Research Of Brain Image Segmentation Algorithm Based On Energy Minimization

Medical imaging technology has become one of the most important means of clinical diagnosis of brain disease.Medical image processing is difficult because of variable imaging parameters,overlapping intensities,noise,partial voluming,gradients,motion,echoes,blurred edges,normal anatomical variations and susceptibility artifacts.Multiple sclerosis(MS)is the most frequent,non-traumatic,neurological disease capable of causing disability in young adults.MS is a chronic,persistent inflammatory-demyelinating and degenerative disease of the central nervous system(CNS),characterized pathologically by areas of inflammation,demyelination,axonal loss,and gliosis scattered throughout the CNS,often causing motor,sensorial,vision,coordination,deambulation,and cognitive impairment.Furthermore,vascular diseases are one of the major sources of deaths in the world.Quantitative modeling of brain vasculature is important for diagnosis of vessel pathologies as well as for surgery treatment planning.1.In this chapter,we extend the multiplicative intrinsic component optimization(MICO)algorithm to multichannel MR image segmentation,with focus on segmentation of multiple sclerosis(MS)lesions for normal brain tissue segmentation and intensity inhomogeneity correction of a single channel MR image,which exhibits desirable advantages over other methods for MR image segmentation and intensity inhomogeneity correction in terms of segmentation accuracy and robustness.We assign different weights for different channels to control the impact of each channel.The weighted channels allow the enhancement of the impact of the FLAIR image on the segmentation of MS lesions by assigning a larger weight to the FLAIR image channel than the other channels.With the inherent mechanism of estimation of the bias field,our method is able to deal with the intensity inhomogeneity in the input multichannel MR images.In the application of our method,we only use Tl-w and FLAIR images as the input two channel MR images.Especially,our method is robust and totally automatic.Experimental results show this promising result of our method.2.This chapter proposes a level set method for Multiple Sclerosis(MS)Lesion segmentation from FLAIR Images in the presence of intensity inhomogeneities.We use a three-phase level set formulation of segmentation and bias field estimation to segment MS lesions and normal tissue region and Cerebrospinal fluid and the background from FLAIR images.To save computational load,we derive a two-phase formulation from the original three-phase formulation to segment the MS lesions and non-lesion regions.Our method is able to deal with intensity inhomogeneity,which is a common image artifact in MRI.More importantly,our method exhibits desirable segmentation accuracy,and allow the use of flexible contour initialization.3.In this part,firstly,the modified LBF model is applied to the segmentation of MR images of cerebrovascular.In order to reduce the error of segmentation of small blood vessels,the improved method will be used to calculate the fitting function in the process of level set evolution.we define the Heaviside function’s parameter value be a smaller number,so that the zero-level set curve converges more precisely to the target area.Secondly,we also make full use of the local average gray scale of the blood vessel in the MRA image,which is higher than the local average gray scale of the background,giving a forced sequential rearrangement of the fitting functions f1and f2 such that f1<f2.In addition,the level set regularization term preserves the regularity of the level set function in essence,ensures the accuracy of the computation,and avoids the large number of re-initialization of the evolving level set function.Due to allow a very flexible initialization,the practical application of a great convenience.4.Accurate ventricle segmentation is a challenge technique for the development of detection system of ischemic stroke in computed tomography(CT),as ischemic stroke regions are generally adjacent to the brain ventricle with similar intensity.To address this problem,we developed an objective segmentation system of brain ventricle region in CT.Brain image was first aligned based on the midsagittal line(MSL)detection.The intensity distribution of the ventricle was estimated based on clustering technique,connectivity,and domain knowledge,and the initial ventricle segmentation results were then obtained.To exclude the stroke regions from initial segmentation,a combined segmentation strategy were proposed,which is composed of three different schemes.The proposed method yielded a high reliabilityfor ventricle segmentation,and obtained a high correlation coefficient between segmentation result and reference standard.Therefore,the segmentation system can offer a desirable performance.