DCCRF:
TUD-Crossing
Benchmark:
Short name:
DCCRF
Detector:
Public
Description:
Reference:
H. Zhou, W. Ouyang, J. Cheng, X. Wang, H. Li. Deep Continuous Conditional Random Fields with Asymmetric Inter-object Constraints for Online Multi-object Tracking. In IEEE Transactions on Circuits and Systems for Video Technology, 2018.
Last submitted:
June 26, 2017 (3 years ago)
Published:
April 12, 2018 at 08:22:16 CET
Submissions:
2
Project page / code:
n/a
Open source:
No
Hardware:
1*Titanx
Runtime:
0.1 Hz
Benchmark performance:
| Sequence | MOTA | IDF1 | MOTP | MT | ML | FP | FN | Recall | Precision | FAF | ID Sw. | Frag |
| 2D MOT 2015 | 33.6 | 39.1 | 70.9 | 75 (10.4) | 271 (37.6) | 5,917 | 34,002 | 44.7 | 82.3 | 1.0 | 866 (19.4) | 1,566 (35.1) |
Detailed performance:
| Sequence | MOTA | IDF1 | MOTP | MT | ML | FP | FN | Recall | Precision | FAF | ID Sw. | Frag |
| ADL-Rundle-1 | 28.3 | 40.0 | 71.4 | 3 | 9 | 1,300 | 5,312 | 42.9 | 75.4 | 2.6 | 59 | 125 |
| ADL-Rundle-3 | 38.7 | 39.3 | 73.2 | 3 | 11 | 553 | 5,601 | 44.9 | 89.2 | 0.9 | 74 | 103 |
| AVG-TownCentre | 32.3 | 43.2 | 68.9 | 28 | 66 | 777 | 3,831 | 46.4 | 81.0 | 1.7 | 229 | 395 |
| ETH-Crossing | 22.9 | 29.3 | 74.3 | 2 | 16 | 45 | 714 | 28.8 | 86.5 | 0.2 | 14 | 15 |
| ETH-Jelmoli | 43.5 | 61.7 | 72.4 | 9 | 14 | 374 | 1,040 | 59.0 | 80.0 | 0.9 | 19 | 49 |
| ETH-Linthescher | 24.1 | 32.4 | 74.1 | 14 | 132 | 309 | 6,382 | 28.5 | 89.2 | 0.3 | 85 | 97 |
| KITTI-16 | 42.0 | 54.1 | 72.5 | 2 | 2 | 155 | 812 | 52.3 | 85.2 | 0.7 | 19 | 37 |
| KITTI-19 | 26.1 | 45.4 | 65.6 | 5 | 11 | 1,088 | 2,797 | 47.7 | 70.1 | 1.0 | 66 | 236 |
| PETS09-S2L2 | 45.6 | 29.9 | 69.6 | 4 | 4 | 664 | 4,335 | 55.0 | 88.9 | 1.5 | 245 | 416 |
| TUD-Crossing | 58.6 | 50.5 | 73.6 | 3 | 1 | 96 | 340 | 69.1 | 88.8 | 0.5 | 20 | 24 |
| Venice-1 | 24.8 | 31.9 | 70.8 | 2 | 5 | 556 | 2,838 | 37.8 | 75.6 | 1.2 | 36 | 69 |
Raw data:
