CorePal

Automotive parts remanufacturing factories adjust their production models based on changes in the used parts market. For example, in the case of transmission boxes, the market typically sees no more than 500 types of common used parts. However, each year, the 50 oldest types are phased out, and 50 new types are introduced. If each part is photographed from various angles and backgrounds, approximately 1,000 images are needed for training per type of transmission box. To facilitate continuous learning for traditional object classification AI, about 50,000 training images are prepared annually. Through the use of generative AI, clients only need to perform handheld 3D scanning and modeling for 50 types of parts each year. Multimodal AI can directly complete the training. The model identification AI can aid remanufacturing factories in identifying the raw materials on-site, assisting in material management while also reducing communication costs in procurement from retailers.