[Searching for LAB] Professor Jeon Jae-wook's Team at Sungkyunkwan University Wins 1st Place in International AI Competition

The research team led by Professor Jeon Jae-wook from the Department of Semiconductor Systems Engineering at Sungkyunkwan University announced that they won first place in the 7th ‘AI City Challenge’ competition hosted by CVPR, the world’s top academic conference in the fields of artificial intelligence (AI) and computer vision, held in Vancouver, Canada, last June. Professor Jeon’s team participated in the category of ‘Detection of Motorcycle Riders Violating Helmet Wearing Regulations.’

Professor Jeon Jae-wook of the Department of Semiconductor Systems Engineering at Sungkyunkwan University (center) and his research team, composed of his students, are posing for a photo.

The competition, which took place over six months from January this year, involved 39 teams worldwide. The theme was the development of technology to detect whether individual motorcycle riders were wearing helmets. Professor Jeon’s team developed a two-stage AI technology that first identifies the locations of multiple motorcycles and their riders, then detects whether each rider is wearing a helmet, achieving the highest detection performance score.

This research is expected to be utilized to enhance the efficiency of various types of intelligent traffic surveillance systems in the future. It can be applied comprehensively to various traffic monitoring systems that can be verified through CCTV, such as motorcycle helmets, seatbelt non-wearing, and illegal tuning. Professor Jeon explained, "We have made it possible to detect traffic violations that previously required human visual inspection using only CCTV. While directly viewing inside vehicles raises privacy concerns, violations that are externally visible, such as illegal vehicle tuning, can be monitored with current technology."

Professor Jeon’s research area is ‘road environment recognition using deep learning technology.’ The team consists of Professor Jeon and about ten graduate students and postdoctoral researchers from the Department of Electrical, Electronics, and Computer Engineering participating in the BK21 program. This technology is also the result of a research project titled ‘Real-time Intelligent Traffic Surveillance System on Edge Devices’ under the Software StarLab (SW StarLab) initiative by the Ministry of Science and ICT and the Institute for Information & Communications Technology Planning & Evaluation, aimed at securing world-class foundational software technology and training master’s and doctoral-level talent.

Professor Jeon said, "We have been continuously researching foundational technology and accumulating core technologies. Regarding autonomous driving, we have previously developed deep learning technology to detect people, passenger cars, and motorcycles on roads, and have been continuously developing technology capable of detection even in challenging environments such as darkness or heavy snowfall. I feel proud that our research has been recognized at a world-class academic conference, and I hope this will serve as an opportunity for my students who worked with me to advance globally."