Table 3 Comparison with other methods on MOT15, MOT16 and MOT17 train and test datasets at small 576x320 resolution.
From: Dual attention for multi object tracking with intra sample context and cross sample interaction
Dataset | Method | MOTA \(\uparrow\) | IDF1 \(\uparrow\) | MT \(\uparrow\) | ML \(\downarrow\) | IDs \(\downarrow\) | FPS \(\uparrow\) |
|---|---|---|---|---|---|---|---|
MOT15-Train | JDE(Wang et al., 2020)10 | 48.7 | 58.9 | 328 | 80 | 753 | 44.2 |
| Â | TADN(Psalta et al., 2024)6 | 47.7 | 54.6 | 277 | 105 | 781 | 13.8 |
| Â | Ours | 52.4 | 59.9 | 320 | 77 | 760 | 56.8 |
MOT16-Train | JDE(Wang et al., 2020)10 | 61.6 | 61.8 | 199 | 75 | 1483 | 40.7 |
| Â | TADN(Psalta et al., 2024)6 | 60.1 | 60.4 | 132 | 164 | 1414 | 12.1 |
| Â | Ours | 61.9 | 63.6 | 195 | 79 | 1303 | 53.2 |
MOT15-Test | JDE(Wang et al., 2020)10 | 36.3 | 47.9 | 158 | 168 | 1453 | 44.8 |
| Â | TADN(Psalta et al., 2024)6 | 42.9 | 49.0 | 112 | 241 | 1595 | 14.4 |
| Â | Ours | 44.6 | 50.9 | 174 | 122 | 1548 | 57.5 |
MOT16-Test | JDE(Wang et al., 2020)10 | 45.7 | 44.5 | 122 | 309 | 1768 | 41.6 |
| Â | TADN(Psalta et al., 2024)6 | 46.1 | 43.5 | 137 | 296 | 1832 | 12.3 |
| Â | Ours | 50.3 | 47.6 | 152 | 272 | 1811 | 53.5 |
MOT17-Test | JDE(Wang et al., 2020)10 | 44.5 | 43.8 | 369 | 999 | 5286 | 40.3 |
| Â | TADN(Psalta et al., 2024)6 | 43.7 | 43.3 | 352 | 1011 | 5312 | 11.6 |
| Â | Ours | 49.0 | 46.8 | 456 | 885 | 5463 | 52.9 |
