A Real-time Deep Graph Matching for Multi-object Tracking
Short name:
youtu_face
Benchmark:
Description:
Hardware:
2.5GHz,2 Core
Detector:
Private
Processing:
Batch
Last submitted:
April 26, 2018 (1 year ago)
Published:
May 06, 2019 at 04:27:37 CET
Submissions:
1
Open source:
No
Project page / code:
n/a
Reference:
. A Real-time Deep Graph Matching for Multi-object Tracking. In Tech Report, Tencent, 2018.
Benchmark performance:
| MOTA | MOTP | FAF | MT | ML | FP | FN | ID Sw. | Frag | Specifications | Detector |
| 56.8 | 75.6 | 1.1 | 43.3 % | 21.2 % | 6,452 | 19,642 | 459 | 900 | 2.5GHz,2 Core | Private |
| IDF1 | ID Precision | ID Recall |
| 59.4 | 67.6 | 53.1 |
Detailed performance:
| Sequence | MOTA | IDF1 | MOTP | FAF | GT | MT | ML | FP | FN | ID Sw | Frag |
| TUD-Crossing | 87.6 | 78.2 | 74.5 | 0.1 | 13 | 76.9 % | 7.7 % | 13 | 120 | 4 | 6 |
| PETS09-S2L2 | 67.8 | 44.7 | 75.1 | 1.4 | 42 | 33.3 % | 4.8 % | 629 | 2,286 | 187 | 225 |
| ETH-Jelmoli | 52.5 | 66.3 | 78.4 | 1.2 | 45 | 44.4 % | 33.3 % | 525 | 670 | 11 | 27 |
| ETH-Linthescher | 66.3 | 67.1 | 79.5 | 0.7 | 197 | 41.6 % | 28.9 % | 876 | 2,078 | 59 | 71 |
| ETH-Crossing | 76.2 | 80.2 | 84.8 | 0.3 | 26 | 38.5 % | 34.6 % | 58 | 176 | 5 | 8 |
| AVG-TownCentre | 68.6 | 77.7 | 71.4 | 1.6 | 226 | 53.5 % | 15.5 % | 728 | 1,466 | 50 | 226 |
| ADL-Rundle-1 | 45.2 | 54.3 | 75.5 | 1.5 | 32 | 28.1 % | 21.9 % | 773 | 4,295 | 32 | 66 |
| ADL-Rundle-3 | 52.5 | 52.0 | 78.9 | 1.5 | 44 | 29.5 % | 13.6 % | 943 | 3,829 | 58 | 93 |
| KITTI-16 | 49.5 | 64.6 | 71.1 | 1.6 | 17 | 41.2 % | 5.9 % | 339 | 504 | 16 | 45 |
| KITTI-19 | 39.0 | 59.3 | 68.4 | 1.2 | 62 | 33.9 % | 22.6 % | 1,218 | 2,010 | 31 | 119 |
| Venice-1 | 43.8 | 54.8 | 74.7 | 0.8 | 17 | 29.4 % | 35.3 % | 350 | 2,208 | 6 | 14 |
Raw data:
