fragment finding : fgfgmfndsci

1. Make the angle histogram for each hit using the population of other hits in 2 degree angular window with the center at the hit. Find the most probable hit( h1) as a track start point and the most probable angle(${\bf\theta}$) as a direction of the track.
2. From this hit( h1) and to this angle(${\bf\theta}$), start searching a hit( h2) in the nearby plane within the certain angular window.
3. From the 2nd hit( h2), repeat the process '2' and continue till no more hit is found.
4. If the track is long enough(at least track should pass 3 layers of SCIFI out of 20 layers), save the hit list as a track.
5. Repeat process '1', using the remaining hits.
6. If the angle to the beam direction $>$ 60 degree : nhit/length $>$ 0.6, or
   If the angle to the beam direction $<$ 60 degree : nhit/length $>$ 0.75,
   then accepted. For example,

   
           Scifi layer : 1 2 3 4 5 6 7 8 9 0 1 2 3 4 .....
             fragment1 :     o o x o x o x o o        ==> nhit/length(6/9) = 0.67
          or fragment2 :   o x x o x o o x o x o o    ==> nhit/length(7/12) = 0.58