In July, CVPR 2023 Art of Robustness: Vehicle Detection in the Physical-World Challenge concluded with a global championship victory for the 'Huge' team representing the Huazhong University of Science and Technology (HUST). The triumphant team consisted of Associate Researcher Hu Shengshan and graduate student Wang Haoyu from the School of Cyber Science and Engineering, in addition to Lecturer Li Minghui from the School of Software Engineering.

With a principal mission of promoting academic exchange and advancing innovative research in the field of computer vision and pattern recognition, CVPR (Conference on Computer Vision and Pattern Recognition) stands as a preeminent academic conference in the realm. The conference is organized by the IEEE (Institute of Electrical and Electronics Engineers).

This challenge session, jointly organized by Beihang University and SenseTime Technology, attracted nearly 100 teams from universities and enterprises. The competition simulated real-world scenarios of autonomous vehicle target detection, encouraging participants to enhance their detection algorithms under unknown testing conditions.

Team leader Hu Shengshan (fourth from the left in the front row) and team members

In the session, ‘Huge’ team achieved the global championship by securing first place in the preliminary round and leading the second-place team by more than 10% in the finals. The team was thereafter invited to present its findings at the CVPR Workshop Forum.

The Trustworthy Artificial Intelligence Team from the School of Cyber Science and Engineering is affiliated with the Key Laboratory of Service Computing Technologies and Systems, Ministry of Education. For years, the team has been deeply involved in areas including AI adversarial attacks and defenses and autonomous driving safety. The team has published a number of research findings in the first half of 2023 at international conferences in the field of security and artificial intelligence such as IEEE S&P 2024, ICCV 2023, CVPR 2023, AAAI 2023, and ACM MM 2023. The team has also submitted relevant patent applications, attracting extensive attention from the industry.

Written by: Zhang Shiyu

Edited by: Yang Kunjie, Chang Wen, Peng Yumeng