@InProceedings{hurych-cvww2011,
  IS = { zkontrolovano 10 Jan 2012 },
  UPDATE  = { 2011-04-11 },
  author =      {Hurych, David and Zimmermann, Karel and Svoboda, Tom{\' a}{\v s}},
  title =       {Detection of unseen patches trackable by linear predictors},
  year =        {2011},
  pages =       {107-114},
  booktitle =   {CVWW '11: Proceedings of the 16th Computer Vision Winter Workshop},
  editor =      {Wendel, Andreas and Stering, Sabine and Godec, Martin},
  publisher =   {Graz University of Technology},
  address =     {Inffeldgasse 16/II, Graz, Austria},
  isbn =        {978-3-85125-129-6},
  book_pages =  {165},
  month =       {February},
  day =         {2-4},
  venue =       {Mitterberg, Austria},
  organization ={Graz University of Technology - Institute for 
                 Computer Graphics and Vision},
  annote =      {Linear predictors (LPs) are being used for tracking
    because of their computational efficiency which is better than
    steepest descent methods (e.g. Lucas-Kanade).  The only
    disadvantage of LPs is the necessary learning phase which hinders
    the predictors applicability as a general patch tracker.  We
    address this limitation and propose to learn a bank of LPs
    off-line and develop an on-line detector which selects image
    regions that could be tracked by some predictor from the bank.
    The proposed detector differs significantly from the usual
    solutions that attempt to find the closest match between a
    candidate patch and a database of exemplars. We construct the
    detector directly from the learned linear predictor.  The detector
    positively detects the learned patches, but also many other image
    patches, which were not used in LP learning phase. This means,
    that the LP is able to track also previously unseen image patches,
    the appearances of which are often significantly diverse from the
    patches used for learning. We propose a fast LP-structure-based
    detection method, which is in computational cost comparable with
    standard appearance-based detectors and is easy to construct
    directly from a trained LP without any further learning.},
  keywords =    {detection, tracking, linear predictors, learning},
  prestige =    {international},
  authorship =  {50-30-20},
  project =     {GACR P103/10/1585, GACR P103/11/P700, FP7-ICT-247870 NIFTi},
  url =         {ftp://cmp.felk.cvut.cz/pub/cmp/articles/hurycd1/cvww2011.pdf},
  psurl       = {[hurych-CVWW2011.pdf] },
}