3.1. Image thresholding
The study of the Islamic Geometric Patterns images shows that they AZD0530
By the segmentation procedure, our objective is not to detect a particular type of furmahs. The goal is to realize a segmentation that preserves the spatial arrangement of furmahs having the same shape and the same color. For this purpose, we propose to use a global thresholding technique. This choice is justified by the fact that colors are used by craftsmen in order to emphasize certain aspects or certain geometrical structures. Often they influence the way in which a composition is perceived by using bright and dark colors. It is obvious to see that a geometrical structure such as the geometric rosette is composed by the both bright and dark furmahs.
Several methods of image thresholding are proposed in the literature: parametric and non parametric methods. In parametric methods , a statistical model is first assumed to fit the gray level
distribution of an image, and a set of parameters that control the fitness of the model are found using a histogram. In non parametric methods, thresholds are chosen by optimizing an objective function, such as maximizing interclass variance , or minimizing entropy