TDANet: Transformer-Based Detection and Association Network for Multi-Object Tracking
MOT16-07
Short name:
TDANet
Detector:
Private
Description:
Reference:
Yu-Lei Li, Xiamen University, chaorengui@gmail.com
Last submitted:
July 02, 2021 (2 years ago)
Published:
July 02, 2021 at 04:37:53 CET
Submissions:
1
Project page / code:
n/a
Open source:
No
Hardware:
a 2080TI gpu
Runtime:
21.7 Hz
Benchmark performance:
| Sequence | MOTA | IDF1 | HOTA | MT | ML | FP | FN | Rcll | Prcn | AssA | DetA | AssRe | AssPr | DetRe | DetPr | LocA | FAF | ID Sw. | Frag |
| MOT16 | 77.7 | 76.2 | 62.4 | 348 (45.8) | 99 (13.0) | 9,571 | 30,417 | 83.3 | 94.1 | 61.7 | 63.3 | 67.0 | 78.5 | 69.1 | 78.0 | 83.4 | 1.6 | 682 (0.0) | 1,956 (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 |
| MOT16-01 | 62.9 | 63.6 | 50.8 | 12 | 1 | 574 | 1,779 | 72.2 | 88.9 | 49.2 | 52.6 | 54.6 | 75.7 | 58.9 | 72.6 | 82.0 | 1.3 | 19 | 109 |
| MOT16-03 | 91.2 | 85.3 | 71.7 | 132 | 1 | 3,721 | 5,431 | 94.8 | 96.4 | 68.4 | 75.3 | 74.1 | 80.7 | 80.7 | 82.0 | 85.1 | 2.5 | 86 | 449 |
| MOT16-06 | 63.1 | 67.2 | 52.8 | 86 | 49 | 633 | 3,496 | 69.7 | 92.7 | 54.0 | 52.0 | 61.3 | 75.2 | 56.7 | 75.4 | 82.3 | 0.5 | 125 | 230 |
| MOT16-07 | 68.1 | 63.2 | 49.5 | 25 | 1 | 986 | 4,131 | 74.7 | 92.5 | 45.3 | 54.3 | 49.0 | 70.5 | 59.5 | 73.7 | 80.9 | 2.0 | 85 | 271 |
| MOT16-08 | 52.4 | 52.1 | 41.1 | 21 | 2 | 1,936 | 5,883 | 64.9 | 84.9 | 38.5 | 44.5 | 42.7 | 66.0 | 50.6 | 66.2 | 78.4 | 3.1 | 154 | 355 |
| MOT16-12 | 57.0 | 66.8 | 51.8 | 28 | 15 | 866 | 2,677 | 67.7 | 86.6 | 55.3 | 49.0 | 59.9 | 78.4 | 55.5 | 71.0 | 82.6 | 1.0 | 22 | 123 |
| MOT16-14 | 56.4 | 65.0 | 44.9 | 44 | 30 | 855 | 7,020 | 62.0 | 93.1 | 48.2 | 42.7 | 50.8 | 76.9 | 46.1 | 69.2 | 77.1 | 1.1 | 191 | 419 |
Raw data:
