Generalizing an interval-valued image magnification algorithm using homogeneity measures and interval fusion functions
Daniel Paternain, Aránzazu Jurío, Miguel Pagola, Edurne Barrenechea, Humberto Bustince · 2017
In this work we study and generalize an image magnification algorithm based on the use of interval-valued fuzzy sets. The first proposed generalization incorporates an homogeneity measure that allows to model the length of the intervals generated by the algorithm. The second one makes use of several homogeneity measures and, by means of a fusion function, it combines the intervals generated by each individual homogeneity measure. The results show that our generalization outperforms the original algorithm when an appropriate homogeneity measure is used. Moreover, experiments have demonstrated that the second generalization, based on interval fusion functions, avoids low quality results due to bad homogeneity measures.