Press Release

ComputerMate (CEO Sang-in Seo and Seong-ho Kim), an AI technology company, announced on the 11th that it has developed a 'deep learning-based universal vision inspection platform' that has dramatically improved inspection accuracy in automobile parts manufacturing sites.

ComputerMate applied a deep learning-based vision inspection solution to the cowl cross inspection process of Sewon Precision, an auto parts manufacturer in Daegu. In the existing process, normal products were often judged as defective, which resulted in inefficiency as workers had to visually re-inspect them. In addition, there were limitations in that accurate inspection was difficult with the existing vision system due to various inspection conditions such as the location and shape of the parts. To solve this problem, Sewon Precision participated in the 'Demand-tailored AI Solution Development and Verification Support Project' promoted by the Daegu Digital Innovation Promotion Agency.

The 'Demand-tailored AI solution development and verification support project' is part of the 'Manufacturing AI Convergence Base Creation Project', which is managed by the National IT Industry Promotion Agency with the support of the Ministry of Science and ICT and the Daegu Metropolitan City. This project aims to solve various problems occurring in intelligent machine parts manufacturing sites in the Daegu area with AI technology, thereby increasing production capacity and efficiency and strengthening the safety of the work environment.

Computermate's platform is designed based on data set construction, image processing, and transfer learning technology using generative AI. In particular, it focuses on detecting non-standard errors and minimizing false positives through deep learning technology. Through this, it has overcome the limitations of existing systems and has universal applicability to various parts and production lines.

The platform continuously improves AI accuracy through periodic data collection and model retraining, and supports rapid decision-making by systematizing inspection process data and quantifying quality data.

Through the introduction of this solution, Sewon Precision aims to ‘zero out’ defects in the inspection process and improve productivity by more than 5%.

Kim Yong-hwan, head of ComputerMate Research Lab, said, “I am delighted that we can simultaneously improve the quality and efficiency of the manufacturing industry through a deep learning-based vision inspection universal platform,” adding, “It will greatly contribute to solving the problem of poor cost-effectiveness and improving productivity.”

Founded in 1992, ComputerMate is a smart factory solution specialist that has been developing products optimized for manufacturing sites for over 30 years. Its flagship product, the Mate series, provides a variety of solutions tailored to corporate needs, including a manufacturing execution system (MES), enterprise resource planning (ERP), supply chain management (SCM), warehouse management system (WMS), and quality management system (QMS).

[Source] Maeil Kyungjae, Reporter Jung Mooyong jung, mooyong@mkinternet.com