Textile defect detection using YOLOv5 on AITEX Dataset

  • Rodolfo Seidel IFES
  • Hilário Seibel Júnior IFES
  • Karin Satie Komati IFES

Resumo


Devido à identificação manual de defeitos têxteis ainda nos dias atuais, é necessário encontrar meios de detectar defeitos de forma automatizada e eficiente. Para isso, este trabalho se propõe a aplicar o modelo YOLOv5 na base de dados AITEX, usando a abordagem de detecção de objetos para localizar e identificar defeitos, avaliando diferentes técnicas de anotação de objetos e data augmentation. Com os resultados obtidos, concluiu-se que o YOLOv5 adaptou-se muito bem a outro contexto com objetos distintos do prétreinamento, as anotações com Bounding Boxes permitiram maior aprendizado e reconhecimento dos defeitos, mesmo com diferentes formas e tamanhos, e por fim, a combinação de data augmentation potencializam seu desempenho.

Referências

Arlen, T. C. (2018). Understanding the map evaluation metric for object detection. [link].

Bochkovskiy, A., Wang, C.-Y., and Liao, H.-Y. M. (2020). Yolov4: Optimal speed and accuracy of object detection. arXiv preprint arXiv:2004.10934.

Conceição, A. M. G. (1998). Critérios de classificação de tecidos quanto à qualidade. [link].

Gevorgyan, Z. (2022). Siou loss: More powerful learning for bounding box regression. arXiv preprint arXiv:2205.12740.

Jin, R. and Niu, Q. (2021). Automatic fabric defect detection based on an improved yolov5. Mathematical Problems in Engineering, 2021.

Jing, J., Wang, Z., Rätsch, M., and Zhang, H. (2022). Mobile-unet: An efficient convolutional neural network for fabric defect detection. Textile Research Journal, 92(12):30-42.

Jing, J., Zhuo, D., Zhang, H., Liang, Y., and Zheng, M. (2020). Fabric defect detection using the improved yolov3 model. Journal of engineered fibers and fabrics, 15:1558925020908268.

Jocher, G., Nishimura, K., Mineeva, T., and Vilariño, R. (2020). Yolov5 (2020). GitHub repository: https://github.com/ultralytics/yolov5.

Jun, X., Wang, J., Zhou, J., Meng, S., Pan, R., and Gao, W. (2021). Fabric defect detection based on a deep convolutional neural network using a two-stage strategy. Textile Research Journal, 91(1-2):130-142.

Lin, T.-Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., Dollár, P., and Zitnick, C. L. (2014). Microsoft coco: Common objects in context. In European conference on computer vision, pages 740-755. Springer.

Ouyang, W., Xu, B., Hou, J., and Yuan, X. (2019). Fabric defect detection using activation layer embedded convolutional neural network. IEEE Access, 7:70130-70140.

Petronas, I. I. (2020). Indústria têxtil 4.0: quais são as novidades para este setor? https://inovacaoindustrial.com.br/industria-textil-40/.

Popkova, E. G., Ragulina, Y. V., and Bogoviz, A. V. (2019). Fundamental differences of transition to industry 4.0 from previous industrial revolutions. In Industry 4.0: Industrial Revolution of the 21st Century, pages 21-29. Springer.

Redmon, J., Divvala, S., Girshick, R., and Farhadi, A. (2015). You look only once: unified real-time object detection. arXiv preprint arXiv:1506.02640.

Rezatofighi, H., Tsoi, N., Gwak, J., Sadeghian, A., Reid, I., and Savarese, S. (2019). Generalized intersection over union: A metric and a loss for bounding box regression. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 658-666.

Rong-qiang, L., Ming-hui, L., Jia-chen, S., and Yi-bin, L. (2021). Fabric defect detection method based on improved u-net. In Journal of Physics: Conference Series, volume 1948, page 012160. IOP Publishing.

Sakkos, D., Shum, H. P., and Ho, E. S. (2019). Illumination-based data augmentation for robust background subtraction. In 2019 13th International Conference on Software, Knowledge, Information Management and Applications (SKIMA), pages 1-8. IEEE.

Seçkin, A. Ç. and Seçkin, M. (2022). Detection of fabric defects with intertwined frame vector feature extraction. Alexandria Engineering Journal, 61(4):2887-2898.

Silvestre-Blanes, J., Albero Albero, T., Miralles, I., Pérez-Llorens, R., and Moreno, J. (2019). A public fabric database for defect detection methods and results. Autex Research Journal, 19(4):363-374.

Wang, Y., Hao, Z., Zuo, F., and Pan, S. (2021). A fabric defect detection system based improved yolov5 detector. In Journal of Physics: Conference Series, volume 2010, page 012191. IOP Publishing.

Wieler, M., Hahn, T., and Hamprecht, F. A. (2007). Weakly supervised learning for industrial optical inspection. [dataset]. [link].

Zheng, L., Wang, X., Wang, Q., Wang, S., and Liu, X. (2021). A fabric defect detection method based on improved yolov5. In 2021 7th International Conference on Computer and Communications (ICCC), pages 620-624. IEEE.

Zheng, Z., Wang, P., Liu, W., Li, J., Ye, R., and Ren, D. (2020). Distance-iou loss: Faster and better learning for bounding box regression. In Proceedings of the AAAI conference on artificial intelligence, volume 34, pages 12993-13000.

Zhou, Q., Mei, J., Zhang, Q., Wang, S., and Chen, G. (2021). Semi-supervised fabric defect detection based on image reconstruction and density estimation. Textile Research Journal, 91(9-10):962-972.
Publicado
28/11/2022
SEIDEL, Rodolfo; SEIBEL JÚNIOR, Hilário; KOMATI, Karin Satie. Textile defect detection using YOLOv5 on AITEX Dataset. In: ENCONTRO NACIONAL DE INTELIGÊNCIA ARTIFICIAL E COMPUTACIONAL (ENIAC), 19. , 2022, Campinas/SP. Anais [...]. Porto Alegre: Sociedade Brasileira de Computação, 2022 . p. 763-774. ISSN 2763-9061. DOI: https://doi.org/10.5753/eniac.2022.227396.