HGFMOT: optimize the backbone structures and loss functions based on anchor free framework
MOT16-08
Short name:
HGFMOT
Detector:
Private
Description:
Reference:
submission by AI Lab, Lenovo Research
Last submitted:
January 12, 2021 (3 years ago)
Published:
November 26, 2020 at 11:30:10 CET
Submissions:
3
Project page / code:
n/a
Open source:
No
Hardware:
V100
Runtime:
9.9 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 | 77.6 | 74.9 | 61.5 | 362 (47.7) | 93 (12.3) | 9,620 | 30,073 | 83.5 | 94.1 | 59.7 | 63.5 | 65.2 | 77.7 | 69.3 | 78.0 | 83.4 | 1.6 | 1,067 (12.8) | 2,201 (26.4) |
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 | 62.8 | 63.4 | 51.2 | 11 | 1 | 498 | 1,824 | 71.5 | 90.2 | 49.8 | 52.9 | 54.2 | 78.5 | 58.9 | 74.4 | 83.1 | 1.1 | 54 | 145 |
| MOT16-03 | 91.0 | 83.9 | 69.7 | 134 | 1 | 3,339 | 5,883 | 94.4 | 96.7 | 66.3 | 73.5 | 72.0 | 80.0 | 78.8 | 80.8 | 84.0 | 2.2 | 157 | 499 |
| MOT16-06 | 63.6 | 65.9 | 52.9 | 97 | 43 | 789 | 3,199 | 72.3 | 91.4 | 52.6 | 53.6 | 59.4 | 75.9 | 59.2 | 74.8 | 82.6 | 0.7 | 207 | 244 |
| MOT16-07 | 68.0 | 65.7 | 52.2 | 25 | 1 | 939 | 4,178 | 74.4 | 92.8 | 48.9 | 55.9 | 53.0 | 73.5 | 61.1 | 76.3 | 82.9 | 1.9 | 107 | 302 |
| MOT16-08 | 52.7 | 45.8 | 40.0 | 21 | 2 | 2,086 | 5,576 | 66.7 | 84.3 | 33.3 | 48.7 | 39.2 | 63.6 | 56.0 | 70.8 | 83.0 | 3.3 | 248 | 422 |
| MOT16-12 | 56.2 | 64.1 | 52.5 | 29 | 14 | 960 | 2,640 | 68.2 | 85.5 | 55.1 | 50.2 | 60.6 | 78.3 | 57.6 | 72.2 | 84.5 | 1.1 | 35 | 150 |
| MOT16-14 | 56.5 | 67.1 | 47.6 | 45 | 31 | 1,009 | 6,773 | 63.4 | 92.1 | 50.1 | 45.6 | 53.9 | 75.7 | 49.6 | 72.1 | 80.0 | 1.3 | 259 | 439 |
Raw data:
