The MRI-based Segmentation And Three-dimensional Reconstruction Of Brain Tumor

As a common nervous system disease,brain tumors have seriously threatened the health of the patients.Megnetic Resonance Imaging(MRI)technology is widely used in the brain tumors diagnosis and surgical treatment because of its high resolution,no damage and multi parameter and multi-directional imaging characteristics.The study of tumor segmentation in the brain MRI image has become a hot topic at home and abroad,and it is also the practical need of clinical application.The segmentation results of the tumor image only show the focal area information on the two-dimensional level,but the three-dimensional structure of the tumor can only rely on the subjective imagination.3D reconstruction techniques can be used to display the features of the tumor in a three-dimensional way and facilitate the doctor to plan the surgical path.Therefore,the study of MRI brain tumor segmentation and three-dimensional reconstruction is of great value for computer aided treatment of brain tumors.This paper presents a fully automatic segmentation method for brain tumors,this method uses the symmetry of human brain structure information,through the regional growth method to achieve the brain tumors rough segmentation,then take the coarse segmentation region as the initial level set contour,using geodesic active contour(geodesic active improved contours,GAC)model for further accurate segmentation.Through the experimental analysis,it can be found that the model has good weak edge ability to divide the gray scale inhomogeneous.Finally,the image of the segmented brain tumor sequence was reconstructed by surface drawing and visualized to provide more information for the brain tumor research.The contents and results of this study are as follows:(1)according to the characteristics of MRI brain image sequence,this paper uses symmetry operation to preprocess the tumor image,the tumor area is effectively preserved by symmetry operation,and the tumor was clearly displayed by image enhancement and morphological processing process,provide the initial location for the subsequent rough segmentation process,plays a role in the automatic segmentation of medical sequence images.(2)in this paper,an improved GAC model is proposed,it integrates the region gray information of the image and improves the edge stopping function in the GAC model byconsidering the edge information of the image,which greatly improves the accuracy of segmentation.At the same time,the segmentation results of regional growth method are used as the initial contour of the improved GAC model,solved the initial location problem.The experimental results show that the proposed method has better segmentation performance.(3)in order to make full use of the 3D body data information of MRI medical images,make it better to assist surgeons to understand the three-dimensional structure of brain tumors In this paper,MC(Marching Cubes,MC)algorithm is applied to 3D reconstruction of brain and segmented tumor sequence images,showing the three-dimensional shape of tumor.The accuracy of reconstruction is verified by experimental comparison.