Recently, the high-profile leading international conference in the field of  artificial intelligence (AI) and computer vision, CVPR 2020 (IEEE Conference on Computer Vision and Pattern Recognition), was held online for the first time. The team, jointly formed by Prof. Zhong Bineng's group, from College of Computer Science and Technology, Huaqiao University (HQU), and Didi Map Vision, participated in the Track 1 competition, namely Multi-Class Multi-Movement Vehicle Counting and won the second prize among more than 100 teams in the world. Three papers of Prof. Zhong’s group have been included in CVPR 2020.

The AI City Challenge, hosted by CVPR Workshop, is known as the "ImageNet Competition in Intelligent Traffic Video Analytics" and has attracted much attention since it was held in 2017. There are four challenge tracks in the competition: Vehicle Counting, Vehicle Re-identification, Vehicle Tracking, and Traffic Anomaly Detection.

This competition attracted active participation of more than 100 top AI and vision R&D teams in academia and the industrial community, including Carnegie Mellon University, University of Illinois, State University of New York, University of Technology Sydney, ENGIE Group (France), Brno University of Technology (Czech), Korea Advanced Institute of Science and Technology, Baidu, and Didi.

Prof. Zhong's research group is composed of Zhou Qinqin and other postgraduates under his supervision. Their main research areas are computer vision and pattern recognition, covering visual object tracking, confrontation learning and attack, and medical endoscope image processing. Relevant codes, models and data of the research direction are open-source on the Github website, and have received extensive attention from the industry and academia.

Screenshot of the Competition