Study On Vascular Structure Recognition In Coronary Angiograms

Coronary angiogram is the “gold standard” for the diagnosis of coronary artery disease.Quantitative analysis on the location,extent and type of coronary artery disease,which is the key to accurately evaluate the severity of coronary artery disease,is significant for treatment decisions.The pinpoint of lesion needs to specify the location of lesion in the coronary artery branches.The evaluation of lesion extent relies on the measurements of coronary branches diameters.However,automatic method which is used to identify coronary structure and measure diameters of coronary artery has not been yet fully researched.In this paper,based on image segmentation technologies,it is automatic to identify the structure of coronary artery tree,separate the coronary artery tree according to the bifurcation structures,and evaluate the degree of stenosis.The main contents of this paper are as follows:(1)The pretreatment of coronary angiograms.First of all,in order to solve the problem of high noise and low contrast in coronary angiograms,the adaptive fuzzy switching median filter(NAFSM)is used to remove the salt and pepper noises from images,and the contrast limited adaptive histogram equalization(CLAHE)is adopted to enhance the contrast of images.Then,in view of the complexity of coronary tree structure,multi-scale linear enhancement method based on the Hessian matrix is used to emphasize vascular structures and inhibit non-vascular structures by analyzing the characteristics of the Hessian matrix.(2)The extraction of the vascular structures and the automatic separation of coronary arterial tree.Firstly,the narrow band level set method is used to extract the coronary contours.Secondly,based on the characteristics of the Hessian matrix analysis,the seed points and the initial tracking directions are detected,and then the adaptive tracking method is adopted to extract coronary skeleton,bifurcation points and endpoints.Finally,according to the characteristics of coronary bifurcation structures,the main stem and its branches of vascular tree are extracted,and the coronary artery tree is separated.(3)The measurement of vascular diameters and stenosis degree.First of all,in order to optimize the diameter measuring results,a diameter measuring method based on the interaction design is proposed.Then,according to the skeleton and contour structures of coronary artery,the diameter of the target vascular is measured.Secondly,based on the direction of coronary skeleton,the contour of coronary artery,and the separation of vascular tree,the mean diameter and minimum diameter of the vascular segments are assessed,and then the extent of vascular stenosis is evaluated according to the stenosis evaluating definition.(4)The development of a coronary angiography image processing system.A coronary angiography image processing system is developed under the MATLAB environment including three functional modules: the user login interface,the data reader,and the segmentation interface.The image processing module achieves image preprocessing,coronary structure recognition,vascular diameter measurement,and stenosis assessment.The results show that: firstly,the preprocessing of coronary angiograms can inhibit the image noises and enhance the vascular structures.Secondly,the coronary artery structure including skeleton,contour,bifurcation points and endpoints can be accurately extracted.Then according to the structure of coronary artery,coronary arterial tree can be separated.Thirdly,based on the interactive design,the diameter of the cross section of an arbitrary vessel can be measured,and then the degree of stenosis of the target blood vessel segment can be properly evaluated.Finally,a coronary angiograms processing system is developed to realize coronary angiograms processing functions including the image preprocessing,the identification of coronary arterial structures,the separation of vascular tree,the measurement of vessel diameters,and the evaluation of vascular stenosis.This paper implements the automatic identification of coronary structure including skeleton,contour,bifurcation points and endpoints,the separation of vascular tree,the measurement of vascular diameter,and the evaluation of coronary stenosis.The separation of vascular tree can accurately match the standard coronary arterial structure.The measurement of vascular diameter and the evaluation of vascular stenosis can quantitatively analyze the degree of coronary artery lesion to evaluate the severity of coronary artery diseases,which is significant to the surgical decision making.