Short name:
ApLift
Detector:
Public
Description:
Reference:
A. Hornakova*, T. Kaiser*, M. Rolinek, B. Rosenhahn, P. Swoboda, R. Henschel. Making Higher Order MOT Scalable: An Efficient Approximate Solver for Lifted Disjoint Paths. In International Conference on Computer Vision (ICCV), 2021.
Last submitted:
March 14, 2021 (3 years ago)
Published:
March 22, 2021 at 11:46:35 CET
Submissions:
1
Project page / code:
Open source:
No
Hardware:
5Ghz 8 Cores
Runtime:
0.6 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 | 61.7 | 66.1 | 51.3 | 260 (34.3) | 237 (31.2) | 9,168 | 60,180 | 67.0 | 93.0 | 53.2 | 49.8 | 59.2 | 73.1 | 53.8 | 74.6 | 80.7 | 1.5 | 495 (0.0) | 802 (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 |
| MOT16-01 | 48.8 | 54.3 | 43.9 | 8 | 9 | 122 | 3,145 | 50.8 | 96.4 | 50.8 | 38.1 | 56.3 | 76.3 | 39.8 | 75.4 | 79.9 | 0.3 | 8 | 21 |
| MOT16-03 | 73.7 | 73.3 | 56.5 | 84 | 17 | 4,497 | 22,923 | 78.1 | 94.8 | 55.6 | 57.7 | 60.9 | 73.2 | 62.2 | 75.5 | 80.5 | 3.0 | 118 | 308 |
| MOT16-06 | 57.4 | 61.9 | 49.2 | 98 | 75 | 1,233 | 3,605 | 68.8 | 86.5 | 49.5 | 49.2 | 67.7 | 60.4 | 55.7 | 70.2 | 80.9 | 1.0 | 72 | 92 |
| MOT16-07 | 46.9 | 52.8 | 40.9 | 11 | 9 | 1,090 | 7,499 | 54.1 | 89.0 | 41.5 | 40.7 | 44.9 | 74.2 | 44.0 | 72.4 | 80.0 | 2.2 | 79 | 133 |
| MOT16-08 | 41.7 | 51.8 | 43.9 | 17 | 24 | 499 | 9,220 | 44.9 | 93.8 | 53.7 | 36.0 | 56.3 | 81.8 | 37.9 | 79.1 | 84.0 | 0.8 | 47 | 66 |
| MOT16-12 | 49.4 | 63.4 | 49.5 | 23 | 32 | 580 | 3,597 | 56.6 | 89.0 | 57.0 | 43.1 | 64.9 | 75.5 | 47.2 | 74.2 | 82.9 | 0.6 | 20 | 32 |
| MOT16-14 | 37.8 | 51.0 | 36.7 | 19 | 71 | 1,147 | 10,191 | 44.9 | 87.8 | 41.1 | 32.9 | 46.2 | 68.2 | 35.6 | 69.7 | 79.2 | 1.5 | 151 | 150 |
Raw data: