DOI: 10.17587/prin.15.589-599

Technology of Artificial Data Generation for Neural Network Training (using the Example of Quality Control of Laminate Production)

A. V. Zaripov, Postgraduate Student, aleksei.v.zaripov@tusur.ru, R. S. Kulshin, Graduate Student, roman.s.kulshin@tusur.ru, A. A. Sidorov, Cand. Sci. (Econ.), Head of Department, anatolii.a.sidorov@tusur.ru, Tomsk State University of Control Systems and Radioelectronics, Tomsk, 634050, Russian Federation

Corresponding author: Anatoly A. Sidorov, Cand. Sci. (Econ.), Head of Department, Tomsk State University of Control Systems and Radioelectronics, 634050, Tomsk, Russian Federation, E-mail: anatolii.a.sidorov@tusur.ru

The paper presents the results of a study on artificial data generation methods for computer vision tasks in the operation of conveyor-type technological lines. It identifies the problematic issues of high cost and low efficiency of traditional data collection methods such as video recording and manual markup. A generalized model of synthetic data generation using game engines and 3D modeling is proposed as a solution to these problems. The model was experimentally tested in the context of the laminate coating determination process. Synthetic data allowed the YOLOv8 neural network to be trained with high accuracy, achieving an mAP50 of 0.95. The results indicate that synthetic data can be used to improve the quality of models and optimize learning processes for neural networks in situations where real datasets are limited.

Keywords: data generation, neural network, synthetic data, computer vision, Yolo, Unity, conveyor, laminate, defect

pp. 589—599

For citation:
Zaripov A. V., Kulshin R. S., Sidorov A. A. Technology of Artificial Data Generation for Neural Network Training (using the Example of Quality Control of Laminate Production), Programmnaya Ingeneria, 2024, vol. 15, no. 11, pp. 589—599. DOI: 10.17587/prin.15.589-599. (in Russian).

The work was carried out within the framework of the state task "Science"; project FEWM-2023-0013.

References:

1. Greeshma C. A., Nidhindas K. R., Sreejith P. Traffic control using computer vision, International Journal of Advanced Research in Computer and Communication Engineering, 2019, vol. 8, no. 4, pp. 39—47.
2. Rusanovsky M., Beeri O., Oren G. An end-to-end computer vision methodology for quantitative metallography, Scientific Reports, 2022, vol. 12, no. 1, article 4776. DOI: 10.1038/s41598-022-08651-w.
3. Aydin I., Othman N. A. A new IoT combined face detection of people by using computer vision for security application, 2017 International Artificial Intelligence and Data Processing Symposium IDAP, 2017, pp. 1—6. DOI: 10.1109/IDAP.2017.8090171.
4. Costa C., Antonucci F., Pallottino F. et al. Shape analysis of agricultural products: a review of recent research advances and potential application to computer vision, Food and Bioprocess Technology, 2011, vol. 4, no. 5, pp. 673—692. DOI: 10.1007/s11947-011-0556-0.
5. Video analytics in production. How does artificial intelligence work where humans can't, available at: https://habr.com/ru/companies/evraz/articles/581882/ (date of access 31.04.2024) (in Russian).
6. Esteva A., Chou K., Yeung S. et al. Deep learning-enabled medical computer vision, NPJ digital medicine, 2021, vol. 4, no. 1, article 5. DOI: 10.1038/s41746-020-00376-2.
7. Villalba-Diez J., Schmidt D. et al. Deep learning for industrial computer vision quality control in the printing industry 4.0, Sensors, 2019, vol. 19, no. 18, article 3987. DOI: 10.3390/s19183987.
8. Kiefer B., Ott D., Zell A. Leveraging synthetic data in object detection on unmanned aerial vehicles, 2022 26th international conference on pattern recognition (ICPR), 2022, pp. 3564—3571. DOI: 10.1109/ICPR56361.2022.9956710.
9. Chumak R. A synthetic data generator. Training of neural networks for industrial flaw detection, available at: https://medium.com/phygitalism/synthetic-data-generator-a052d347468 (date of access 5.05.2024) (in Russian).
10. The Perception Camera Component, available at: https://docs.unity3d.com/Packages/com.unity.perception@1.0/manual/Per-ceptionCamera.html. (date of access 20.05.24).
11. Reutov I., Moskvin D., Voronova A., Venediktov M. Generating Synthetic Data To Solve Industrial Control Problems By Mod­eling A Belt Conveyor, Procedia Computer Science, 2022, vol. 212, pp. 264—274. DOI: 10.1016/j.procs.2022.11.010.
12. Pchelintsev S., Yulyashkov M. A., Kovaleva O. A. A method for creating synthetic datasets for training neural network models to recognize objects, Information management systems, 2022, no. 3, pp. 9—19. DOI: 10.31799/1684-8853-2022-3-9-19 (in Russian).
13. Magnus D., Predrag B., Hannie P. C++ Application of the fractal Perlin noise algorithm for the generation of simulated breast tissue, Medical Imaging 2015: Physics of Medical Imaging, 2015, vol. 9412, pp. 844—852.
14. Bazuhair W., Lee W. Detecting malign encrypted network traffic using perlin noise and convolutional neural network, 2020 10th Annual Computing and Communication Workshop and Con­ference (CCWC), 2020, pp. 0200—0206.
15. Li H., Tuo X., Liu Y., Jiang X. A parallel algorithm using Perlin noise superposition method for terrain generation based on CUDA architecture, International Conference on Materials Engineering and Information Technology Applications (MEITA 2015), 2015, pp. 967—974.
16. Ying X. An overview of overfitting and its solutions, Journal of physics: Conference series, 2019, vol. 1168, no. 2. DOI: 10.1088/1742­6596/1168/2/022022.
17. About Cinemachine, available at: https://docs.unity3d.com/Packages/com.unity.cinemachine@2.8/manual/index.html (date of access 26.05.24).
18. Particle system, available at: https://docs.unity3d.com/ru/530/Manual/ParticleSystems.html (date of access 26.05.2024).
19. Sohan M., Sai Ram T., Reddy R., Venkata C. A review on yolov8 and its advancements, International Conference on Data Intelligence and Cognitive Informatics, 2024, pp. 529—545. DOI: 10.1007/978-981-99-7962-2_39.