Magnetic Resonance Image Segmentation Technique by using Hurst Coefficients

S. Ezekiel and M. Schuette (USA)


MRI, Hurst Exponent, Fractal Dimension, Rescaled Range Analysis, T1-Controlled Image


Segmentation of Magnetic Resonance Imaging (MRI) images into tissue types, e.g. gray matter (GM), white matter (WM) or cerebrospinal fluid (CSF), is traditionally based on selecting histogram modes to represent “pure” tissue signal values or manual image segmentation. In clinical situations where large numbers of data sets must be segmented, manual segmentation results are not easily reproducible. Also, the histogram modes may be biased by partial voluming and there may be no peaks for some pure tissue values. Automatic image segmentation can eliminate the problems that manual segmentation presents while also expediting the process. In this paper, we present a technique to reliably segment MRI images by using rescaled range analysis. The results demonstrate the power of this technique and its capability to analyze a broad range of images. The discussed method segments images quickly and without human error. Also, this method can be implemented for virtual surgical environments.

Important Links:

Go Back

Rotating Call For Paper Image