On the Use of Fuzzy Texture Spectrum for Homogeneous and Textured Images Discrimination

Identifying the perceived qualities and properties of texture in images is an important step for image analysis, however the vague definition of texture has led to a huge variety of ways to analyze and characterize texture. In the ninetieths He and Wang (1990) proposed the use of the texture spectrum for extracting texture features. Recently we have developed a fuzzy-based texture spectrum coding approach, the fuzzy texture spectrum coding (Barcelo et al., 2006), in which a texture image could be described by the occurrence frequency function of all its fuzzy texture feature vectors. Later on, in the work of Barcelo et al. (2006), we proved the robustness of the proposed approach, particularly for representing homogeneity features. As an extension of our previous study on the analysis of the fuzzy texture spectrum coding for extracting texture features, in this paper we present the first results of our ongoing work for obtaining the classes encoded by the Reduced Fuzzy Texture Spectrum allowing to determine images´ texture characteristics as well as the degrees to which these classes determine image´s texture and homogeneity.