Short name:
MOTer
Detector:
Public
Description:
TransCenter on public detections trained with COCO, CrowdedHuman and MOT.
Reference:
Y. Xu, Y. Ban, G. Delorme, C. Gan, D. Rus, X. Alameda-Pineda. TransCenter: Transformers with Dense Queries for Multiple-Object Tracking. In arXiv, 2021.
Last submitted:
July 15, 2021 (3 years ago)
Published:
March 28, 2021 at 15:55:33 CET
Submissions:
4
Project page / code:
Open source:
Yes
Hardware:
RTX Titan
Runtime:
1.0 Hz
Benchmark performance:
Sequence | MOTA | IDF1 | HOTA | MT | ML | FP | FN | Rcll | Prcn | AssA | DetA | AssRe | AssPr | DetRe | DetPr | LocA | FAF | ID Sw. | Frag |
MOT20 | 59.1 | 49.9 | 43.6 | 598 (48.1) | 189 (15.2) | 59,624 | 147,333 | 71.5 | 86.1 | 37.0 | 51.8 | 45.2 | 58.8 | 58.9 | 71.0 | 81.2 | 13.3 | 4,597 (0.0) | 9,232 (0.0) |
Detailed performance:
Sequence | MOTA | IDF1 | HOTA | MT | ML | FP | FN | Rcll | Prcn | AssA | DetA | AssRe | AssPr | DetRe | DetPr | LocA | FAF | ID Sw. | Frag |
MOT20-04 | 74.8 | 61.2 | 52.4 | 416 | 48 | 22,616 | 45,230 | 83.5 | 91.0 | 43.9 | 63.1 | 53.2 | 65.7 | 69.7 | 75.9 | 82.0 | 10.9 | 1,319 | 3,083 |
MOT20-06 | 42.4 | 33.2 | 28.9 | 79 | 73 | 18,103 | 56,401 | 57.5 | 80.8 | 21.2 | 40.0 | 27.5 | 43.3 | 46.2 | 65.0 | 79.6 | 18.0 | 1,907 | 3,526 |
MOT20-07 | 73.3 | 58.2 | 50.0 | 67 | 7 | 2,480 | 6,070 | 81.7 | 91.6 | 41.2 | 61.2 | 49.1 | 60.8 | 67.9 | 76.2 | 82.8 | 4.2 | 280 | 433 |
MOT20-08 | 26.2 | 29.6 | 25.9 | 36 | 61 | 16,425 | 39,632 | 48.9 | 69.7 | 21.6 | 31.7 | 26.5 | 44.4 | 38.9 | 55.5 | 78.6 | 20.4 | 1,091 | 2,190 |
Raw data: