Automatic determination of MS lesion subtypes based on fractal analysis in brain MR images英文文献资料.docVIP

J. Biomedical Science and Engineering, 2012, 5, 162-169 JBiSE /10.4236/jbise.2012.54021 Published Online April 2012 (http://www.SciRP.org/journal/jbise/) Automatic determination of MS lesion subtypes based on fractal analysis in brain MR images Mahdi Mohamadkhanloo, Farzad Mehrabi, Abdolhamid Sohrabi Department of Medical Science, Aja University, Tehran, Iran Email: mahdi.khanlo@, df_mehrabi2002@, sohrabind@ Received 29 November 2011; revised 25 December 2011; accepted 29 January 2012 ABSTRACT of MS is MRI. Recently, the study of MRI in MS has been flourishing in the field of research due to its ability in generating multiple images of the same tissue with different contrast mechanisms using different image ac- quisition protocols and parameters. In its application to MS, macroscopic areas of damage or loss of myelin with hyper or hypo intensities relative to the surrounding tis- sues are shown by MRI. Different MRI protocols in- cluding T1-weighted (T1-w), T2-weighted (T2-w), PD- weighted (PD-w), and fluid-attenuated inversion reco- very (FLAIR) T2 images are being investigated for im- proving the performance in the early inflammatory stage and in the advanced stage of the diseases [1]. Figure 1 shows a slice of a multi-parametric MRI of two MS pa- tients. The lesions appear as regions with increased sig- nal intensity on the T2-w and FLAIR images and de- creased signal intensity (hypointense) on the T1-w image. There are two major categories of lesions which can be identified with conventional MR imaging: acute lesions demonstrating blood-brain barrier leakage on contrast- enhanced MR imaging (enhancing lesions), chronic se- verely damaged lesions that are hypointense, so-called “black holes” on T1-w MR images [2]. Acute plaques appear with less signal changes in T1-w images due to more inflammation, edema, and little demyelination. The borders of acute plaqu