A novel methodology for dataset augmentation in the semantic segmentation of coil-coated surface degradation is presented in this study. Deep convolutional generative adversarial networks (DCGAN) are employed to generate synthetic input-target pairs, which closely resemble real-world data, with the goal of expanding an existing dataset. These augmented datasets are used to train two state-of-the-art models, U-net, and DeepLabV3, for the precise detection of degradation areas around scribes. In a series of experiments, it was demonstrated that the introduction of synthetic data improves the models' performance in detecting degradation, especially when the ratio of synthetic to real data is carefully managed. Results indicate that optimal improvements in accuracy and F1-score are achieved when the ratio of synthetic to original data is between 0.2 and 0.5. Moreover, the advantages and limitations of different GAN architectures for dataset expansion are explored, with specific attention to their ability to produce realistic and diverse samples. This work offers a scalable solution to the challenges associated with creating large and diverse annotated datasets for industrial applications of coil coating degradation assessment. The proposed approach provides a significant contribution by improving model generalization and segmentation accuracy while reducing the burden of manual data annotation. These findings have important implications for industries relying on coil coatings, as more efficient and accurate degradation detection methods are enabled.

Faculty of Electrical Engineering and Informatics University of Pardubice Pardubice Czechia

Zobrazit více v PubMed

Allah A. M. G., Sarhan A. M., Elshennawy N. M. (2023). Edge U-Net: brain tumor segmentation using MRI based on deep U-Net model with boundary information. Expert Syst. Applic. 213:118833. 10.1016/j.eswa.2022.118833 DOI

Allen-Zhu Z., Li Y. (2022). “Feature purification: how adversarial training performs robust deep learning,” in 2021 IEEE 62nd Annual Symposium on Foundations of Computer Science (FOCS) (IEEE: ), 977–988. 10.1109/FOCS52979.2021.00098 DOI

Bastos A., Simões A. (2009). Effect of deep drawing on the performance of coil-coatings assessed by electrochemical techniques. Progr. Organ. Coatings 65, 295–303. 10.1016/j.porgcoat.2009.01.002 DOI

Blanchard P. J., Hill D. J., Bretz G. T., McCune R. C. (2014). “Evaluation of corrosion protection methods for magnesium alloys in automotive applications,” in Essential Readings in Magnesium Technology, eds. S. N. Mathaudhu, A. A. Luo, N. R. Neelameggham, E. A. Nyberg, W. H. Sillekens (Cham: Springer; ).

Camargo A. M., Olveres J., Escalante-Ramírez B. (2024). “Neural style transfer in tiny sets of ultrasound images for data augmentation,” in Optics, Photonics, and Digital Technologies for Imaging Applications VIII (SPIE: ), 23–31. 10.1117/12.3017658 DOI

Chen L., Papandreou G., Schroff F., Adam H. (2017). Rethinking atrous convolution for semantic image segmentation. arXiv preprint arXiv:1706.05587.

Chollet F. (2017). “Xception: deep learning with depthwise separable convolutions,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 1251–1258). 10.1109/CVPR.2017.195 DOI

Cringasu E. C., Dragomirescu A., Safta C. A. (2017). “Image processing approach for estimating the degree of surface degradation by corrosion,” in 2017 International Conference on ENERGY and ENVIRONMENT (CIEM), 275–278. 10.1109/CIEM.2017.8120791 DOI

Dolezel P., Rozsivalova V., Pakosta M., Stursa D. (2023). “Automated dataset enhancement using GAN for assessment of degree of degradation around scribe,” in 2023 9th International Conference on Control, Decision and Information Technologies (CoDIT) (IEEE: ), 1454–1458. 10.1109/CoDIT58514.2023.10284338 DOI

Goodfellow I., Pouget-Abadie J., Mirza M., Xu B., Warde-Farley D., Ozair S., et al. . (2014). “Generative adversarial nets,” in Advances in Neural Information Processing Systems, 27.

Hanus J. (2011). Selection and evaluation of singlelayer coating compositions in corrosive environments. Acta Univer. Agric. Silvicult Mendelianae Brunensis 59, 53–64. 10.11118/actaun201159050053 DOI

Islam T., Hafiz M. S., Jim J. R., Kabir M. M., Mridha M. (2024). A systematic review of deep learning data augmentation in medical imaging: Recent advances and future research directions. Healthcare Analytics 5:100340. 10.1016/j.health.2024.100340 DOI

Jandel A.-S. (2019). Sustainability in the coil coatings sector [nachhaltigkeit in der coil-coating-branche]. JOT, J. Oberflaechentechnik 59, 10–13. 10.1007/s35144-019-0274-3 DOI

Kang D., Lai J., Han Y. (2024). Accurate detection of surface defects by decomposing unreliable tasks under boundary guidance. Expert Syst. Applic. 244:122977. 10.1016/j.eswa.2023.122977 DOI

Kapsalas P., Zervakis M., Maravelaki-Kalaitzaki P. (2007). Evaluation of image segmentation approaches for non-destructive detection and quantification of corrosion damage on stonework. Corros. Sci. 49, 4415–4442. 10.1016/j.corsci.2007.03.049 DOI

Karras T., Laine S., Aila T. (2021). A style-based generator architecture for generative adversarial networks. IEEE Trans. Pattern Anal. Mach. Intell. 43, 4217–4228. 10.1109/TPAMI.2020.2970919 PubMed DOI

Linder-Norén E. (2019). Keras-GAN. GitHub. Available at: https://github.com/eriklindernoren/Keras-GAN/tree/master (accessed August 13, 2024).

Mirza M., Osindero S. (2014). Conditional generative adversarial nets. arXiv preprint arXiv:1411.1784.

Nanni L., Paci M., Brahnam S., Lumini A. (2021). Comparison of different image data augmentation approaches. J. Imaging 7:254. 10.3390/jimaging7120254 PubMed DOI PMC

Niu S., Peng Y., Li B., Wang X. (2023). A transformed-feature-space data augmentation method for defect segmentation. Comput. Ind. 147:103860. 10.1016/j.compind.2023.103860 DOI

Radford A., Metz L., Chintala S. (2016). Unsupervised representation learning with deep convolutional generative adversarial networks. arXiv preprint arXiv:1511.06434.

Ronneberger O., Fischer P., Brox T. (2015). “U-net: convolutional networks for biomedical image segmentation,” in Medical image computing and computer-assisted intervention MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceedings, part III 18 (Springer International Publishing: ), 234–241. 10.1007/978-3-319-24574-4_28 DOI

Rozsivalova V., Dolezel P., Stursa D., Rozsival P. (2022). Sequence of u-shaped convolutional networks for assessment of degree of delamination around scribe. Int. J. Comput. Intell. Syst. 15:76. 10.1007/s44196-022-00141-1 DOI

Sandler M., Howard A., Zhu M., Zhmoginov A., Chen L. C. (2018). “Mobilenetv2: Inverted residuals and linear bottlenecks,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 4510–4520. 10.1109/CVPR.2018.00474 PubMed DOI

Shorten C., Khoshgoftaar T. M. (2019). A survey on image data augmentation for deep learning. J. Big Data 6, 1–48. 10.1186/s40537-019-0197-0 PubMed DOI PMC

Tao X., Gao H., Yang K., Wu Q. (2024). Expanding the defect image dataset of composite material coating with enhanced image-to-image translation. Eng. Appl. Artif. Intell. 133:108590. 10.1016/j.engappai.2024.108590 DOI

Tekchandani H., Verma S., Londhe N. (2020). Performance improvement of mediastinal lymph node severity detection using gan and inception network. Comput. Methods Programs Biomed. 194:105478. 10.1016/j.cmpb.2020.105478 PubMed DOI

The National Coil Coating Association (2020). NCCA promotes growth of coil coating industry. Light Metal Age 78, 66–67. Available at: https://store.lightmetalage.com/index.php?_a=product&product_id=1271

Wang Q., Zhou X., Wang C., Liu Z., Huang J., Zhou Y., et al. . (2019). Wgan-based synthetic minority over-sampling technique: improving semantic fine-grained classification for lung nodules in CT images. IEEE Access 7, 18450–18463. 10.1109/ACCESS.2019.2896409 DOI

Xu M., Yoon S., Fuentes A., Park D. S. (2023). A comprehensive survey of image augmentation techniques for deep learning. Pattern Recognit. 137:109347. 10.1016/j.patcog.2023.109347 DOI

Yuan L.-X., Xu H.-M., Zhang Z.-Y., Liu X.-W., Li J.-X., Wang J.-H., et al. . (2023). High precision tracking analysis of cell position and motion fields using 3d u-net network models. Comput. Biol. Med. 154:106577. 10.1016/j.compbiomed.2023.106577 PubMed DOI

Zhu J. Y., Park T., Isola P., Efros A. A. (2017). “Unpaired image-to-image translation using cycle-consistent adversarial networks,” in Proceedings of the IEEE International Conference on Computer Vision, 2223–2232. 10.1109/ICCV.2017.244 DOI