SiamAtt

This is an official implemention for “SiamAtt: Siamese attention network for visual tracking”.

Dependencies

Python 3.7
PyTorch 1.0.0
numpy
CUDA 10
skimage
matplotlib

Prepare training dataset

Prepare training dataset, detailed preparations are listed in training_dataset directory.

VID
YOUTUBEBB (BaiduYun, extract code: h964.)
DET
COCO

Training:

CUDA_VISIBLE_DEVICES=0,1
python -m torch.distributed.launch \
    --nproc_per_node=2 \
    --master_port=2333 \
    ../../tools/train.py --cfg config.yaml

Testing:

python ../tools/test.py

References

[1]SiamRPN++: Evolution of Siamese Visual Tracking with Very Deep Networks. Bo Li, Wei Wu, Qiang Wang, Fangyi Zhang, Junliang Xing, Junjie Yan. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2019.

Acknowledgment

Our SiamAtt tracker is based on PySot. We sincerely thank the authors Bo Li for providing these great works.

Citation

If you're using this code in a publication, please cite our paper.

@article{yang2020siamatt,
  title={SiamAtt: Siamese attention network for visual tracking},
  author={Yang, Kai and He, Zhenyu and Zhou, Zikun and Fan, Nana},
  journal={Knowledge-based systems},
  volume={203},
  pages={106079},
  year={2020},
  publisher={Elsevier}
}