SpringerLink
Log in

Presswork defect inspection using only defect-free high-resolution images

  • Original article
  • Published:
The Visual Computer Aims and scope Submit manuscript

Abstract

Efficient automated presswork defect detection is valuable to the printing industry since such defects dramatically depress the presswork grad and manually detecting them is cost-ineffective. Instead of using real defect annotating data, the goal of this paper is to just use defect-free high-resolution images to complete the detection of defect in presswork (SWDF). In this paper, we first propose an end-to-end crop** method to balance the number of samples which cropped from different patterns in the defect-free high-resolution images, and thus, we can solve the quantity imbalance between different patterns (image level). Secondly, we designed an simple but effective loss function to solve the severe imbalance which caused by the larger quantity difference between the defective pixels and background pixels (pixel level). In order to solve the absence of the real defect annotating data and replace the time-consuming acquisition of defect data, we designed a high-speed defect generation algorithm, which directly generate defect on defect-free samples (obtained from the crop** process). As for the detection of defects, we use self-attention to design a novel and effective semantic segmentation head (GST), which can exploit global information from the feature map to repair the detection results, so as to obtain a better performance. In the experimental part, we will use a presswork defect dataset from the Zhentu Cup Competition (ZCC) and a public dataset DAGM 2007 to test the effect of our scheme. Especially, compared with existing supervised models, our model has also reached the state of the art in DAGM 2007.

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

Access this article

Log in via an institution

Subscribe and save

Springer+ Basic
EUR 32.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or Ebook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Price includes VAT (Germany)

Instant access to the full article PDF.

Institutional subscriptions

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

Similar content being viewed by others

Notes

  1. Esca** rate is the ability of a model to identify only the relevant objects. It is the percentage of relevant ground truths not detected among all samples and is given by:Esca** rate = FN/all.

  2. Overkill is the ability of a model to identify only the relevant objects. It is the percentage of wrong positive predictions among all samples and is given by:Overkilling rate = FP/all.

  3. Intersection Over Union (IOU) is a measure based on Jaccard index that evaluates the overlap between two regions.

References

  1. Božič, J., Tabernik, D., Skočaj, D.: End-to-end training of a two-stage neural network for defect detection. ar**v preprint ar**v:2007.07676 (2020)

  2. Božič, J., Tabernik, D., Skočaj, D.: Mixed supervision for surface-defect detection: from weakly to fully supervised learning. Comput. Ind. 129, 103459 (2021)

    Article  Google Scholar 

  3. Chen, L.C., Zhu, Y., Papandreou, G., Schroff, F., Adam, H.: Encoder-decoder with atrous separable convolution for semantic image segmentation. In: Proceedings of the European conference on computer vision (ECCV), pp. 801–818 (2018)

  4. Dai, W., Erdt, M., Sourin, A.: Detection and segmentation of image anomalies based on unsupervised defect reparation. The Visual Computer pp. 1–10 (2021)

  5. Kim, S., Kim, W., Noh, Y.K., Park, F.C.: Transfer learning for automated optical inspection. In: 2017 International Joint Conference on Neural Networks (IJCNN), pp. 2517–2524. IEEE (2017)

  6. Kingma, D.P., Welling, M.: Auto-encoding variational bayes. ar**v preprint ar**v:1312.6114 (2013)

  7. Li, B., Liu, Y., Wang, X.: Gradient harmonized single-stage detector. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 33, pp. 8577–8584 (2019)

  8. Li, X., Huang, H., Zhao, H., Wang, Y., Hu, M.: Learning a convolutional neural network for propagation-based stereo image segmentation. Vis. Comput. 36(1), 39–52 (2020)

    Article  Google Scholar 

  9. Lin, T.Y., Goyal, P., Girshick, R., He, K., Dollár, P.: Focal loss for dense object detection. In: Proceedings of the IEEE international conference on computer vision, pp. 2980–2988 (2017)

  10. Mei, S., Wang, Y., Wen, G.: Automatic fabric defect detection with a multi-scale convolutional denoising autoencoder network model. Sensors 18(4), 1064 (2018)

    Article  Google Scholar 

  11. Mei, S., Yang, H., Yin, Z.: An unsupervised-learning-based approach for automated defect inspection on textured surfaces. IEEE Trans. Instrum. Meas. 67(6), 1266–1277 (2018)

    Article  Google Scholar 

  12. Qin, Y., Mo, L., Li, C., Luo, J.: Skeleton-based action recognition by part-aware graph convolutional networks. Vis. Comput. 36(3), 621–631 (2020)

    Article  Google Scholar 

  13. Racki, D., Tomazevic, D., Skocaj, D.: A compact convolutional neural network for textured surface anomaly detection. In: 2018 IEEE Winter Conference on Applications of Computer Vision (WACV), pp. 1331–1339. IEEE (2018)

  14. Shrivastava, A., Gupta, A., Girshick, R.: Training region-based object detectors with online hard example mining. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 761–769 (2016)

  15. Shyam, P., Yoon, K.J., Kim, K.S.: Towards domain invariant single image dehazing. ar**v preprint ar**v:2101.10449 (2021)

  16. Timm, F., Barth, E.: Non-parametric texture defect detection using weibull features. In: Image Processing: Machine Vision Applications IV, vol. 7877, p. 78770J. International Society for Optics and Photonics (2011)

  17. Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A.N., Kaiser, L., Polosukhin, I.: Attention is all you need. ar**v preprint ar**v:1706.03762 (2017)

  18. Wan, J., Mougeot, G., Yang, X.: Dense feature pyramid network for cartoon dog parsing. Vis. Comput. 36(10), 2471–2483 (2020)

    Article  Google Scholar 

  19. Wang, X., Girshick, R., Gupta, A., He, K.: Non-local neural networks. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 7794–7803 (2018)

  20. Weimer, D., Scholz-Reiter, B., Shpitalni, M.: Design of deep convolutional neural network architectures for automated feature extraction in industrial inspection. CIRP Ann. 65(1), 417–420 (2016)

    Article  Google Scholar 

  21. Weimer, D., Thamer, H., Scholz-Reiter, B.: Learning defect classifiers for textured surfaces using neural networks and statistical feature representations. Proc. CIRP 7, 347–352 (2013)

    Article  Google Scholar 

  22. Yang, F., Yang, H., Fu, J., Lu, H., Guo, B.: Learning texture transformer network for image super-resolution. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 5791–5800 (2020)

  23. Yuan, Y., Chen, X., Wang, J.: Object-contextual representations for semantic segmentation. ar**v preprint ar**v:1909.11065 (2019)

  24. Zhao, H., Shi, J., Qi, X., Wang, X., Jia, J.: Pyramid scene parsing network. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 2881–2890 (2017)

  25. Zhao, Z., Li, B., Dong, R., Zhao, P.: A surface defect detection method based on positive samples. In: Pacific Rim International Conference on Artificial Intelligence, pp. 473–481. Springer (2018)

Download references

Funding

No funding was received.

Author information

Authors and Affiliations

  1. Jiangsu Key Laboratory of Big Data Analysis Technology (B-DAT), Nan**g University of Information Science & Technology, Nan**g, China

    Zhenyu Guan, Ziqi Wang, Yisheng Zhu & Guangcan Liu

Authors
  1. Zhenyu Guan
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Ziqi Wang
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Yisheng Zhu
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Guangcan Liu
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Zhenyu Guan.

Ethics declarations

Conflict of interest

Zhenyu Guan declares that he has no conflict of interest. Ziqi Wang declares that he has no conflict of interest. Yisheng Zhu declares that he has no conflict of interest. Guangcan Liu declares that he has 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

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Guan, Z., Wang, Z., Zhu, Y. et al. Presswork defect inspection using only defect-free high-resolution images. Vis Comput 39, 1271–1282 (2023). https://doi.org/10.1007/s00371-022-02403-7

Download citation

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00371-022-02403-7

Keywords

Search

Navigation