Fractal Dimension Based on Pixel Covering Method

Abstract

Fractal dimension is an important feature of images, which is considered as a basic concept in fractal geometry used to measure the geometrical complexity of fractal set. In fractal geometry theory, the fundamental definition of fractal dimension have been based on Hausdorff dimension that is not easy to be estimated in most cases. There are many approaches to estimate the fractal dimension of an object, they compute inefficiently and the present of the local features of image invalidly. This paper addresses this problem by presenting a new estimated algorithm based on pixel covering method. The proposed approach will serve as an important characteristic for several applications in medical, engineering, and sciences, it helps to determine the local structure feature of image upon other conventional approaches used to determine the fractal dimension for the whole image. Experimental investigations indicate the efficiency of this approach compared with a well known widely used approaches such as; the box counting dimension, and the escape time dimension.