Project DEMONSTRATE was featured in the paper “Lacunarity Analysis of Microvascular Morphology in Human Retina” by I. Konatar, N. Popovic, T. Popovic, M. Radunovic, and B. Vukcevic was presented at the 2020 IcETRAN conference today. More about the conference at the following link.
ABSTRACT – Fractal analysis provides means for the quantitative assessment of geometric patterns in one, two, and three dimensions. It is aimed at analysis of graphical shapes that belong to a class of fractal objects that are characterized by the self-similarity over different scales. Various structures in nature are fractals and fractal analysis techniques are widely used for analysis of biomedical images. One such example of application is analyzing blood vessel structure in the human retina that can be extracted from digital images captured by fundus camera. The most commonly used fractal analysis is estimation of fractal dimension using various box-counting methods for mono-and multi-fractals. Although two fractal images can have the same fractal dimension they can have very different appearance and structure. One can appear as a structure that fills most of the available space, while the other can have a lot of empty areas. These differences can be quantified by lacunarity parameter, which has greater value in images with less space-filling properties. This paper focuses on the estimation of the lacunarity parameter implemented in the Python programming language, which is aimed at lacunarity analysis of microvaculae morphology in human retina. The implementation is validated by comparison with the results obtained by ImageJ, a commonly used software for analysis of biomedical images. The value of the lacunarity analysis is demonstrated on a set of actual images of human retina associated with different medical conditions.