A new stopping criterion for the mean shift iterative algorithm
This work is incremental, as it modifies an existing stopping criterion for image segmentation algorithms.
The paper tackles the problem of improving the mean shift iterative algorithm by proposing a new stopping criterion based on entropy, which leads to better segmentation and stability in experiments with standard images.
The mean shift iterative algorithm was proposed in 2006, for using the entropy as a stopping criterion. From then on, a theoretical base has been developed and a group of applications has been carried out using this algorithm. This paper proposes a new stopping criterion for the mean shift iterative algorithm, where stopping threshold via entropy is used now, but in another way. Many segmentation experiments were carried out by utilizing standard images and it was verified that a better segmentation was reached, and that the algorithm had better stability. An analysis on the convergence, through a theorem, with the new stopping criterion was carried out. The goal of this paper is to compare the new stopping criterion with the old criterion. For this reason, the obtained results were not compared with other segmentation approaches, since with the old stopping criterion were previously carried out.