@Article{kalal-pami2012,
  IS = { zkontrolovano 14 Jan 2013 },
  UPDATE  = { 2013-01-14 },
  author    = { Kalal, Zdenek and
                Mikolajczyk, Krystian and
                Matas, Ji{\v r}{\'\i} },
  title =     { Tracking-Learning-Detection },
  journal =   { {IEEE} Transactions 
                on Pattern Analysis and Machine Intelligence },
  volume =    { 34 },
  number =    { 7 },
  year =      { 2012 },
  month =     { July },
  publisher = { IEEE Computer Society },
  address =   { Ney York, USA },
  issn =      { 0162-8828 },
  pages     = { 1409-1422 },
  authorship ={ 34-33-33 },
  project   = { GACR P103/10/1585, FP7-ICT-270138 DARWIN },
  doi       = { http://doi.ieeecomputersociety.org/10.1109/TPAMI.2011.239 },
  keywords  = { tracking, long-term tracking },
  psurl     = { http://cmp.felk.cvut.cz/~matas/papers/kalal-2010-tld-pami.pdf },
  annote = { This paper investigates long-term tracking of unknown
    objects in a video stream. The object is defined by its location
    and extent in a single frame. In every frame that follows, the
    task is to determine the object?s location and extent or indicate
    that the object is not present. We propose a novel tracking
    framework (TLD) that explicitly decomposes the long-term tracking
    task into tracking, learning, and detection. The tracker follows
    the object from frame to frame. The detector localizes all
    appearances that have been observed so far and corrects the
    tracker if necessary. The learning estimates the detector?s errors
    and updates it to avoid these errors in the future. We study how
    to identify the detector?s errors and learn from them. We develop
    a novel learning method (P-N learning) which estimates the errors
    by a pair of ?experts?: 1) P-expert estimates missed detections,
    and 2) N-expert estimates false alarms. The learning process is
    modeled as a discrete dynamical system and the conditions under
    which the learning guarantees improvement are found. We describe
    our real-time implementation of the TLD framework and the P-N
    learning. We carry out an extensive quantitative evaluation which
    shows a significant improvement over state-of-the-art approaches.},
  ut_isi    = { 000304138300012 },
}