Color textured image segmentation wavelet based histogram analysis by thresholding and clustering with performance measures

Abstract

In the field of Computer vision texture analysis has historically been an important topic of study Texture is one of the major primary visual cues to image understanding The human visual system relies on texture among other types of cues, to effectively interpret the information contained in an image Therefore it is imperative that a fully functional artificial vision newlinesystem must perform texture analysis This science is better developed in the case of gray scale image analysis as opposed to color texture analysis It is only recently that attempts are being made to combine both aspects in a single method by extending gray level texture analysis methods to color images This combination can be made more formal by defining color texture as the set of local statistical properties of the colors of image regions Often in image analysis and interpretation tasks including multimedia applications and visual inspection color textured images must be segmented for recognition purposes newline