Fractal analysis as a method of quantitative assessment of medical images heterogeneity

Abstract

The algorithm of quantitative assessment of structural heterogeneity of medical images based on fractal analysis was developed. Digital magnetic resonance images of brain were used to develop the algorithm for the brightness heterogeneity assessment of achromatic grayscale medical images. The analysis of the quantitative distribution of the number of the image pixels by brightness values (from 0 to 255) was performed. The graph curve of the distribution of the number of pixels by brightness levels was considered as a linear fractal and the fractal dimension of this curve was quantified. Image heterogeneity can be quantified using a fractal index, which values may vary from 1 to 2. This index allows to assess the homogeneity or heterogeneity of transitions between adjacent values of the pixel brightness of digital image. The developed algorithm for determining of the image heterogeneity can be used to interpret the data of various diagnostic methods involving the visualization of the object (ultrasound, radiography, various types of tomography) to assess the morphofunctional state of various structures and organs. Keywords: heterogeneity, brightness, fractal analysis, magnetic-resonance imaging, brain, cerebellum.