This page has the most up-to-date information for our challenges. For detailed information on a method, please click the method name. To sort by a specific metric, click on the header in the table. For further questions, please contact us at jrdb@cs.stanford.edu.

Additional Information Used

  • Individual Image: Method uses individual images from each camera
  • Stitched Image: Method uses stitched images combined from the individual cameras
  • Pointcloud: Method uses 3D pointcloud data
  • Online Tracking: Method does frame-by-frame processing with no lookahead
  • Offline Tracking: Method does not do in-order frame processing
  • Public Detections: Method uses publicly available detections
  • Private Detections: Method uses its own private detections

2D Tracking Leaderboard

Name MOTA ↑ MOTP ↑ IDs ↓ False Positives ↓ False Negatives ↓ Runtime ↓ CPU/GPU
DeepSORT
| |
23.20 24.58 5296 78947 650478 0.025 s 1 GPU (Titan X)
N. Wojke, A. Bewley and D. Paulus. Simple Online and Realtime Tracking with a Deep Association Metric. In ICIP, 2017.
JRMOT2D
| |
22.54 23.62 7719 65550 667783 0.06 s 1 GPU (Titan X)
A. Shenoi, M. Patel, J. Gwak, P. Goebel, A. Sadeghian, H. Rezatofighi, R. Martín-Martín and S. Savarse. JRMOT: A Real-Time 3D Multi-Object Tracker and a New Large-Scale Dataset. In IROS, 2020.
Tracktor++
| |
19.70 26.92 7026 79573 681672 0.2 s 1 GPU (Titan X)
P. Bergmann, T. Meinhardt, L. Leal-Taixé. Tracking without bells and whistles. In ICCV, 2019.

3D Tracking Leaderboard

Name MOTA ↑ MOTP ↑ IDs ↓ False Positives ↓ False Negatives ↓ Runtime ↓ CPU/GPU
JRMOT
| |
20.15 42.46 4207 19711 765907 0.06 s 1 GPU (Titan X)
A. Shenoi, M. Patel, J. Gwak, P. Goebel, A. Sadeghian, H. Rezatofighi, R. Martín-Martín and S. Savarse. JRMOT: A Real-Time 3D Multi-Object Tracker and a New Large-Scale Dataset. In IROS, 2020.
AB3DMOT
| |
19.35 42.02 6177 13664 777946 0.01 s 1 GPU (Titan X)
X. Weng and K. Kitani. A Baseline for 3D Multi-Object Tracking. In IROS, 2020.