HEURISTICALLY IMPROVED BAYESIAN SEGMENTATION OF BRAIN MR IMAGES

Abstract

One of the major tasks or even the most prevalent task in medical image
processing is image segmentation.  Among them, brain MR images suffer
from  some  difficulties  such  as  intensity  inhomogeneity  of  tissues,  partial
volume  effect,  noise  and  some  other  imaging  artifacts  and  so  their
segmentation  is  more  challenging.  Therefore,  brain  MRI  segmentation
based  on  just  gray  values  is  prone  to  error.  Hence  involving  problem
specific  heuristics  and  expert  knowledge  in  designing  segmentation
algorithms seems to be useful. A two-phase segmentation algorithm based
on Bayesian method is proposed in this paper. The Bayesian part uses the
gray value in segmenting images and the segmented image is used as the
input to  the second phase to improve the misclassified pixels especially in
borders between tissues. Similarity index is used to compare our algorithm
with the well known method of Ashburner which has been implemented in
Statistical  Parametric  Mapping  (SPM)  package.  Brainweb  as  a  simulated
brain  MRI  dataset  is  used  in  evaluating  the  proposed  algorithm.  Results
show  that  our  algorithm  performs  well  in  comparison  with  the  one
implemented  in  SPM.  It  can  be  concluded  that  incorporating  expert
knowledge  and  problem  specific  heuristics  improve  segmentation  result.
The  major  advantage  of  proposed  method  is  that  one  can  update  the
knowledge  base  and  incorporate  new  information  into  segmentation
process by adding new rules.