An adaptive fuzzy approach for texture modelling

In this paper, a method to adapt the fuzzy sets that models the ”coarseness” texture property according to different coarse-fine perceptions is proposed. The membership functions associated to these fuzzy sets are obtained on the basis of the human assessments given by a pool, that represent the average coarse-fine perception degree. In the proposed method, the user´s particular perception about coarseness-fineness and the changes in perception influenced by the image context have been taken into account to adapt the models. The membership functions are automatically adapted by means of a transformation of functions on the basis of the new coarse-fine perception.