An extended DBC approach by using maximum Euclidian distance for fractal dimension of color images

The Fractal Dimension (FD) of digital image represents the roughness in terms of real number with self similarity property in order to correlate with human perception of surface roughness. FD has broadly adopted in various kinds of applications in different field of computer graphics and image processing such as texture analysis, classification, segmentation and many more that establish in various literatures. For estimating FD there are several techniques were presented in gray scale domain, out of which differential box counting (DBC) is repeatedly used algorithm, but in case of color domain there are few and countable research has been done because the natural color images exhibits a non trivial and self-similar and scale invariance feature. This article presents a new color FD estimation algorithm by extending original DBC algorithm by implementing maximum color Euclidian distance from each non overlapping box block of RGB components. All the experimental work has been made on one set of standard brodatz color texture images and one set of known fractal dimension smooth color images used for showing feasibility of the proposed technique. The experimental result proved that the proposed algorithm efficiently captures the surface roughness of RGB color images. The computational time of this proposed method yields quit less than that of other existing algorithms. This is more reliable and precise method for color images.