Multimodality Medical Images Registration Based On Mutual Information

Multimodal medical image registration is becoming increasingly important in diagnosis,treatment planning,functional studies,computer-guided therapies,and in biomedical research.The matching algorithm based on mutual information,which has the advantages of high speed,good automation and superion accuracy,is widely used in medical image registration.So this paper mainly discusses multimodality image registraion based on mutual imformation.Firstly,it is introduced in the paper that the significance,development and clinical applications of medical image registration.Then the paper describes a simple summary of medical image registration.The konwledge about basic principle,implementation steps,and main methods of it.Meanwhile, both the classification of the registration methods and the concerned evaluation are surveyed.Secondly,the paper deeply studies the medical image registration based on mutual information,the background and the foundation concept of it are introduce.And detail discusses the afection of interpolation,outlier strategy and sampling technology.To solve the problem of local maxima in image registraiton based on MI,this paper analyses the reasons of the local maxima and investigates a approach of decreasing the gray levels of registration images according to the basic properties of the information entropy.Simulations have been done to to illustrate that this method can restrain the local maxima efficiency and convenient for the optimization algorithm to search the correct translation parameters.Finally in the rearch of optimized searching arithmetic,this paper introduces the Genetic Algorithm (GA).And because of the optimized objective function in image registraiton,an adaptive Real-coded Genetic Algorithm(ARGA) was proposed which is improved on the coding modes,genetic operators,etc.Then according to the two algorithms advantages we use the mixed optimization algorithm based on ARGA and Powell algorithm to optimize the function of information entropy.Experiment results show subvoxel accuracy can be achieved and the feasibility and efficiency of this method are verified.