Research On Thickness Measurement And Registration Of Cerebral Cortex Based On MRI

The brain is the most important part in human body,which controls the regular running of various tissue systems of the human body.Therefore,the occurrence of many diseases will be manifested through some changes in the morphological characteristics of the brain.The thickness of the cerebral cortex is an important indicator in the study of brain morphology.Doing research about it that can find the occurrence and evolution convention of the disease.Medical image registration technology plays an extraordinary role in computer-aided diagnosis,and it is an indispensable part in medical image processing.Through the registration,the image is presented in a more visualized form,which can provide the doctor with disease information and anatomical information of the lesion more accurately.Thereby,getting the doctors diagnosis efficiency improved.Nowadays,with the computer and image processing technology developing so fast,the two technologies can be combined and applied in various aspects.About the research on measurement of cortex thickness and image registration technology based on MRI images has become one of the main methods used in clinical diagnosis.According to the statistics from the World Health Organization,there are at least2.2 billion people visually impaired or blind in the world,of which at least 1 billion people have visual impairment that could have been prevented or has yet to be resolved.This thesis,by measuring and counting the thickness of the brain cortex after the reconstruction of the human brain MRI images based on the mesh surface model,analyzes the changes in the brain cortex thickness of the blind and reveals the blindness impact on the morphology of the cerebral cortex.And using the spherical parameterization method to realize spherical registration of the cerebral cortex,which overcomes research difficulties in cerebral cortex including complex structure,high curl and so on.The main research content and innovations of the thesis are as follows:(1)In the measurement of cerebral cortex thickness,this thesis optimizes the image segmentation steps in the image preprocessing.By combining two different types of segmentation methods,deformation model method and statistical classification method,moderating a single segmentation’s deficiencies.Then,the result of accurate segmentation is obtained.And through experiments,the effectiveness of the method in this thesis is verified,which provides the accuracy of the subsequent cerebral cortex thickness measurement with a guarantee.After that,with the cortex surface obtained by the three-dimensional reconstruction,the cortex thickness is calculated using the Laplace equation method;(2)Regarding the registration of the cerebral cortex,this thesis transforms the surface of the cerebral cortex into a mesh for spherical registration.Aiming at the distortion problem existing in the spherical parameterization algorithms,a conformal and area-preserving parameterization method is proposed.First,in order to solve the problem of the loss of model appearance details during simplification,this thesis introduces a curvature factor into the classic quadratic error metric simplification algorithm to ensure the surface characteristics of the model;then,the spherical parameterization is carried out by using the nuclear sampling method;finally,through discrete AMIPS energy is constructed to optimize the global distortion,and an effective parameterization result without flipping is obtained,and the distortion of the angle and area before and after the parameterization is minimum;(3)Design experiments to verify the effectiveness of this method.In the cerebral cortex thickness measurement,this thesis calculates the cortical thickness values of all the subjects based on the accurate segmentation results.Finally,comparing the cortical thickness differences between congenital blind,acquired blind and normal people with statistical analysis.In the registration of the cerebral cortex,first comparing the spherical parameterization algorithm with others.The result shows that our method has lower distortion.Finally,through this spherical parameterization method,the cerebral cortex surface is registered to the unit sphere,and the three-dimensional visualization of the cerebral cortex is realized.