Most of the existing box-counting methods for measuring fractal features are only applicable to square images or images with each dimension equal to the power of 2 and require that the box at the top of the box stack of each image block is of the same height as that of other boxes in the same stack, which gives rise to inaccurate estimation of fractal dimension. In this paper, we propose a more accurate box-counting method for images of arbitrary size, which allows the height of the box at the top of each grid block to be adaptable to the maximum and minimum gray-scales of that block so as to circumvent the common limitations of existing box-counting methods.
展开▼
