On the performance of connected components grouping

Grouping processes may benefit computationally when simple algorithms are u sed as part of the grouping process. In this paper we consider a common and extremely fast grouping process based on the connected components algorith m. Relying on a probabilistic model, we focus on analyzing the algorithm's performance. In particular, we derive the expected number of addition error s and the group fragmentation rate. We show that these performance figures depend on a few inherent and intuitive parameters. Furthermore, we show tha t it is possible to control the grouping process so that the performance ma y be chosen within the bounds of a given tradeoff. The analytic results are supported by implementing the algorithm and testing it on synthetic and re al images.