ABSTRACT

Texture is an important spatial feature useful for identifying objects or regions of interest in an image. This paper presents a novel texture characterization method based on Fuzzy Texture Unit (FTU) for texture synthesis. The proposed fuzzy texture characterization approach, takes into account the vagueness introduced by noise and the different caption and digitization processes, for defining the texture unit, by which texture synthesis is obtained. FTU extracts local texture information from the pixels for characterizing the textural aspect of a digital image and FTU detects the classes having highest relevance for deciding the existence of a concrete textural feature within an image by which patches can be identified easily. A patch that satisfies the minimum error criterion computed using L2norm is selected as the candidate patch. Edging two patches along the minimum error boundary cut is considered to be one of the most efficient candidate patches till recently. However, in real textures, especially the semi-stochastic ones, perfect cuts do not exist; therefore, the defects due to the inconsistency between the patches become an important problem for the synthesized texture quality. For this reason the present paper adopts a modified MEBC determine the boundary between the patches. Experimental results on wide range of texture samples demonstrate that the performance of the present approach is superior to many other existing methods of texture synthesis.