Application of Mep-Based Fuzzy Clustering to the Segmentation of Multivariate Medical Images

摘要

In this article, we present some results on the application of fuzzy methods to the segmentation of multivariate medical images. We report the results obtained by using the method by K.Rose, E.Gurewitz and G.Fox (RGF) based on the Maximum Entropy Principle (MEP). In particular we study the effect of using, for each epoch of the algorithm, a Reduced Data Base (RDB) obtained through an uniform random sampling of the original data base. As shown by our results, this approach reduces the execution times of the clustering algorithm, maintaining in the same time a high segmentation reliability of multivariate images.