vision Research, 120:93–107, 2016.
  34. A ranking-based, balanced loss function unifying classification and localisation in object detection. Adv. Neural Inform. Process. Syst., 33:15534–15545, 2020a.
  35. Imbalance problems in object detection: A review. IEEE Trans. Pattern Anal. Mach. Intell., 43(10):3388–3415, 2020b.
  36. Rank & sort loss for object detection and instance segmentation. In Int. Conf. Comput. Vis., pages 3009–3018, 2021.
  37. On the uncertainty of self-supervised monocular depth estimation. In IEEE Conf. Comput. Vis. Pattern Recog., pages 3227–3237, 2020.
  38. Rindnet: Edge detection for discontinuity in reflectance, illumination, normal and depth. In Int. Conf. Comput. Vis., pages 6879–6888, 2021.
  39. Edter: Edge detection with transformer. In IEEE Conf. Comput. Vis. Pattern Recog., pages 1402–1412, 2022.
  40. DR loss: Improving object detection by distributional ranking. In IEEE Conf. Comput. Vis. Pattern Recog., pages 12164–12172, 2020.
  41. Optimizing rank-based metrics with blackbox differentiation. In IEEE Conf. Comput. Vis. Pattern Recog., pages 7620–7630, 2020.
  42. Deepcontour: A deep convolutional feature learned by positive-sharing loss for contour detection. In IEEE Conf. Comput. Vis. Pattern Recog., pages 3982–3991, 2015.
  43. Indoor segmentation and support inference from rgbd images. In Eur. Conf. Comput. Vis., pages 746–760. Springer, 2012.
  44. Pixel difference networks for efficient edge detection. In Int. Conf. Comput. Vis., pages 5117–5127, 2021.
  45. Generalised dice overlap as a deep learning loss function for highly unbalanced segmentations. In Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support: Third International Workshop, DLMIA 2017, and 7th International Workshop, ML-CDS 2017, Held in Conjunction with MICCAI 2017, Québec City, QC, Canada, September 14, Proceedings 3, pages 240–248. Springer, 2017.
  46. Review of classification methods on unbalanced data sets. IEEE Access, 9:64606–64628, 2021.
  47. Deep crisp boundaries. In IEEE Conf. Comput. Vis. Pattern Recog., pages 3892–3900, 2017.
  48. Discriminatively trained sparse code gradients for contour detection. Adv. Neural Inform. Process. Syst., 25, 2012.
  49. Holistically-nested edge detection. In Int. Conf. Comput. Vis., pages 1395–1403, 2015.
  50. Learning deep structured multi-scale features using attention-gated crfs for contour prediction. Adv. Neural Inform. Process. Syst., 30, 2017.
  51. Fcl-net: Towards accurate edge detection via fine-scale corrective learning. Neural Networks, 145:248–259, 2022.
  52. Object contour detection with a fully convolutional encoder-decoder network. In IEEE Conf. Comput. Vis. Pattern Recog., pages 193–202, 2016.
  53. Deep long-tailed learning: A survey. IEEE Trans. Pattern Anal. Mach. Intell., 2023.
  54. Rectifying pseudo label learning via uncertainty estimation for domain adaptive semantic segmentation. Int. Conf. Comput. Vis., 129(4):1106–1120, 2021.
  55. The treasure beneath multiple annotations: An uncertainty-aware edge detector. In IEEE Conf. Comput. Vis. Pattern Recog., pages 15507–15517, 2023.
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