New measures of homogeneity for image processing: an application to fingerprint segmentation.

In this work, we axiomatically describe the concepts of total homogeneity and local homogeneity of a fuzzy set. We also present different construction methods by means of restricted equivalence functions, restricted dissimilarity functions and implication operators. We show that the most common homogeneity expressions can be obtained as particular cases of the constructions developed here. The usefulness of these functions is shown in segmentation of fingerprint images. [ABSTRACT FROM AUTHOR]