MCIT: Multi-level cross-modal interactive transformer for RGBT tracking

The paper was accepted by the Neurocomputing.

Citation

If our work is useful for your research, please consider citing:

@article{MCIT,
title = {MCIT: Multi-level cross-modal interactive transformer for RGBT tracking},
journal = {Neurocomputing},
volume = {649},
pages = {130758},
year = {2025},
issn = {0925-2312},
doi = {https://doi.org/10.1016/j.neucom.2025.130758},
url = {https://www.sciencedirect.com/science/article/pii/S0925231225014304},
author = {Yu Qin and Jianming Zhang and Shimeng Fan and Zikang Liu and Jin Wang},
}

Install the environment

Install virtual environment and dependency packages.

conda create -n MCIT python=3.7
conda activate MCIT
pip install -r requirements.txt

Create the default environment setting files.

# Environment settings for pytracking. Saved at pytracking/evaluation/local.py
python -c "from pytracking.evaluation.environment import create_default_local_file; create_default_local_file()"

# Environment settings for ltr. Saved at ltr/admin/local.py
python -c "from ltr.admin.environment import create_default_local_file; create_default_local_file()"

Then set the paths of the project and dataset in "ltr/admin/local.py" and "pytracking/evaluation/local.py".

Training

Set the training parameters in "ltr/train_settings/MCIT/MCIT_settings.py".

Then run:

python ltr/run_training.py

Testing

Set the model weight path in "pytracing/parameter/MCIT/MCIT.py".

Then run:

python pytracking/run_tracker.py --dataset_name rgbt234

Tracking results

Download the tracking results from Baidu Netdisk code: 87xu

Download the model weights from Baidu Netdisk code: 5m57

Acknowledgments

Thanks for the PyTracking and OSTrack library, which helps us to quickly implement our ideas.