Research On Coronary Heart Disease Computer-aided Diagnosis Methods Based On Computed Tomography Angiography

Coronary heart disease(CHD)is one of the main factors which threaten human health for a long time.Some clinical research shows 64-slice spiral CT can meet the need of CHD screening,which has the same accuracy with coronary angiography.Nowadays,CHD check and diagnosis has a higher demand for the proposal of the new standard of the coronary lesion classification based on plague attenuation modal.This paper researches on methods of the coronary heart disease computer-aided diagnosis.This method aims at the automatic classification of coronary lesion,which can provide the aided diagnostic information for doctors,improving the accuracy and efficient of CHD.In addition,this paper designs the medical image interaction display software with the help of the VTK which has some visualization function.The main work in this paper are shown below.1?Design the CAD algorithm based on CT data.The algorithm can be divided in to two parts,including building the coronary structure and lesion detection and classification.And the first part includes pre-procession,centerline tracking and vessel walls extraction.2?Apply the pre-procession on the 3D CT data combined anisotropic diffusion filter and vessel filter based on Hessian matric.This method can reduce the image noise and increase the vessel contrast.And the result can be valid for centerline tracking algorithm's input.3?Apply some heuristic methods in the traditional minimum cost path centerline extraction algorithms,which can decrease the time and space cost in algorithms and keeps the same accuracy as the traditional algorithms.4?Design the “Ray Casting” method for image sampling with the centerline as the source point,calculate the walls location in the “Rays” based on the grays decrease on the vessel walls,match the vessel walls by the disperse walls location.This method improves the vessel accuracy compared with traditional methods.5?Develop basic functions of medical image visualization,including the DICOM image browsing,image multi-planar reformation and so on,set up the software platform of medical image visualization.