Official implementation of the CVPR2022 paper "Learning of Global Objective for Network Flow in Multi-Object Tracking"
# Install cvxpy and qpth
pip install cvxpy
pip install qpth
# Install pytorch_geometric(Version < 2.0.0) and gurobipy, please check if gurobipy is correctly installed
import gurobipy as gp
gurobi_solver = gp.Model()
# Install torchreid from: https://github.com/KaiyangZhou/deep-person-reid, and put it inside ./lib folder.
Download pre-processed detections&appearance features (~1.4GB): [Google Drive] We aslo provide the tracking results on MOT17/20 test set in txt format, which can replicate the numbers reported in the paper.
Execute run_train.ipynb. Monitor all relevant metrics, using our provided data, the loss converges at around 8 epochs. Training with more data might be helpful, but not tested.
Execute run_test.ipynb, you need to adjust the MOT17/20 data path accordingly. Note that for one specific MOT20 sequence, running time could be long.
If you have any questions using this code, please open an issue. I'll respond ASAP.
If you find this code useful in your research, please consider citing:
@inproceedings{li2022learning,
title={Learning of Global Objective for Network Flow in Multi-Object Tracking},
author={Li, Shuai and Kong, Yu and Rezatofighi, Hamid},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
pages={8855--8865},
year={2022}
}