Abstract. The purpose of the study was to investigate the impact of using deep neural networks on improving the efficiency of automated quality control in contemporary production. The analysis of the application of advanced technologies, such as computer vision, machine learning, and deep neural networks, to automate quality control processes in production conditions was carried out. Successful implementations of automated quality control systems at enterprises such as Bayerische Motoren Werke AG, Siemens, and Nikon were considered as practical examples. The data obtained confirmed that the use of convolutional neural networks for image and video processing, autoencoders, and generative conflicting networks provides high accuracy and speed in detecting defects. In particular, an increase in the accuracy of defect identification was recorded from 80% to 95%, completeness – from 85% to 92%, specificity – from 90% to 98%. The speed of image and video processing increased fivefold – from 5 minutes to 1 minute per unit of production, which significantly reduced the control cycle time. Improving the defect detection rate to 95% helped to reduce the cost of manual verification and minimise the impact of the human factor. The conducted comparative analysis confirmed that automated systems based on deep neural networks significantly outperform conventional methods of monitoring key performance metrics. The study also showed that the integration of such systems with data processing peripherals and cloud platforms provides high flexibility and scalability of production processes. The economic assessment showed a significant reduction in labour costs and the number of errors, and an increase in the overall productivity of enterprises using automated quality control systems. The results obtained can be used as practical recommendations for companies interested in implementing innovative approaches to product quality assurance
Vitalii Yasenenko (Fri,) studied this question.