基于CT影像处理的心脏腔室四维建模及分析系统设计

发布时间：2018-05-15 01:17

本文选题：四维动态模型 + 图像分割　；参考：《南昌航空大学》2017年硕士论文

【摘要】：随着医学成像电子仪器和计算机图像处理技术的快速发展,医学影像的诊断及分析技术也相应得到快速提升。运用计算机对医学影像进行数字化的处理、提取关键信息参数或构建专业化的分析系统等计算机辅助诊断技术在当下医疗领域显得尤为重要。本研究旨在开发一种计算机软件仿真系统用来进行医学影像建模,能够实现心脏内部各个独立腔室的四维重建,从而构建有关局部心肌动量的四维动态显示模型组,达到心脏腔室生理分析及病变评估的功能。心脏疾病是人类非意外死亡的主要死因,但又因心脏生理结构及功能的特殊性而无法做太过深入的侵入式诊断或研究。心脏的医学成像及三维模型仿真虽然为诊断或研究提供了便利,但二维的心脏医学影像及三维化的心脏重建模型并不能多方位、实时的展示心脏各部位状态与其生理参数变化。基于心脏器官具有形态上实时变化的生理特殊性,四维化的仿真模型在分析及评估心脏功能方面有着绝对优势:相比于传统三维模型,四维化可以通过动态模型展示心脏的形态变化,模拟心脏的跳动;在心脏参数的分析上,可以生成整个心跳周期下不同时段的生理数据资讯,实现对心脏生理变化的综合评估。鉴于传统心脏成像与模型仿真技术的不足,本文研究设计了一种心脏腔室的四维建模与分析系统,可以实现对病人心脏生理状况全面且精确的运动模拟及功能评估。在建模系统的内部算法处理上,按照影像重组、图像分割、三维重建和四维化的四个主要流程进行整体设计。以切面构造法实现影像重组的中轴矫正办法,成功获取了目标心室的短轴切面图像;改进了基于区域增长的图像分割法,结合自动分割与人机交互的方式进行心肌边缘资讯获取,增加了图像分割的准确性和抗噪性;开发了以数据点建立向量关联进行三维重建的方法,同时通过过渡模型及颜色梯度处理等方法得到具有心肌局部动量评估功能的四维模型。在系统的外部界面设计上,分阶段的提供了有关图像重组与分割的手动修正界面、三维重建及参数提取界面以及多组四维动态模型同平台显示界面等,实时反馈建模各阶段效果和参数,并在最终分析平台中实现了多种特殊功能四维模型之间的联合显示与对比。该系统成功构建了电烧术后病人的左心室四维重建模型,并通过心肌动量模型与参数分析评估了其生理状况,表明了本系统所设计无创且便捷的心脏腔室四维建模及评估方案有效可行。
[Abstract]:With the rapid development of medical imaging electronic instruments and computer image processing technology, medical image diagnosis and analysis technology has been rapidly improved. It is very important to use computer to digitize medical image, extract key information parameters or build professional analysis system. The purpose of this study is to develop a computer software simulation system for medical image modeling, which can realize the four-dimensional reconstruction of each independent chamber in the heart and construct a four-dimensional dynamic display model group of local myocardial momentum. To achieve the heart chamber physiological analysis and disease evaluation function. Heart disease is the main cause of human non-accidental death, but it can not be too deep invasive diagnosis or research because of the particularity of cardiac physiological structure and function. Although the medical imaging and three-dimensional model simulation of the heart provide convenience for the diagnosis or research, the two-dimensional cardiac medical image and the three-dimensional heart reconstruction model can not show the state of each part of the heart and the changes of its physiological parameters in real time. Based on the physiological characteristics of real-time morphologic changes in cardiac organs, the four-dimensional simulation model has an absolute advantage in analyzing and evaluating cardiac function: compared with the traditional three-dimensional model, In the analysis of heart parameters, the physiological data of different periods in the whole heart cycle can be generated, and the comprehensive evaluation of the physiological changes of the heart can be realized by using the dynamic model to show the changes of the heart morphology and to simulate the beating of the heart. In view of the deficiency of traditional cardiac imaging and model simulation techniques, a four-dimensional modeling and analysis system of cardiac chamber is designed in this paper, which can realize the comprehensive and accurate motion simulation and functional evaluation of the patient's heart physiological condition. According to the four main processes of image recombination, image segmentation, 3D reconstruction and four-dimensional design, the internal algorithm of the modeling system is designed as a whole. In this paper, we successfully obtain the short axis section image of the target ventricle, and improve the image segmentation method based on the region growth by using the method of section construction to realize the central axis correction of the image recombination, and to obtain the short axis section image of the target ventricle successfully. Combining automatic segmentation and human-computer interaction to obtain myocardial edge information, the accuracy and anti-noise of image segmentation are increased, and the method of constructing vector association based on data points for 3D reconstruction is developed. At the same time, a four-dimensional model with the function of evaluating the local momentum of myocardium was obtained by the methods of transition model and color gradient processing. In the external interface design of the system, the manual correction interface about image recombination and segmentation, the 3D reconstruction and parameter extraction interface, and the display interface of multi-group four-dimensional dynamic model with the platform are provided in stages. The effect and parameters of each stage are modeled with real-time feedback, and the joint display and comparison among four dimensional models with special functions are realized in the final analysis platform. The system successfully constructed the left ventricular four-dimensional reconstruction model of the patients after electrocautery, and evaluated the physiological status of the patients by myocardial momentum model and parameter analysis. The results show that the non-invasive and convenient four-dimensional modeling and evaluation scheme is effective and feasible.
【学位授予单位】：南昌航空大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R816.2;TP391.41

【参考文献】