Box-Counting Method in Python for Fractal Analysis of Biomedical Images

Abstract

This paper presents the implementation of a Python-based library with a purpose to determine fractal dimension of biomedical images. The described method is based on the assumption that the images are already pre-processed and contain binarized version of fractal-like structures that can often be found in biomedical images. Three variants of the box-counting method were implemented using different ways for selecting and sampling the boxes: standard non-overlapping box scanning, gliding or overlapping box scanning, and random box sampling. The utility of the proposed software was validated through the analysis of an open access library of binarized images of retinal microvasculature and by the comparison of these results with those obtained by using ImageJ program, that is commonly used for this purpose.