Research And Implementation Of Brain Tissue MRI Image Segmentation Algorithm Based On MRF Method

Currently,the prevalence of brain disease is a serious threat to people's health and life.In the medical field,nuclear magnetic resonance imaging is a key technology for the detection of brain diseases,MRI image segmentation of brain tissue is of great significance for the detection and treatment of brain diseases.This thesis completed skull image segmentation based on BET algorithm and segmentation of MRI image of brain tissue based on GMM after analyzing MRI imaging features of brain tissue.On the basis of the analysis of Markov random field MRF model and gray feature model of MRI image,the double characteristics random field model of MRF was established.Completed the segmentation of brain MRI image and experimental comparison and analysis of improved algorithm based on double features MRF.The innovation of this paper lies in: The EM algorithm based on genetic algorithm(GA)is improved by using the method of initializing K parameters automatically,improved the convergence of the algorithm.Proposed the improved method for brain tissue segmentation of gray features and texture features based on MRF,improved the accuracy of brain tissue segmentation.The research work and primary coverage of this thesis:(1)Completed the algorithm flow design of brain tissue segmentation based on MRI image feature analysis.On the basis of completing the threshold method and region growing method of MRI image segmentation methods and the realization of algorithmic programming.Completed the research and implementation of BET algorithm based on segmentation of the cranium,and revised the calculation of force in the BET algorithm.Completed the segmentation by improved BET algorithm and completed the experimental comparison and analysis,and the feasibility of the algorithm is verified by the experimental results.(2)Designed GMM algorithm segmentation process based on analyzing GMM algorithm.Complete GMM parameter estimation and simulation analysis based on EM algorithm which is based on EM algorithm,proposed improved EM algorithm based on GA,and through programming and simulation analysis verified that the improved EM algorithm has improved the convergence.Completed the segmentation algorithm of brain tissue MRI image based on the GMM model of brain MRI image.(3)Designed brain tissue segmentation algorithm based on MRF after analyzing MRF model.Completed research and analysis of brain tissue segmentation algorithm based on MRF.On the base of analyzing the brain MRI image grayscale characteristics with the airport and texture feature random field model and the label field,the improved model based on MRF double characteristic random field was proposed.Completed the design of brain tissue segmentation algorithm process based on MRF double characteristic field.And used MAP and ICM to optimize the solution of the algorithm.Completed research and comparative analysis of brain tissue segmentation algorithm based of MRF double characteristic field and it is verified that the improved algorithm has high segmentation accuracy.