The Shape Analysis Of Ventricular Septums Based On Image Data

WHO authoritative report shows cardiovascular diseases seriously endanger human health. Computer aided diagnosis of heart has an important supporting role in the diagnosis of cardiovascular diseases. Ventricular septum is an important medium for ventricular interaction. It can reflect the ventricular systolic movement. Around the ventricular septums, combined with the medical image digital image processing and simulation modeling technology provide efficient and accurate algorithm and simulation research platform for cardiovascular diseases.Based on the short axis cardiac magnetic resonance images, through ventricle segmentation and morphological image processing algorithm to extract ventricular endocardial boundary and ventricular septum skeleton.According to the different characteristics of left and right ventricle, ventricular segmentation using region growing algorithm and feasible model algorithm. Different algorithms of segmentation results will be used for different follow-up work. Strive to ensure analysis and calculation accuracy while reducing medical image analysis for segmentation requirements Based on the results of regional growth algorithm, using the morphological algorithm for ventricular endocardial boundary extraction; ventricular wall was established based on the result of feasible model. And then using skeleton extraction algorithm to obtain ventricular membrane skeleton. In the morphological analysis, an analysis of ventricular systolic movement by calculating changes in the ventricular endocardial boundary perimeter; The calculation of ventricular interval length provide reference for the diagnosis of myocardial hypertrophy; Realize the automatic calculation of ventricular septum curvature; proposed the concept of ventricular septum deviation, and calculate the mean value and variance of offset. Through three-dimensional reconstruction algorithm, using two-dimensional image data to construct three-dimensional motif; through meshing algorithm, convert the3D model into finite element model.All the work in this thesis is based on the actual image data. In the forward process or the realization of algorithm, try to keep the automation and accuracy of this algorithm, so as to meet the need of computer aided diagnosis. Experimental results show that:through automatic or semi-automatic algorithm to obtain and analyze the important parameters of ventricular septum that can be evaluated ventricular septum morphology can achieve the purpose of calculating interest diagnosis.