DSST
the implementation of DSST(Accurate scale estimation for robust visual tracking )
Install / Use
/learn @ml422524/DSSTREADME
This MATLAB code implements the Discriminative Scale Space Tracker (DSST) [1]. The implementation is built upon the code provided by [2]. The code provided by [3] is used for computing the HOG features.
The project website is: http://www.cvl.isy.liu.se/research/objrec/visualtracking/index.html
Installation: To be able to run the "mexResize" function, try to use either one of the included mex-files or compile one of your own. OpenCV is needed for this. The existing compile scripts "compilemex.m" or "compilemex_win.m" can be modified for this purpose.
Instructions:
- Run the "run_tracker.m" script in MATLAB.
- Choose sequence (only "dog1" is included).
Contact: Martin Danelljan martin.danelljan@liu.se
[1] Martin Danelljan, Gustav Häger, Fahad Shahbaz Khan and Michael Felsberg. "Accurate Scale Estimation for Robust Visual Tracking". Proceedings of the British Machine Vision Conference (BMVC), 2014.
[2] J. Henriques, R. Caseiro, P. Martins, and J. Batista. "Exploiting the circulant structure of tracking-by-detection with kernels." In ECCV, 2012.
[3] Piotr Dollár. "Piotr’s Image and Video Matlab Toolbox (PMT)." http://vision.ucsd.edu/~pdollar/toolbox/doc/index.html.
