ProContEXT

ProContEXT: Exploring Progressive Context Transformer for Tracking

ProContEXT achieves SOTA performance on multiple benchmarks.

Tracker	GOT-10K (AO)	TrackingNet (AUC)
ProContEXT	74.6	84.6

Quick Start

Installation

You can refer to OSTrack to install the whole environments and prepare the data.

Set project paths

Run the following command to set paths for this project

python tracking/create_default_local_file.py --workspace_dir . --data_dir ./data --save_dir ./output

Train

We use models offered by MAE as our pretrained models. Put it under directory:

${PROJECT_ROOT}
    -- pretrained_models
      | -- mae_pretrain_vit_base.pth

Run the following command to train the model:

python tracking/train.py --script procontext --config procontext_got10k --save_dir ./output --mode multiple --nproc_per_node 4 # GOT-10k model
python tracking/train.py --script procontext --config procontext --save_dir ./output --mode multiple --nproc_per_node 4 # TrackingNet model

Test

You can download the trained model and put them under directory path:

${PROJECT_ROOT}/output/checkpoints/train/procontext/procontext_got10k/ProContEXT_ep0100.pth.tar # GOT-10k model
${PROJECT_ROOT}/output/checkpoints/train/procontext/procontext/ProContEXT_ep0300.pth.tar # TrackingNet model

Run the following command to test the model:

python tracking/test.py procontext procontext_got10k --dataset got10k_test --threads 16 --num_gpus 4
python tracking/test.py procontext procontext --dataset trackingnet --threads 16 --num_gpus 4

Acknowledgment

Our implementation is mainly based on OSTrack, Stark, pytracking, and Timm. We gratefully thank the authors for their wonderful works.

Citation

Please cite the following paper if this repo helps your research:

@article{ProContEXT,
  author    = {Jin{-}Peng Lan and
               Zhi{-}Qi Cheng and
               Jun{-}Yan He and
               Chenyang Li and
               Bin Luo and
               Xu Bao and
               Wangmeng Xiang and
               Yifeng Geng and
               Xuansong Xie},
  title     = {ProContEXT: Exploring Progressive Context Transformer for Tracking},
  journal   = {CoRR},
  volume    = {abs/2210.15511},
  year      = {2022},
  url       = {https://doi.org/10.48550/arXiv.2210.15511},
  doi       = {10.48550/arXiv.2210.15511},
}

License

This repo is released under the MIT license. Please see the LICENSE file for more information.

MSathishkumar1990/ProContEXT