4. Fractal analysis of the liver on CT

Abstract

Fractal theory gives us an effective method to describe quantitatively the complicated shape in nature. We applied the theory to analyze CT images. We utilized a scale conversion method to obtain fractal dimensions. Regions of interest were set on the right lobe of the liver on CT and the images were converted to digital data with pixels of 256x256 and densities of 6 bits. The scale was converted from 1x1 to 168x168 with 9 steps and the surface areas of the density profile were counted for each scale. The numbers of the surface areas and the scale were plotted against doubly logarithmic scale and a straight line was drawn by least square fits of data points. The fractal dimension was obtained from the slope of the line. As a result, the CT images of the liver showed fractality. Fractal dimensions of the normal and cirrhotic livers were compared.