This is the implementation of our VITAL paper. The project page can be found here: https://ybsong00.github.io/cvpr18_tracking/index.html

The pipeline is built upon the MDNet tracker for your reference: http://cvlab.postech.ac.kr/research/mdnet/

Try 'tracking/demo_tracking.m' to see the tracker performance on the Bolt sequences.

A pytorch implemenation is provided here: https://github.com/abnerwang/py-Vital. Thanks to David Wang.

If you find the code useful, please cite both VITAL and MDNet:

@inproceedings{nam-cvpr16-MDNET,
    author    = {Nam, Hyeonseob and Han, Bohyung}, 
    title     = {Learning multi-domain convolutional neural networks for visual tracking}, 
    booktitle = {IEEE Conference on Computer Vision and Pattern Recognition},    
    pages     = {4293--4302},
    year      = {2016},
}

@inproceedings{song-cvpr18-VITAL,
    author    = {Song, Yibing and Ma, Chao and Wu, Xiaohe and Gong, Lijun and Bao, Linchao and Zuo, Wangmeng and Shen, Chunhua and Lau, Rynson and Yang, Ming-Hsuan}, 
    title     = {VITAL: VIsual Tracking via Adversarial Learning}, 
    booktitle = {IEEE Conference on Computer Vision and Pattern Recognition},    
    year      = {2018},
}