Brain mri segmentation based on the rényi's fractal dimension

Abstract

This paper presents a novel technique to perform segmentation on brain Magnetic Resonance Imaging (MRI). In this work, the multifractal behavior of the different textures present in brain MRI images was characterized. This process was done by calculating the Rényi's entropy function using different mask sizes. The average Rényi's fractal dimension was then determined and used to separate the gray and white matter regions in the image. The proposed method yielded promising results when applied to a real MRI data set. © 2009 Springer Berlin Heidelberg.