MOTer: TransCenter_Pub_CH
MOT17-01-FRCNN
Short name:
MOTer
Detector:
Public
Description:
TransCenter on public detections trained with COCO, CrowdedHuman and MOT.
Reference:
Y. Xu, Y. Ban, G. Delorme, C. Gan, D. Rus, X. Alameda-Pineda. TransCenter: Transformers with Dense Queries for Multiple-Object Tracking. In arXiv, 2021.
Last submitted:
March 12, 2021 (3 years ago)
Published:
March 28, 2021 at 15:54:32 CET
Submissions:
3
Project page / code:
Open source:
Yes
Hardware:
RTX TiTan
Runtime:
1.0 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 | 71.9 | 62.3 | 54.1 | 894 (38.0) | 534 (22.7) | 17,378 | 137,008 | 75.7 | 96.1 | 49.9 | 59.0 | 54.3 | 76.0 | 63.0 | 80.0 | 83.7 | 1.0 | 4,046 (0.0) | 7,929 (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 | 49.8 | 40.3 | 37.1 | 8 | 9 | 448 | 2,745 | 57.4 | 89.2 | 33.2 | 42.0 | 35.4 | 72.4 | 46.2 | 71.8 | 81.6 | 1.0 | 46 | 75 |
| MOT17-01-FRCNN | 49.8 | 40.1 | 36.3 | 8 | 9 | 480 | 2,709 | 58.0 | 88.6 | 31.5 | 42.3 | 35.9 | 69.1 | 46.7 | 71.4 | 81.6 | 1.1 | 47 | 79 |
| MOT17-01-SDP | 50.4 | 39.7 | 36.4 | 8 | 9 | 490 | 2,662 | 58.7 | 88.5 | 31.2 | 42.8 | 35.6 | 69.1 | 47.4 | 71.4 | 81.5 | 1.1 | 49 | 86 |
| MOT17-03-DPM | 89.6 | 73.8 | 63.8 | 126 | 5 | 2,514 | 8,167 | 92.2 | 97.5 | 56.3 | 72.5 | 60.7 | 78.5 | 77.2 | 81.6 | 84.1 | 1.7 | 226 | 620 |
| MOT17-03-FRCNN | 88.2 | 73.7 | 63.4 | 123 | 8 | 2,505 | 9,619 | 90.8 | 97.4 | 56.4 | 71.6 | 60.6 | 79.1 | 76.1 | 81.7 | 84.2 | 1.7 | 224 | 600 |
| MOT17-03-SDP | 88.9 | 73.0 | 63.4 | 122 | 9 | 2,731 | 8,697 | 91.7 | 97.2 | 56.0 | 72.0 | 60.5 | 78.3 | 76.8 | 81.4 | 84.2 | 1.8 | 240 | 658 |
| MOT17-06-DPM | 61.2 | 56.3 | 47.0 | 76 | 56 | 497 | 3,900 | 66.9 | 94.1 | 43.1 | 51.6 | 49.7 | 73.7 | 55.4 | 77.9 | 83.3 | 0.4 | 170 | 242 |
| MOT17-06-FRCNN | 63.5 | 56.8 | 47.5 | 82 | 42 | 543 | 3,552 | 69.9 | 93.8 | 42.5 | 53.5 | 49.0 | 73.2 | 57.7 | 77.5 | 83.1 | 0.5 | 201 | 307 |
| MOT17-06-SDP | 62.9 | 56.9 | 47.6 | 84 | 50 | 556 | 3,617 | 69.3 | 93.6 | 43.0 | 53.0 | 49.4 | 73.6 | 57.3 | 77.4 | 83.2 | 0.5 | 194 | 284 |
| MOT17-07-DPM | 58.7 | 48.1 | 40.8 | 15 | 8 | 683 | 6,100 | 63.9 | 94.0 | 35.3 | 48.2 | 38.8 | 69.9 | 51.9 | 76.4 | 82.1 | 1.4 | 190 | 439 |
| MOT17-07-FRCNN | 58.8 | 48.5 | 40.7 | 15 | 6 | 674 | 6,085 | 64.0 | 94.1 | 35.1 | 48.4 | 38.5 | 71.6 | 52.0 | 76.5 | 82.1 | 1.3 | 194 | 438 |
| MOT17-07-SDP | 59.8 | 47.7 | 40.8 | 16 | 6 | 702 | 5,885 | 65.2 | 94.0 | 34.7 | 49.1 | 38.3 | 70.2 | 52.9 | 76.3 | 82.0 | 1.4 | 197 | 451 |
| MOT17-08-DPM | 46.2 | 36.1 | 34.6 | 22 | 21 | 422 | 10,662 | 49.5 | 96.1 | 30.8 | 39.6 | 34.9 | 65.8 | 41.4 | 80.4 | 84.0 | 0.7 | 280 | 424 |
| MOT17-08-FRCNN | 45.7 | 36.5 | 34.4 | 21 | 21 | 395 | 10,815 | 48.8 | 96.3 | 30.9 | 39.1 | 35.1 | 66.8 | 40.9 | 80.7 | 84.1 | 0.6 | 269 | 411 |
| MOT17-08-SDP | 46.6 | 36.1 | 34.5 | 22 | 20 | 427 | 10,571 | 50.0 | 96.1 | 30.5 | 39.8 | 34.9 | 65.5 | 41.7 | 80.3 | 83.9 | 0.7 | 279 | 432 |
| MOT17-12-DPM | 59.5 | 62.3 | 52.4 | 30 | 28 | 334 | 3,121 | 64.0 | 94.3 | 53.4 | 51.5 | 61.2 | 77.3 | 55.0 | 81.1 | 85.9 | 0.4 | 51 | 127 |
| MOT17-12-FRCNN | 59.3 | 61.8 | 52.1 | 30 | 29 | 272 | 3,208 | 63.0 | 95.3 | 53.3 | 51.0 | 60.7 | 77.4 | 54.1 | 81.9 | 85.9 | 0.3 | 50 | 108 |
| MOT17-12-SDP | 59.7 | 61.7 | 52.4 | 30 | 27 | 361 | 3,077 | 64.5 | 93.9 | 53.6 | 51.5 | 60.7 | 77.4 | 55.2 | 80.4 | 85.8 | 0.4 | 53 | 125 |
| MOT17-14-DPM | 34.9 | 36.5 | 28.9 | 17 | 61 | 674 | 11,050 | 40.2 | 91.7 | 28.7 | 29.9 | 31.5 | 66.8 | 31.6 | 72.1 | 79.6 | 0.9 | 316 | 585 |
| MOT17-14-FRCNN | 36.9 | 37.1 | 29.5 | 19 | 55 | 843 | 10,445 | 43.5 | 90.5 | 27.9 | 32.1 | 31.3 | 64.4 | 34.2 | 71.2 | 79.1 | 1.1 | 383 | 697 |
| MOT17-14-SDP | 37.6 | 38.2 | 30.1 | 20 | 55 | 827 | 10,321 | 44.2 | 90.8 | 28.7 | 32.6 | 31.6 | 65.2 | 34.7 | 71.4 | 79.1 | 1.1 | 387 | 741 |
Raw data:
