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 11, 2021 (3 years ago)
Published:
March 22, 2021 at 11:46:23 CET
Submissions:
1
Project page / code:
Open source:
No
Hardware:
5 GHZ, 8 Cores
Runtime:
1.8 Hz
Benchmark performance:
| Sequence | MOTA | IDF1 | HOTA | MT | ML | FP | FN | Rcll | Prcn | AssA | DetA | AssRe | AssPr | DetRe | DetPr | LocA | FAF | ID Sw. | Frag |
| MOT17 | 60.5 | 65.6 | 51.1 | 798 (33.9) | 728 (30.9) | 30,609 | 190,670 | 66.2 | 92.4 | 53.5 | 49.1 | 59.6 | 73.4 | 53.2 | 74.2 | 80.7 | 1.7 | 1,709 (0.0) | 2,672 (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 | 48.8 | 54.3 | 44.0 | 8 | 10 | 113 | 3,181 | 50.7 | 96.7 | 51.1 | 38.0 | 56.4 | 76.5 | 39.6 | 75.6 | 79.8 | 0.3 | 8 | 21 |
| MOT17-01-FRCNN | 47.0 | 57.5 | 46.7 | 9 | 10 | 360 | 3,050 | 52.7 | 90.4 | 56.2 | 39.0 | 62.0 | 75.8 | 42.1 | 72.2 | 80.4 | 0.8 | 11 | 21 |
| MOT17-01-SDP | 45.2 | 55.4 | 44.9 | 9 | 10 | 488 | 3,033 | 53.0 | 87.5 | 53.0 | 38.3 | 56.6 | 78.4 | 42.1 | 69.5 | 80.1 | 1.1 | 13 | 29 |
| MOT17-03-DPM | 73.8 | 73.4 | 56.6 | 85 | 17 | 4,360 | 22,905 | 78.1 | 94.9 | 55.7 | 57.8 | 60.9 | 73.3 | 62.3 | 75.7 | 80.6 | 2.9 | 118 | 261 |
| MOT17-03-FRCNN | 72.8 | 74.7 | 57.5 | 74 | 17 | 3,471 | 24,883 | 76.2 | 95.8 | 58.1 | 57.1 | 62.5 | 76.9 | 61.0 | 76.7 | 80.8 | 2.3 | 109 | 234 |
| MOT17-03-SDP | 77.7 | 75.4 | 58.1 | 94 | 13 | 4,676 | 18,482 | 82.3 | 94.9 | 55.8 | 60.8 | 60.9 | 73.7 | 65.6 | 75.6 | 80.3 | 3.1 | 139 | 386 |
| MOT17-06-DPM | 57.7 | 61.2 | 48.9 | 94 | 76 | 1,142 | 3,765 | 68.0 | 87.5 | 49.0 | 49.1 | 66.9 | 60.5 | 55.2 | 71.0 | 80.9 | 1.0 | 77 | 91 |
| MOT17-06-FRCNN | 57.3 | 58.4 | 47.7 | 102 | 59 | 1,652 | 3,279 | 72.2 | 83.7 | 45.1 | 50.6 | 66.4 | 56.2 | 59.0 | 68.5 | 81.1 | 1.4 | 102 | 140 |
| MOT17-06-SDP | 57.2 | 59.5 | 47.6 | 107 | 58 | 1,700 | 3,251 | 72.4 | 83.4 | 45.4 | 50.3 | 67.7 | 55.5 | 58.8 | 67.7 | 80.8 | 1.4 | 87 | 125 |
| MOT17-07-DPM | 45.7 | 52.5 | 40.4 | 11 | 15 | 1,062 | 8,038 | 52.4 | 89.3 | 41.6 | 39.6 | 45.1 | 74.4 | 42.6 | 72.6 | 80.1 | 2.1 | 80 | 126 |
| MOT17-07-FRCNN | 45.1 | 53.1 | 41.2 | 11 | 15 | 1,345 | 7,862 | 53.5 | 87.0 | 42.8 | 40.0 | 46.9 | 73.1 | 43.7 | 71.1 | 79.9 | 2.7 | 75 | 135 |
| MOT17-07-SDP | 46.6 | 53.8 | 42.1 | 13 | 11 | 1,622 | 7,310 | 56.7 | 85.5 | 42.6 | 42.0 | 46.6 | 71.6 | 46.4 | 70.0 | 79.6 | 3.2 | 87 | 166 |
| MOT17-08-DPM | 33.7 | 44.3 | 39.3 | 17 | 37 | 421 | 13,533 | 35.9 | 94.7 | 53.2 | 29.1 | 55.8 | 81.7 | 30.3 | 79.8 | 84.0 | 0.7 | 48 | 67 |
| MOT17-08-FRCNN | 31.5 | 42.1 | 38.1 | 17 | 37 | 462 | 13,948 | 34.0 | 94.0 | 52.9 | 27.5 | 56.1 | 79.4 | 28.6 | 79.0 | 83.9 | 0.7 | 53 | 74 |
| MOT17-08-SDP | 34.4 | 45.2 | 39.4 | 18 | 34 | 445 | 13,339 | 36.9 | 94.6 | 52.6 | 29.6 | 56.0 | 80.4 | 30.8 | 79.0 | 83.5 | 0.7 | 63 | 85 |
| MOT17-12-DPM | 47.6 | 61.9 | 48.6 | 23 | 36 | 563 | 3,959 | 54.3 | 89.3 | 57.0 | 41.5 | 64.9 | 75.6 | 45.3 | 74.4 | 82.9 | 0.6 | 20 | 32 |
| MOT17-12-FRCNN | 47.8 | 62.3 | 48.4 | 18 | 40 | 296 | 4,219 | 51.3 | 93.8 | 58.2 | 40.3 | 65.7 | 78.0 | 42.7 | 78.1 | 83.4 | 0.3 | 13 | 24 |
| MOT17-12-SDP | 50.0 | 66.1 | 50.8 | 19 | 42 | 488 | 3,836 | 55.7 | 90.8 | 60.8 | 42.5 | 68.4 | 77.1 | 46.1 | 75.0 | 82.7 | 0.5 | 11 | 31 |
| MOT17-14-DPM | 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 |
| MOT17-14-FRCNN | 33.9 | 48.4 | 35.8 | 25 | 62 | 2,369 | 9,636 | 47.9 | 78.9 | 39.1 | 33.1 | 45.4 | 63.0 | 37.8 | 62.3 | 78.1 | 3.2 | 206 | 228 |
| MOT17-14-SDP | 37.1 | 49.9 | 37.2 | 25 | 58 | 2,427 | 8,970 | 51.5 | 79.7 | 39.6 | 35.2 | 47.0 | 62.5 | 40.4 | 62.5 | 78.2 | 3.2 | 238 | 246 |
Raw data: