An improved differential box-counting approach to compute fractal dimension of gray-level image

The box counting dimension is widely used for the image processing task. In this paper, The key problems involved in the computation of differential box counting are discussed, which are the range of linear scales and the precise number of boxes, etc, then the precise computing method for box counting dimension is presented. Experiments have been performed on three kinds of images, i.e., synthetic noise images, natural textured images and infrared smoke screen jamming images. The results show that, compared with the traditional differential box counting dimension, the proposed approach not only has more precise estimated value of fractal dimension, but also consumes less computational time.