@techreport{TR-IC-14-04,
   number = {IC-14-04},
   author = {Fábio {Dias} and Neucimar Jerônimo {Leite}},
   title    =    {{Accurate    Tridimensional    Reconstruction   with  
                   Unsynchronized    Cameras    Regardless   of   Time  
                   Information}},
   month = {February},
   year = {2014},
   institution = {Institute of Computing, University of Campinas},
   note = {In English, 10 pages.
    \par\selectlanguage{english}\textbf{Abstract}
       In  this  work,  we  approach the problem of the tridimensional
       reconstruction   of   the   trajectory   of   an  object  using 
       unsynchronized  cameras. While most methods from the literature
       try   to   measure,  as  accurately  as  possible,  the  timing 
       difference between the cameras, we chose a very different path:
       we  ignore  most,  or  all,  of  the  time information from the
       videos.  The  idea  is  quite  simple.  Each pair of camera and
       projected  trajectory  generate  a  surface  on the volume. The
       intersection  of these curves is the desired result. Since this
       intersection  can  be very complex, we considered a Monte Carlo
       approach  for  the  reconstruction, using random tridimensional
       points  to  estimate the region of intersection. Therefore, any
       camera  calibration  schema can be used. These points are later
       used  to  compute a continuous curve, which is the final result
       of  the  method.  We  compared  this  method  to  a very simple
       reconstruction   approach,   which   assumes   the  frames  are 
       synchronized,    and    obtained   outstanding   results.   Our  
       implementation     and     data     are     freely    available   
       on~\url{https://code.google.com/p/ucr-timeless/}.
  }
}