Skip to content

Latest commit

 

History

History
 
 

deepsort

README.md

Deep SORT

Introduction

@inproceedings{bewley2016simple,
  title={Simple online and realtime tracking},
  author={Bewley, Alex and Ge, Zongyuan and Ott, Lionel and Ramos, Fabio and Upcroft, Ben},
  booktitle={2016 IEEE International Conference on Image Processing (ICIP)},
  pages={3464--3468},
  year={2016},
  organization={IEEE}
}
@inproceedings{wojke2017simple,
  title={Simple online and realtime tracking with a deep association metric},
  author={Wojke, Nicolai and Bewley, Alex and Paulus, Dietrich},
  booktitle={2017 IEEE international conference on image processing (ICIP)},
  pages={3645--3649},
  year={2017},
  organization={IEEE}
}

Results and models on MOT17

We implement SORT and DeepSORT with independent detector and ReID models. To train a model by yourself, you need to train a detector following here and also train a ReID model. The configs in this folder are basiclly for inference.

Currently we do not support training ReID models. We directly use the ReID model from Tracktor. These missed features will be supported in the future.

Detector ReID Train Set Test Set Public Inf time (fps) MOTA IDF1 FP FN IDSw. Config Download
R50-FasterRCNN-FPN - half-train half-val Y 28.3 46.0 46.6 289 82451 4581 config detector
R50-FasterRCNN-FPN - half-train half-val N 18.6 62.0 57.8 15171 40437 5841 config detector
R50-FasterRCNN-FPN R50 half-train half-val Y 20.4 48.1 60.8 283 82445 1199 config detector reid
R50-FasterRCNN-FPN R50 half-train half-val N 13.8 63.8 69.6 15060 40326 3183 config detector reid
R50-FasterRCNN-FPN - train train Y 28.3 50.9 44.5 1108 153950 10522 config detector
R50-FasterRCNN-FPN - train train N 18.6 80.8 61.3 21537 29280 13947 config detector