Benchmark:
MOT16 |
Short name:
SDMT
Detector:
Public
Description:
Reference:
M. Thoreau, N. Kottege. Deep Similarity Metric Learning for Real-Time Pedestrian Tracking. In arXiv, 2018.
Last submitted:
March 12, 2018 (6 years ago)
Published:
November 11, 2019 at 13:42:38 CET
Submissions:
2
Project page / code:
Open source:
No
Hardware:
i7 3.6 GHz - gtx 1080 ti
Runtime:
19.8 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 | 39.6 | 42.3 | 32.8 | 89 (11.7) | 373 (49.1) | 11,130 | 98,343 | 46.1 | 88.3 | 32.1 | 33.6 | 35.0 | 70.1 | 36.2 | 69.4 | 78.6 | 1.9 | 602 (13.1) | 772 (16.8) |
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 | 26.7 | 35.8 | 28.4 | 4 | 12 | 169 | 4,500 | 29.6 | 91.8 | 36.6 | 22.1 | 44.0 | 61.9 | 22.9 | 70.9 | 76.9 | 0.4 | 19 | 26 |
| MOT16-03 | 45.6 | 44.3 | 34.3 | 26 | 33 | 8,308 | 48,194 | 53.9 | 87.2 | 30.6 | 38.7 | 33.3 | 68.4 | 42.5 | 68.7 | 78.7 | 5.5 | 325 | 405 |
| MOT16-06 | 40.4 | 45.6 | 34.3 | 26 | 124 | 490 | 6,333 | 45.1 | 91.4 | 36.5 | 32.3 | 44.1 | 66.4 | 34.3 | 69.5 | 77.2 | 0.4 | 56 | 77 |
| MOT16-07 | 34.1 | 40.7 | 30.6 | 6 | 20 | 836 | 9,861 | 39.6 | 88.5 | 32.5 | 28.9 | 34.1 | 74.5 | 30.6 | 68.4 | 76.8 | 1.7 | 67 | 107 |
| MOT16-08 | 33.7 | 37.2 | 31.5 | 11 | 26 | 528 | 10,503 | 37.2 | 92.2 | 34.6 | 28.7 | 36.6 | 76.2 | 30.1 | 74.5 | 80.7 | 0.8 | 72 | 74 |
| MOT16-12 | 39.8 | 51.1 | 37.7 | 11 | 46 | 372 | 4,599 | 44.6 | 90.9 | 43.5 | 32.8 | 45.4 | 77.9 | 34.9 | 71.2 | 79.4 | 0.4 | 23 | 29 |
| MOT16-14 | 19.8 | 29.5 | 22.3 | 5 | 112 | 427 | 14,353 | 22.3 | 90.6 | 29.8 | 16.9 | 31.3 | 74.8 | 17.4 | 70.5 | 78.0 | 0.6 | 40 | 54 |
Raw data: