Deep similarity trackers are able to track above realtime speed. However, their accuracy is considerably lower than deep classification based trackers since they avoid valuable online cues. To feed the target-specific information for real-time object tracking, we propose a novel Siamese attention network. Different types of attention mechanisms are used to capture different contexts of target information and then learned knowledge is used to feed target cues at different representation levels of similarity tracking. In addition, an online learning mechanism is employed to utilise the available target-specific data. The proposed tracker reduces the impact of noise in the target template and improves the accuracy of similarity tracking by feeding target cues into the similarity search. Extensive evaluation performed on OTB- 2013/50/100 and VOT2018 benchmark datasets demonstrate the proposed tracker outperforms state-of-the-art approaches while maintaining real-time tracking speed.

Paper

Please cite the above publication if you use the code or compare with the SiamAtt tracker in your work. Bibtex entry:

@ARTICLE{8804242,
author={K. {Thanikasalam} and C. {Fookes} and S. {Sridharan} and A. {Ramanan} and A. {Pinidiyaarachchi}},
journal={IEEE Transactions on Information Forensics and Security},
title={Target-Specific Siamese Attention Network for Real-time Object Tracking},
year={2019},
volume={},
number={},
pages={1-1},
keywords={Visual object tracking;Deep neural networks;Siamese network;Attention network},
doi={10.1109/TIFS.2019.2935871},
ISSN={},
month={},}

If you have any other question, you can contact me by email: [email protected], [email protected]