Reconstruction Model Of Coronary Lesions Based On Ultrasound Images

Coronary heart disease seriously jeopardizes human health.The diagnosis and treatment of coronary heart disease has always been the focus of medical research.Intravascular ultrasound(IVUS)is one of the most important methods for clinical diagnosis of coronary heart disease.Through the ultrasound image,the doctor can observe the cross-sectional information of the lumen in the blood vessel,and analyse the thickness of the blood vessel wall,the distribution,size and composition of the plaque in the lumen.However,doctors have limitations in observing vascular lesions based on ultrasound images,and cannot observe the three-dimensional model of blood vessels.Therefore,the visualization of vascular models is the key point of current research.This paper focuses on the following points:Regarding the existence of interference factors in the image acquisition process(i.e.noise,spot,artefact and so on),this paper evaluates the image filtering algorithm.Since a single filtering method cannot meet the requirements,this paper proposes to optimise the wavelet filtering algorithm in order to eliminate the noisy point in the ultrasound images.In addition,this paper chooses region growing segmentation algorithm which is a more effective method to extract the blood vessel contour.This paper remodels the three-dimensional construction.Due to the single structure inside the blood vessel,the Marching Cubes(MC)algorithm,which is an algorithm in the surface rendering method,will be used.Nevertheless,the current calculation operation for the MC algorithm is inefficient and ambiguous in multiple inheritance.Therefore,this paper will evaluate and optimise the MC algorithm to convert the surface with ambiguity problem and eliminate the ambiguity in multipath inheritance.Because the vascular model reconstructed by the MC algorithm is composed of a large number of triangular mesh data,the internal it contains a large number of redundant triangular patches.It reduces the computational efficiency,slows down the model rendering and has a negative impact on human-computer interaction.Furthermore,it has limitation on data storing and transmission.Regarding all these drawbacks,the current model needs to be simplified.By comparing and evaluating various simplified algorithms,it is found that the Quadric Error Metrics(QEM)algorithm based on edge folding has an ideal outcome.Therefore,I choose to use QEM algorithm to optimize the vascular mesh model.In addition,in regard to the limitation of QEM algorithm,I decide to add variables like the vertex curvature and partial area in the algorithm to improve the efficiency of the QEM algorithm.Finally,by using the VS 2010 platform,a three-dimensional reconstruction system of blood vessels is designed,which mainly includes image pre-processing,reconstruction of blood vessel model and simplification of the model.This system could converse the digital image data to visual model,and generate a three-dimensional model of blood vessels,and lay the foundation for the establishment of a virtual training platform for cardiac intervention.