Skip to main content
SpringerLink
Log in

How to use extra training data for better edge detection?

  • Published:
Applied Intelligence Aims and scope Submit manuscript

Abstract

Prevalent paradigms for edge detection tend to use extra data in a mixed training manner, which can increase the data diversity of training samples; however, a part of extra data may improve their performances, while the other will degrade their performances. This paper first proposes a selective training method to select positive data for improving the corresponding performance. Secondly, to properly use the negative data which may degrade the performance in previous schemes, an adaptive training method is also proposed by introducing a similarity-preserving self-distillation mechanism based on a teacher-student network, which can maintain equivalent performance to the selective one, and simultaneously achieves better generalization ability. Compared with the state-of-the-art method in CVPR 2022, our schemes achieve 0.5% and 7.6% improvement of the ODS F-scores for the data sets BSDS and Pascal, respectively. Experimental results on the benchmarks verify the effectiveness of the proposed schemes. Our source code and model parameters will be released at https://github.com/wenya1994/Boosting-ED.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Algorithm 1
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10

Similar content being viewed by others

Data Availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

References

  1. Al-Amaren A, Ahmad MO, Swamy M (2022) A low-complexity residual deep neural network for image edge detection. Appl Intell 1–18

  2. Arbelaez P, Maire M, Fowlkes C et al (2010) Contour detection and hierarchical image segmentation. IEEE Trans Pattern Anal Mach Intell 33(5):898–916

    Article  Google Scholar 

  3. Bateson M, Dolz J, Kervadec H et al (2021) Constrained domain adaptation for image segmentation. IEEE Trans Med Imaging 40(7):1875–1887

    Article  Google Scholar 

  4. Canny J (1986) A computational approach to edge detection. IEEE Trans Pattern Anal Mach Intell 8(6):679–698

    Article  Google Scholar 

  5. Chen G, Zhang G, Yang Z et al (2023) Multi-scale patch-GAN with edge detection for image inpainting. Appl Intell 53(4):3917–3932

    Article  Google Scholar 

  6. Chen Z, Zhou H, Lai J et al (2020) Contour-aware loss: boundary-aware learning for salient object segmentation. IEEE Trans Image Process 30:431–443

    Article  Google Scholar 

  7. Dlamini S, Kuo CFJ, Chao SM (2023) Developing a surface mount technology defect detection system for mounted devices on printed circuit boards using a MobileNetV2 with feature pyramid network. Eng Appl Artif Intell 121:105,875–105,887

    Article  Google Scholar 

  8. Dollár P, Zitnick CL (2014) Fast edge detection using structured forests. IEEE Trans Pattern Anal Mach Intell 37(8):1558–1570

    Article  Google Scholar 

  9. Dosovitskiy A, Beyer L, Kolesnikov A et al (2021) An image is worth 16x16 words: Transformers for image recognition at scale. In: Proceedings of the international conference on learning representations

  10. Gal R, Patashnik O, Maron H et al (2022) StyleGAN-NADA: CLIP-guided domain adaptation of image generators. ACM Trans Graph 41(4):1–13

    Article  Google Scholar 

  11. Hallman S, Fowlkes CC (2015) Oriented edge forests for boundary detection. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp 1732–1740

  12. He J, Zhang S, Yang M et al (2019) Bi-directional cascade network for perceptual edge detection. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp 3828–3837

  13. Hendrycks D, Gimpel K (2017) A baseline for detecting misclassified and out-of-distribution examples in neural networks. In: Proceedings of the international conference on learning representations

  14. Huan L, Xue N, Zheng X et al (2021) Unmixing convolutional features for crisp edge detection. IEEE Trans Pattern Anal Mach Intell 44(10):6602–6609

    Article  Google Scholar 

  15. Jais IKM, Ismail AR, Nisa SQ (2019) Adam optimization algorithm for wide and deep neural network. Knowl Eng Data Sci 2(1):41–46

    Article  Google Scholar 

  16. Kittler J (1983) On the accuracy of the sobel edge detector. Image Vis Comput 1(1):37–42

    Article  Google Scholar 

  17. LeCun Y, Chopra S, Hadsell R et al (2006) A tutorial on energy-based learning. Predicting Structured Data 1

  18. Li C, Xia W, Yan Y et al (2020) Segmenting objects in day and night: Edge-conditioned cnn for thermal image semantic segmentation. Trans Neural Netw Learn Syst 32(7):3069–3082

    Article  Google Scholar 

  19. Li Z, Liu H, Zhang Z et al (2021) Learning knowledge graph embedding with heterogeneous relation attention networks. IEEE Trans Neural Netw Learn Syst 33(8):3961–3973

    Article  MathSciNet  Google Scholar 

  20. Liang J, Chen G, Wang Y, et al. (2022) EGSANEt: edge-guided sparse attention network for improving license plate detection in the wild. Appl Intell 52(4):4458–4472

    Article  Google Scholar 

  21. Lin C, Zhang Z, Hu Y (2022) Bio-inspired feature enhancement network for edge detection. Appl Intell 52(10):1–16

    Article  Google Scholar 

  22. Liu H, Fang S, Zhang Z et al (2021a) MFDNEt: collaborative poses perception and matrix fisher distribution for head pose estimation. IEEE Trans Multimedia 24:2449–2460

  23. Liu H, Nie H, Zhang Z et al (2021b) Anisotropic angle distribution learning for head pose estimation and attention understanding in human-computer interaction. Neurocomputing 433:310–322

  24. Liu H, Zheng C, Li D et al (2021c) EDMF: efficient Deep matrix factorization with review feature learning for industrial recommender system. IEEE Trans Ind Inf 18(7):4361–4371

  25. Liu H, An Q, Liu T et al (2022a) An infrared image denoising model with unidirectional gradient and sparsity constraint on biomedical images. Infrared Phys Technol 126:104,348–104,354

  26. Liu H, Liu T, Chen Y et al (2022b) EHPE: skeleton cues-based gaussian coordinate encoding for efficient human pose estimation. IEEE Trans Multimedia 1–12

  27. Liu H, Liu T, Zhang Z et al (2022c) ARHPE: asymmetric relation-aware representation learning for head pose estimation in industrial human–computer interaction. IEEE Trans Ind Inf 18(10):7107–7117

  28. Liu H, Zheng C, Li D et al (2022d) Multi-perspective social recommendation method with graph representation learning. Neurocomputing 468:469–481

  29. Liu H, Zhang C, Deng Y et al (2023) TransIFC: invariant cues-aware feature concentration learning for efficient fine-grained bird image classification. IEEE Trans Multimedia 1–14

  30. Liu T, Wang J, Yang B et al (2021d) NGDNEt: nonuniform gaussian-label distribution learning for infrared head pose estimation and on-task behavior understanding in the classroom. Neurocomputing 436:210–220

  31. Liu W, Wang X, Owens J et al (2020) Energy-based out-of-distribution detection. Adv Neural Inf Process Syst 33:21,464–21,475

    Google Scholar 

  32. Liu Y, Cheng MM, Hu X et al (2017) Richer convolutional features for edge detection. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp 3000–3009

  33. Macêdo D, Ren TI, Zanchettin C et al (2021) Entropic out-of-distribution detection: seamless detection of unknown examples. IEEE Trans Neural Netw Learn Syst 33(6):2350–2364

    Article  MathSciNet  Google Scholar 

  34. Martin DR, Fowlkes CC, Malik J (2004) Learning to detect natural image boundaries using local brightness, color, and texture cues. IEEE Trans Pattern Anal Mach Intell 26(5):530–549

    Article  Google Scholar 

  35. Mottaghi R, Chen X, Liu X et al (2014) The role of context for object detection and semantic segmentation in the wild. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp 891–898

  36. Nguyen A, Yosinski J, Clune J (2015) Deep neural networks are easily fooled: high confidence predictions for unrecognizable images. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp 427–436

  37. Pang G, Shen C, Cao L, et al. (2021) Deep learning for anomaly detection: a review. ACM Comput Surv 54(2):1–38

    Article  Google Scholar 

  38. Pu B, Lu Y, Chen J et al (2022a) MobileUNet-FPN: a semantic segmentation model for fetal ultrasound four-chamber segmentation in edge computing environments. IEEE J Biomed Health Inf 26 (11):5540–5550

  39. Pu M, Huang Y, Liu Y et al (2022b) EDTER: edge detection with transformer. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp 1402–1412

  40. Shen W, Wang X, Wang Y et al (2015) DeepContour: a deep convolutional feature learned by positive-sharing loss for contour detection. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp 3982–3991

  41. Silberman N, Hoiem D, Kohli P et al (2012) Indoor segmentation and support inference from rgbd images. In: Proceedings of the European conference on computer vision, pp 746–760

  42. Soria X, Sappa A, Humanante P et al (2023) Dense extreme inception network for edge detection. Pattern Recogn 139:109,461–109,472

    Article  Google Scholar 

  43. Su Z, Liu W, Yu Z et al (2021) Pixel difference networks for efficient edge detection. In: Proceedings of the IEEE international conference on computer vision, pp 5117–5127

  44. Sun Y, Guo C, Li Y (2021) React: out-of-distribution detection with rectified activations. Adv Neural Inf Process Syst 34:144–157

    Google Scholar 

  45. Tan Z, Gu X (2021) Depth scale balance saliency detection with connective feature pyramid and edge guidance. Appl Intell 51(8):5775–5792

    Article  Google Scholar 

  46. Tu Z, Ma Y, Li C et al (2020) Edge-guided non-local fully convolutional network for salient object detection. IEEE Trans Circuits Syst Video Technol 31(2):582–593

    Article  Google Scholar 

  47. Tung F, Mori G (2019) Similarity-preserving knowledge distillation. In: Proceedings of the IEEE international conference on computer vision, pp 1365–1374

  48. Wu W, Zhang S, Zhou K et al (2021) Shadow removal via dual module network and low error shadow dataset. Comput Graph 95:156–163

    Article  Google Scholar 

  49. Wu W, Zhou K, Chen XD et al (2022) Light-weight shadow detection via GCN-based annotation strategy and knowledge distillation. Comput Vis Image Underst 216:103,341–103,352

    Article  Google Scholar 

  50. Xia Y, Yang D, Yu Z et al (2020) Uncertainty-aware multi-view co-training for semi-supervised medical image segmentation and domain adaptation. Med Image Anal 65:101,766–101,776

    Article  Google Scholar 

  51. Xie J, Zheng Z, Gao R et al (2020) Generative VoxelNet: learning energy-based models for 3d shape synthesis and analysis. IEEE Trans Pattern Anal Mach Intell 44(5):2468–2484

    Google Scholar 

  52. Xie S, Tu Z (2015) Holistically-nested edge detection. In: Proceedings of the IEEE international conference on computer vision, pp 1395–1403

  53. Xu W, Fu YL, Xu H et al (2023) Medical image fusion using enhanced cross-visual cortex model based on artificial selection and impulse-coupled neural network. Comput Methods Prog Biomed 229:107,304–107,312

    Article  Google Scholar 

  54. Yang Q, Hou C, Huang M et al (2022a) Instance interactive association graph convolutional network for domain adaptive person re-identification. Appl Intell 52(7):1–14

  55. Yang Z, Soltanian-Zadeh S, Farsiu S (2022b) Biconnet: an edge-preserved connectivity-based approach for salient object detection. Pattern Recog 121:108,231–108,241

  56. Yu L, Li Z, Xu M et al (2022) Distribution-aware margin calibration for semantic segmentation in images. Int J Comput Vis 130(1):95–110

    Article  MATH  Google Scholar 

Download references

Funding

This research work was partially supported by the National Science Foundation of China (61972120), and the National Key R&D Program of China (2019YFB1405703).

Author information

Authors and Affiliations

  1. School of Computer Science and Technology, Hangzhou Dianzi University, Hangzhou, 310018, China

    Wenya Yang, Wen Wu & Xiao-Diao Chen

  2. School of Media and Design, Hangzhou Dianzi University, Hangzhou, 310018, China

    Xiuting Tao

  3. Department of Computer Science and Engineering, University of Yamanashi, Kofu, 400-8511, Yamanashi, Japan

    Xiaoyang Mao

Authors
  1. Wenya Yang
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Wen Wu
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Xiao-Diao Chen
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Xiuting Tao
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Xiaoyang Mao
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

Wenya Yang: Conceptualization, Writing - Original Draft, Writing - Review & Editing. Wen Wu: Validation, Methodology, Project administration, Funding acquisition. Xiao-Diao Chen: Visualization, Formal analysis, Data Curation. Xiuting Tao: Supervision. Xiaoyang Mao: Supervision.

Corresponding authors

Correspondence to Wenya Yang, Xiao-Diao Chen or Xiaoyang Mao.

Ethics declarations

Ethics approval and consent to participate

The work follows appropriate ethical standards in conducting research and writing the manuscript. This work presents computational models trained with publicly available data, for which no ethical approval was required.

Conflict of Interests

The authors declare that they have no conflict of interest.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Yang, W., Wu, W., Chen, XD. et al. How to use extra training data for better edge detection?. Appl Intell 53, 20499–20513 (2023). https://doi.org/10.1007/s10489-023-04587-4

Download citation

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10489-023-04587-4

Keywords

Search

Navigation