SiamCAR

1. Environment setup

This code has been tested on Ubuntu 16.04, Python 3.6, Pytorch 0.4.1/1.2.0, CUDA 9.0. Please install related libraries before running this code:

bash
pip install -r requirements.txt

2. Test

Download the pretrained model:
general_model code: lw7w
got10k_model code: p4zx
LaSOT_model code: 6wer
(The model in google Driver) and put them into

tools/snapshot

Download testing datasets and put them into

test_dataset

python test.py                                \
    --dataset UAV123                      \ # dataset_name
    --snapshot snapshot/general_model.pth  # tracker_name

The testing result will be saved in the

results/dataset_name/tracker_name

3. Train

Prepare training datasets

Download the datasets： * VID * YOUTUBEBB (code: v7s6) * DET * COCO * GOT-10K * LaSOT

Note:

train_dataset/dataset_name/readme.md

Download pretrained backbones

Download pretrained backbones from google driver or BaiduYun (code: 7n7d) and put them into

pretrained_models

Train a model

To train the SiamCAR model, run

train.py

cd tools
python train.py

4. Evaluation

We provide the tracking results (code: 8c7b) (results in google driver )of GOT10K, LaSOT, OTB and UAV. If you want to evaluate the tracker, please put those results into

results

python eval.py                            \
    --tracker_path ./results          \ # result path
    --dataset UAV123                  \ # dataset_name
    --tracker_prefix 'general_model'   # tracker_name

5. Acknowledgement

The code is implemented based on pysot. We would like to express our sincere thanks to the contributors.

6. Cite

If you use SiamCAR in your work please cite our paper:

@InProceedings{Guo2020CVPR,
author = {Guo, Dongyan and Wang, Jun and Cui, Ying and Wang, Zhenhua and Chen, Shengyong},
title = {SiamCAR: Siamese Fully Convolutional Classification and Regression for Visual Tracking},
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2020}
}