/fucolot

Matlab implementation of FuCoLoT tracker

Primary LanguageC++

FuCoLoT: A Fully-Correlational Long-Term Tracker

Matlab implementation of Fully Correlational Long-Term Tracker published at the 14th Asian Conference on Computer Vision (ACCV) 2018.

Publication

Alan Lukežič, Luka Čehovin Zajc, Tomáš Vojíř, Jiří Matas and Matej Kristan. ''FuCoLoT - A Fully-Correlational Long-Term Tracker.'' Asian Conference on Computer Vision (ACCV), 2018.
Paper

BibTex citation:
@InProceedings{Lukezic_ACCV_2018,
Title = {FuCoLoT - A Fully-Correlational Long-Term Tracker},
Author = {Luke{\v{z}}i{\v{c}}, Alan and {\v{C}}ehovin Zajc, Luka and Voj{'i}{\v{r}}, Tom{'a}{\v{s}} and Matas, Ji{\v{r}}{'i} and Kristan, Matej},
Booktitle = {ACCV},
Year = {2018}
}

Source code running

  • Clone git repository:
    $ git clone https://github.com/alanlukezic/fucolot.git
  • Compile mex files running compile.m command in the CSRDCF folder
    Set opencv_include and opencv_libpath to the correct OpenCV paths in the compile.m script
  • Use run_on_uav.m script for running the tracker on UAV dataset (https://ivul.kaust.edu.sa/Pages/Dataset-UAV123.aspx)
    Set tracker_path variable to the directory where your source code is and dataset_path to the directory where you have stored the UAV dataset. Set results_path variable to the directory where you want to store the raw results. Modify parfor into for command to disable parallel processing of the sequences.

Tracking results

UAV20L benchmark (long-term): Results
OTB100 benchmark: Results
VOT-2016 benchmark (unsupervised experiment): Results