The National University of Korea's automotive artificial intelligence club, ‘KaAI (KAI),’ achieved the remarkable feat of securing 3rd place at the DriveX Grand Challenge held during CVPR 2025 (Computer Vision and Pattern Recognition), one of the world's most prestigious computer vision conferences. CVPR is one of the most influential international academic conferences in the fields of image recognition and artificial intelligence, where leading scholars and industry experts from around the world gather annually to showcase the latest technologies and research findings.

The solution submitted by the undergraduate team from KaAI—consisting of Kang Sang-gu, Park Gyu-min, Seo Eun, Yang Jung-woo, and Jeong Seong-hyun—is a Mixture Query-based Transformer model for V2X-based 3D object detection and tracking. It achieved high accuracy in object detection and tracking performance even in complex intersection environments through multi-vehicle cooperative recognition. In particular, it was evaluated as having achieved outstanding results in various evaluation metrics through precise preprocessing, model optimization, and improved real-time processing performance.

This competition is an official side event of CVPR 2025 and is a global challenge on the convergence of autonomous driving and V2X (vehicle-to-vehicle communication) technologies, with fierce competition among leading universities and research institutions around the world. The KaAI team was recognized internationally for its technical perfection and creative approach, achieving a place in the top three.

Professor Lee Sang-heon, the team's advisor, said, “It is a rare achievement for an undergraduate team to rank third among the world's leading AI research teams. This is proof of Kookmin University's educational capabilities and the potential of its students on the international stage.”

KaAI is a research-oriented club composed primarily of undergraduate students from the Automotive Fusion College at Kookmin University, dedicated to conducting research and practical projects in the fields of autonomous driving and artificial intelligence. The team has achieved outstanding results in international competitions and academic events, including winning first place in the IEEE Intelligent Transportation Systems Society (ITSS) AI Competition and being selected for an oral presentation at the IEEE IV 2025 conference, among other accomplishments.