Elice Group and Hyundai Motor Namyang R&D Center Develop Manufacturing-Specific AI Images

Elice Group, a company specializing in AI solutions, has partnered with Hyundai Motor Company's Namyang R&D Center to unveil an innovative solution that addresses the long-standing problem of image management in manufacturing sites. Through approximately 8 months of close collaboration, they successfully developed a 'manufacturing site-optimized image AI solution' that dramatically reduces the manpower and time wasted on classifying and searching vast amounts of randomly stored test photos. This project was led by Hyundai Motor Company's Namyang R&D Center and was conducted from late March to late November. Elice Group developed advanced deep learning-based image classification and search models and built an integrated desktop program for easy utilization. Particularly prioritizing sensitive data security, model training was conducted in a private cloud environment with Elice's modular security system applied, further enhancing reliability. The solution applies a customized AI model, re-trained based on crash test data from accredited institutions, perfectly optimized for the unique test environment of Hyundai Motor Company's Namyang R&D Center. Remarkably, it demonstrated the ability to classify over 60 types of crash test images with an astonishing accuracy of over 98%. Furthermore, it is noteworthy that the model's completeness was enhanced through an intelligent method of incorporating additional public data for training, even in situations where internal data was insufficient. An advanced image search function, going beyond simple classification, is also a core strength of this solution. When a user selects a specific damaged area or component in a photo, the AI presents the most similar images in order based on that selection. This enables the exploration of highly specific test cases, which was difficult with existing text-based or whole-image search methods, thereby maximizing research efficiency. By integrating the entire process, from image indexing and automatic classification to searching, into a single program, Elice Group has completely automated the workflow from the moment test photos are stored to when they are utilized. This is expected to significantly contribute to accelerating the digital transformation of manufacturing sites. Kim Jae-won, CEO of Elice Group, stated, "We received very positive feedback from Hyundai Motor Company's Namyang R&D Center regarding the image similarity search function and the intuitive user interface (UI)." He further expressed his ambition to "leverage this successful collaboration as a stepping stone to continuously expand AI solutions that will revolutionize work efficiency across various industrial sectors, including manufacturing and mobility."