Authors:
Manuel González-Hidalgo
;
Sebastià Massanet
;
Arnau Mir
and
Daniel Ruiz-Aguilera
Affiliation:
University of the Balearic Islands, Spain
Keyword(s):
Edge Detection, Fuzzy Mathematical Morphology, Uninorms, Fuzzy Implications, Hysteresis.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Computational Intelligence
;
Fuzzy Image, Speech and Signal Processing, Vision and Multimedia
;
Fuzzy Systems
;
Soft Computing
Abstract:
In this paper, we study the performance of the edge detector from the fuzzy mathematical morphology based on conjunctive uninorms. Several different pairs of uninorm and fuzzy implication (configurations) are considered in the fuzzy morphological gradient. The results are compared using an objective edge detection performance measure, the so-called Pratt’s figure of merit. To reinforce the analysis a K-means clustering algorithm has been applied to study the relation between the configurations and to determine which uninorm and implication have to be chosen to obtain an optimal edge detector. According to the analysis of the obtained results, the idempotent uninorm obtained using the classical negation, and its residual implication is the best configuration in this framework.