Overview | Compete

Argoverse 2022 Challenge Winners

Congratulations to the winners of the Argoverse 2022 challenges. To learn about the winning methods, check out our presentation from the CVPR 2022 Workshop on Autonomous Driving and read on for detailed reports from the winning teams:

3D Object Detection

• Detectors. Jin Fang*, Qinghao Meng*, Dingfu Zhou, Chulin Tang, Jianbing Shen, Cheng-Zhong Xu and Liangjun Zhang. Robotics and Autonomous Driving Laboratory, Baidu Research, University of Macau, Beijing Institute of Technology, University of California, Irvine. (Report)

Stereo Depth Estimation

• GMStereo. Haofei Xu, Jing Zhang, Jianfei Cai, Hamid Rezatofighi, Dacheng Tao. Department of Data Science and AI, Monash University, The University of Sydney, JD Explore Academy. (Based on GMFlow)
• MSCLab. Antyanta Bangunharcana, Soohyun Kim, Kyung-Soo Kim. Korea Advanced Institute of Science and Technology. (Report)
• LRM. Luis Rosero, Fernando Osório. Institute of Mathematics and Computer Science, University of São Paulo. (Report)

Motion Forecasting

• BANet. Chen Zhang, Honglin Sun, Chen Chen, Yandong Guo. OPPO Research Institute, Waseda University. (Report)
• QML. Tong Su, Xishun Wang, Xiaodong Yang. QCraft AI. (Report)
• GANet. Mingkun Wang, Changqian Yu, Mingxu Wang, Dongchun Ren, Deheng Qian. Peking University, Meituan, Fudan University. (Report)

Ongoing Argoverse Challenges

Our 2022 Argoverse Challenges have ended, but our leaderboards remain open and we encourage you to use them. Follow the links below for instructions and challenge rules.

Two of our leaderboards use Argoverse 2 datasets:

• 3D Object Detection with the Argoverse 2 Sensor Dataset
• Motion Forecasting with the Argoverse 2 Motion Forecasting Dataset

Four of our leaderboards use Argoverse 1 datasets:

• Stereo Depth Estimation with the Argoverse 1 Stereo Dataset
• Motion Forecasting with the Argoverse 1 Motion Forecasting Dataset
• 3D Tracking with the Argoverse 1 3D Tracking Dataset
• 3D Object Detection with the Argoverse 1 3D Tracking Dataset

Argoverse data is also used by Carnegie Mellon University’s Streaming Perception Challenge. While Argo AI does not endorse the results or conclusions of any third-party, we are pleased to see applications of Argoverse data and encourage participation.

We have hosted three previous rounds of Argoverse competitions. For information on the top performing methods, please see our presentation at the NeurIPS 2019 Workshop on Machine Learning for Autonomous Driving, our presentation at the CVPR 2020 Workshop on Autonomous Driving, and our presentation at the CVPR 2021 Workshop on Autonomous Driving.

To receive Argoverse news and announcements, subscribe to our mailing list.