Online Multi-Object Tracking with the GMPHD Filter Using Occlusion Group Management
Short name:
GMPHDOGM17
Benchmark:
Description:
The source code is available at https://github.com/SonginCV/GMPHD-OGM
Hardware:
Intel i7-7700K @ 4.2GHz, 4 Cores, no GPU
Detector:
Public
Processing:
Online
Last submitted:
June 30, 2019 (3 months ago)
Published:
August 01, 2019 at 08:45:27 CET
Submissions:
2
Open source:
Yes
Project page / code:
n/a
Reference:
Y. Song, K. Yoon, Y. Yoon, K. Yow, M. Jeon. Online Multi-Object Tracking Framework with the GMPHD Filter and Occlusion Group Management. In arXiv:1907.13347, 2019.
Benchmark performance:
| MOTA | MOTP | FAF | MT | ML | FP | FN | ID Sw. | Frag | Specifications | Detector |
| 49.9 | 77.0 | 1.4 | 19.7 % | 38.0 % | 24,024 | 255,277 | 3,125 | 3,540 | Intel i7-7700K @ 4.2GHz, 4 Cores, no GPU | Public |
| IDF1 | ID Precision | ID Recall |
| 47.1 | 63.5 | 37.5 |
Detailed performance:
| Sequence | MOTA | IDF1 | MOTP | FAF | GT | MT | ML | FP | FN | ID Sw | Frag |
| MOT17-01-DPM | 28.0 | 41.7 | 72.9 | 0.2 | 24 | 8.3 % | 54.2 % | 74 | 4,555 | 13 | 25 |
| MOT17-03-DPM | 49.8 | 43.7 | 75.5 | 2.8 | 148 | 16.9 % | 22.3 % | 4,261 | 47,999 | 278 | 304 |
| MOT17-06-DPM | 46.0 | 51.3 | 74.0 | 0.2 | 222 | 12.6 % | 45.9 % | 296 | 5,919 | 151 | 185 |
| MOT17-07-DPM | 36.0 | 34.4 | 74.9 | 1.0 | 60 | 8.3 % | 46.7 % | 501 | 10,132 | 171 | 186 |
| MOT17-08-DPM | 28.1 | 36.0 | 79.8 | 0.9 | 76 | 14.5 % | 52.6 % | 534 | 14,596 | 56 | 62 |
| MOT17-12-DPM | 38.6 | 50.4 | 78.3 | 0.3 | 91 | 12.1 % | 57.1 % | 290 | 4,990 | 39 | 39 |
| MOT17-14-DPM | 21.9 | 31.8 | 74.9 | 0.7 | 164 | 3.7 % | 61.6 % | 533 | 13,760 | 141 | 142 |
| MOT17-01-FRCNN | 35.5 | 43.5 | 76.5 | 2.6 | 24 | 29.2 % | 33.3 % | 1,148 | 2,991 | 19 | 27 |
| MOT17-03-FRCNN | 60.0 | 52.2 | 77.1 | 2.5 | 148 | 30.4 % | 16.9 % | 3,800 | 37,869 | 220 | 324 |
| MOT17-06-FRCNN | 51.5 | 56.0 | 77.8 | 1.2 | 222 | 31.1 % | 20.7 % | 1,478 | 4,016 | 227 | 228 |
| MOT17-07-FRCNN | 33.5 | 40.3 | 76.5 | 1.4 | 60 | 6.7 % | 33.3 % | 688 | 10,441 | 107 | 168 |
| MOT17-08-FRCNN | 25.7 | 35.6 | 79.5 | 1.3 | 76 | 13.2 % | 52.6 % | 825 | 14,817 | 48 | 70 |
| MOT17-12-FRCNN | 35.9 | 51.1 | 79.5 | 0.5 | 91 | 11.0 % | 54.9 % | 423 | 5,108 | 26 | 32 |
| MOT17-14-FRCNN | 22.8 | 32.1 | 72.8 | 1.7 | 164 | 4.9 % | 51.2 % | 1,244 | 12,830 | 196 | 270 |
| MOT17-01-SDP | 44.5 | 41.2 | 74.8 | 1.7 | 24 | 29.2 % | 20.8 % | 780 | 2,756 | 43 | 54 |
| MOT17-03-SDP | 76.5 | 55.8 | 78.0 | 2.2 | 148 | 57.4 % | 8.1 % | 3,366 | 20,760 | 475 | 539 |
| MOT17-06-SDP | 56.3 | 57.3 | 77.1 | 0.8 | 222 | 35.1 % | 29.3 % | 942 | 4,030 | 177 | 165 |
| MOT17-07-SDP | 46.9 | 44.1 | 77.6 | 1.6 | 60 | 18.3 % | 28.3 % | 796 | 8,017 | 160 | 177 |
| MOT17-08-SDP | 33.7 | 41.7 | 79.7 | 0.9 | 76 | 22.4 % | 43.4 % | 543 | 13,354 | 100 | 113 |
| MOT17-12-SDP | 43.1 | 54.3 | 79.7 | 0.7 | 91 | 19.8 % | 41.8 % | 666 | 4,218 | 44 | 44 |
| MOT17-14-SDP | 27.6 | 32.5 | 75.1 | 1.1 | 164 | 4.3 % | 50.6 % | 836 | 12,119 | 434 | 386 |
Raw data:
n/a
