Human Brain MR Image Segmentation Algorithm And Fiber Tracking Based On Fuzzy Clustering

Magnetic Resonance(MR)imaging technology is an important part of modern medical imaging technology.Due to its non-invasive and high-contrast features,its application in clinical medicine is gradually increasing.MR brain image segmentation is mainly to divide the brain tissue,that is,the MR brain image is divided into white matter,gray matter and cerebrospinal fluid and other tissue parts,MR brain image segmentation provides the basis for medical image registration,three-dimensional reconstruction and visualization.However,MR images have some defects,such as partial volume effect,uneven gradation,and noise.Therefore,in practical applications,segmentation of brain MR images is difficult to obtain high accuracy.Therefore,based on the classic fuzzy C-means clustering algorithm,this paper proposes some improvements to the algorithm in order to solve the defects of MR medical images,and studies the issue of brain fiber tracing and visualization.Specifically,this article mainly makes the following research:Firstly,in order to solve the problems of lack spatial information,low accuracy,and noise sensitivity in classical FCM(Fuzzy C Means)image segmentation algorithms,a novel spatial fuzzy clustering(KSFCM)algorithm combined with kernel function is proposed to segment brain MR images.The algorithm maps the elements in the image sample space to the high-dimensional feature vector space through the kernel function,which improves the accuracy of image segmentation.The normal distribution spatial information function is proposed by using the normal distribution characteristic of MR images,and the membership function of the modified spatial information function is combined to effectively solve the problem of image noise interference.A large number of experimental results show that the KSFCM algorithm is more accurate and more efficient in segmenting brain MR images.Secondly,because of the shortcomings of the traditional FCM algorithm such as the random selection of the initial cluster center and the slow convergence of the objective function,this paper proposes a new human brain DTI image segmentation algorithm based on the KSFCM algorithm(C-KSFCM).The algorithm first converts the DTI data into a two-dimensional scalar image,and then down-samples the data to reduce the amount of computation.Based on the downsampled data,the center pointof the class is selected using the density peak algorithm,and the selected point is used as the initial clustering center of the KSFCM algorithm to segment the human brain DTI image.Finally,in order to solve the problem that two-dimensional magnetic resonance images cannot express brain information in detail,this paper using the diffusion tensor data,which has the characteristics that the water molecule dispersion can express the fiber bundle's strike,the fiber tracking of human brain DTI data is performed using three fiber tracking algorithms based on the tensor domain,and then compared these result.The three-dimensional model is displayed intuitively to obtain the structure and direction of the inner nerve fibers of the brain,which can better assist in the diagnosis and treatment of clinical brain diseases,and lay a foundation for subsequent brain tissue research.