BLSTM_MTP_O: Online Greedy Data Association using Bilinear LSTM with Multi-track Pooling
MOT17-08-FRCNN
Short name:
BLSTM_MTP_O
Detector:
Public
Description:
Reference:
C. Kim, F. Li, M. Alotaibi, J. Rehg. Discriminative Appearance Modeling with Multi-track Pooling for Real-time Multi-object Tracking. In CVPR (accepted), 2021.
Last submitted:
March 17, 2021 (3 years ago)
Published:
March 17, 2021 at 03:16:25 CET
Submissions:
1
Project page / code:
n/a
Open source:
No
Hardware:
Titan Xp GPU, 3.6GHZ CPU
Runtime:
20.1 Hz
Benchmark performance:
| Sequence | MOTA | IDF1 | HOTA | MT | ML | FP | FN | Rcll | Prcn | AssA | DetA | AssRe | AssPr | DetRe | DetPr | LocA | FAF | ID Sw. | Frag |
| MOT17 | 51.5 | 54.9 | 41.3 | 481 (20.4) | 836 (35.5) | 29,616 | 241,619 | 57.2 | 91.6 | 41.6 | 41.3 | 44.8 | 72.0 | 44.3 | 70.9 | 77.9 | 1.7 | 2,566 (0.0) | 7,748 (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 |
| MOT17-01-DPM | 32.7 | 42.2 | 32.8 | 4 | 9 | 621 | 3,696 | 42.7 | 81.6 | 36.4 | 29.6 | 39.7 | 68.5 | 32.8 | 62.7 | 76.7 | 1.4 | 25 | 84 |
| MOT17-01-FRCNN | 29.4 | 51.0 | 39.6 | 6 | 8 | 1,410 | 3,130 | 51.5 | 70.2 | 48.1 | 32.9 | 52.1 | 73.0 | 40.3 | 54.9 | 77.0 | 3.1 | 13 | 66 |
| MOT17-01-SDP | 35.2 | 49.3 | 38.8 | 8 | 5 | 1,662 | 2,465 | 61.8 | 70.6 | 39.8 | 38.1 | 42.9 | 73.5 | 48.4 | 55.3 | 77.7 | 3.7 | 52 | 136 |
| MOT17-03-DPM | 56.6 | 57.9 | 43.1 | 38 | 21 | 5,552 | 39,564 | 62.2 | 92.1 | 41.3 | 45.4 | 43.8 | 73.6 | 48.4 | 71.8 | 77.7 | 3.7 | 324 | 953 |
| MOT17-03-FRCNN | 59.7 | 58.8 | 43.9 | 44 | 25 | 3,865 | 38,127 | 63.6 | 94.5 | 41.6 | 46.6 | 44.5 | 71.7 | 49.3 | 73.3 | 78.1 | 2.6 | 211 | 534 |
| MOT17-03-SDP | 73.3 | 66.8 | 50.6 | 71 | 14 | 3,155 | 24,420 | 76.7 | 96.2 | 46.2 | 55.7 | 49.6 | 71.5 | 59.1 | 74.2 | 78.4 | 2.1 | 333 | 1,327 |
| MOT17-06-DPM | 47.6 | 55.2 | 41.2 | 39 | 107 | 317 | 5,788 | 50.9 | 95.0 | 45.4 | 37.5 | 54.3 | 67.3 | 39.4 | 73.5 | 78.6 | 0.3 | 74 | 269 |
| MOT17-06-FRCNN | 55.6 | 59.1 | 44.5 | 54 | 71 | 482 | 4,679 | 60.3 | 93.6 | 44.8 | 44.4 | 54.3 | 65.9 | 47.2 | 73.3 | 79.0 | 0.4 | 74 | 255 |
| MOT17-06-SDP | 56.9 | 63.3 | 47.3 | 73 | 72 | 582 | 4,404 | 62.6 | 92.7 | 49.8 | 45.1 | 55.9 | 71.2 | 48.6 | 71.9 | 79.0 | 0.5 | 89 | 266 |
| MOT17-07-DPM | 43.2 | 47.7 | 34.7 | 6 | 24 | 486 | 9,028 | 46.6 | 94.2 | 35.6 | 34.0 | 37.2 | 75.5 | 35.7 | 72.1 | 77.7 | 1.0 | 85 | 343 |
| MOT17-07-FRCNN | 39.1 | 41.5 | 31.6 | 8 | 16 | 1,126 | 8,942 | 47.1 | 87.6 | 30.1 | 33.6 | 32.3 | 69.3 | 36.3 | 67.6 | 77.4 | 2.3 | 217 | 446 |
| MOT17-07-SDP | 49.5 | 48.4 | 36.8 | 14 | 15 | 779 | 7,606 | 55.0 | 92.3 | 34.7 | 39.4 | 36.8 | 72.6 | 42.1 | 70.7 | 77.8 | 1.6 | 141 | 431 |
| MOT17-08-DPM | 28.3 | 35.3 | 28.2 | 11 | 34 | 680 | 14,360 | 32.0 | 90.9 | 33.0 | 24.2 | 34.9 | 75.0 | 25.3 | 71.7 | 79.2 | 1.1 | 98 | 189 |
| MOT17-08-FRCNN | 24.4 | 33.6 | 28.2 | 9 | 38 | 1,150 | 14,749 | 30.2 | 84.7 | 35.5 | 22.6 | 37.6 | 75.6 | 24.1 | 67.5 | 78.7 | 1.8 | 65 | 118 |
| MOT17-08-SDP | 30.9 | 35.5 | 29.4 | 12 | 34 | 894 | 13,562 | 35.8 | 89.4 | 32.7 | 26.8 | 34.4 | 72.7 | 28.2 | 70.4 | 78.1 | 1.4 | 144 | 295 |
| MOT17-12-DPM | 40.4 | 55.1 | 40.1 | 19 | 40 | 671 | 4,452 | 48.6 | 86.3 | 47.0 | 34.3 | 49.9 | 74.0 | 37.5 | 66.6 | 77.6 | 0.7 | 45 | 176 |
| MOT17-12-FRCNN | 38.0 | 51.2 | 37.8 | 15 | 43 | 646 | 4,700 | 45.8 | 86.0 | 44.0 | 32.5 | 50.3 | 67.9 | 35.4 | 66.6 | 77.7 | 0.7 | 30 | 100 |
| MOT17-12-SDP | 40.3 | 57.1 | 41.0 | 17 | 41 | 803 | 4,328 | 50.1 | 84.4 | 49.2 | 34.4 | 52.8 | 72.1 | 38.3 | 64.5 | 77.4 | 0.9 | 44 | 154 |
| MOT17-14-DPM | 27.3 | 42.0 | 29.2 | 10 | 87 | 1,026 | 12,338 | 33.2 | 85.7 | 35.6 | 24.0 | 37.5 | 72.0 | 25.5 | 65.8 | 76.9 | 1.4 | 78 | 341 |
| MOT17-14-FRCNN | 27.0 | 39.7 | 28.6 | 10 | 74 | 1,864 | 11,365 | 38.5 | 79.2 | 30.9 | 26.7 | 33.2 | 66.7 | 29.4 | 60.5 | 74.7 | 2.5 | 256 | 656 |
| MOT17-14-SDP | 35.5 | 47.3 | 33.4 | 13 | 58 | 1,845 | 9,916 | 46.4 | 82.3 | 35.4 | 31.7 | 37.5 | 71.4 | 35.2 | 62.5 | 75.8 | 2.5 | 168 | 609 |
Raw data:
