This is our standard clone killer algorithm which is run at the end of the pattern recognition in every standard reconstruction job. It works by comparing hit contents of tracks. If they share more than a certain percentage of hits, two tracks are considered clones. The highest quality one is written to the output container (most hits, in case of same number of hits, lower chi^2 wins). Details can be found in:
--
ManuelSchiller - 15 Sep 2008