The code for TransCenter and TransCenter-Lite is now available, you can find the code and pretrained models at https://gitlab.inria.fr/robotlearn/TransCenter_official.
TransCenter: Transformers with Dense Representations for Multiple-Object Tracking
Yihong Xu, Yutong Ban, Guillaume Delorme, Chuang Gan, Daniela Rus, Xavier Alameda-Pineda
[Paper] [Project]
MOT20 example:
If you find this code useful, please star the project and consider citing:
@misc{xu2021transcenter,
title={TransCenter: Transformers with Dense Representations for Multiple-Object Tracking},
author={Yihong Xu and Yutong Ban and Guillaume Delorme and Chuang Gan and Daniela Rus and Xavier Alameda-Pineda},
year={2021},
eprint={2103.15145},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
For TransCenter:
MOT17 public detections:
| Pretrained | MOTA | MOTP | IDF1 | FP | FN | IDS |
|---|---|---|---|---|---|---|
| CoCo | 71.9% | 80.5% | 64.1% | 27,356 | 126,860 | 4,118 |
| CH | 75.9% | 81.2% | 65.9% | 30,190 | 100,999 | 4,626 |
MOT20 public detections:
| Pretrained | MOTA | MOTP | IDF1 | FP | FN | IDS |
|---|---|---|---|---|---|---|
| CoCo | 67.7% | 79.8% | 58.9% | 54,967 | 108,376 | 3,707 |
| CH | 72.8% | 81.0% | 57.6% | 28,026 | 110,312 | 2,621 |
MOT17 private detections:
| Pretrained | MOTA | MOTP | IDF1 | FP | FN | IDS |
|---|---|---|---|---|---|---|
| CoCo | 72.7% | 80.3% | 64.0% | 33,807 | 115,542 | 4,719 |
| CH | 76.2% | 81.1% | 65.5% | 40,101 | 88,827 | 5,394 |
MOT20 private detections:
| Pretrained | MOTA | MOTP | IDF1 | FP | FN | IDS |
|---|---|---|---|---|---|---|
| CoCo | 67.7% | 79.8% | 58.7% | 56,435 | 107,163 | 3,759 |
| CH | 72.9% | 81.0% | 57.7% | 28,596 | 108,982 | 2,625 |
Note:
- The results can be slightly different depending on the running environment.
- We might keep updating the results in the near future.
The code for TransCenterV2, TransCenter-Lite is modified and network pre-trained weights are obtained from the following repositories:
- The PVTv2 backbone pretrained models from PVTv2.
- The data format conversion code is modified from CenterTrack.
CenterTrack, Deformable-DETR, Tracktor.
@article{zhou2020tracking,
title={Tracking Objects as Points},
author={Zhou, Xingyi and Koltun, Vladlen and Kr{\"a}henb{\"u}hl, Philipp},
journal={ECCV},
year={2020}
}
@InProceedings{tracktor_2019_ICCV,
author = {Bergmann, Philipp and Meinhardt, Tim and Leal{-}Taix{\'{e}}, Laura},
title = {Tracking Without Bells and Whistles},
booktitle = {The IEEE International Conference on Computer Vision (ICCV)},
month = {October},
year = {2019}}
@article{zhu2020deformable,
title={Deformable DETR: Deformable Transformers for End-to-End Object Detection},
author={Zhu, Xizhou and Su, Weijie and Lu, Lewei and Li, Bin and Wang, Xiaogang and Dai, Jifeng},
journal={arXiv preprint arXiv:2010.04159},
year={2020}
}
@article{zhang2021bytetrack,
title={ByteTrack: Multi-Object Tracking by Associating Every Detection Box},
author={Zhang, Yifu and Sun, Peize and Jiang, Yi and Yu, Dongdong and Yuan, Zehuan and Luo, Ping and Liu, Wenyu and Wang, Xinggang},
journal={arXiv preprint arXiv:2110.06864},
year={2021}
}
@article{wang2021pvtv2,
title={Pvtv2: Improved baselines with pyramid vision transformer},
author={Wang, Wenhai and Xie, Enze and Li, Xiang and Fan, Deng-Ping and Song, Kaitao and Liang, Ding and Lu, Tong and Luo, Ping and Shao, Ling},
journal={Computational Visual Media},
volume={8},
number={3},
pages={1--10},
year={2022},
publisher={Springer}
}
Several modules are from:
MOT Metrics in Python: py-motmetrics
Soft-NMS: Soft-NMS
DETR: DETR
DCNv2: DCNv2
PVTv2: PVTv2
ByteTrack: ByteTrack
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent