The Research Of Vessel3D Segmentation Algorithm In CTA Images

In the present world, cardiovascular disease is one of the diseases with highestincidence and mortality rates. With the development of medical imaging technology inthe trend of digital and information age, the patients’ diagnosis process has being alsochanged in hospitals. In the field of clinical medicine, medical imaging technology hasbeen widely applied in the observation and diagnosis of disease. Doctors can usemedical image technology to understand the condition of patient’s diseases directly.But the body’s cardiovascular system is very complex, and the medical imageinformation is very large, which is from CT machine scanning the patient’s body. So itis impossible that use image segmentation algorithm to extract vessels for the doctordiagnosing cardiovascular disease.In the thesis, according to the character of CTA (Computed TomographyAngiography) images, and taking the vessel of CTA volume data as the segmentationobject, the paper analyzes the existing image segmentation method, and combining withsegmentation object, it has carried out effective blood vessel3D segmentation algorithmresearch, the research results are as follows:(1) The thesis uses symmetric region growing to achieve3D segmentation of thevessel which is in body abdominal CTA. In the process, this paper simply introduces thebasic theory and method of symmetric region growing, and it analyzes the advantagesof symmetric region growing in3D segmentation algorithm by comparing with thetraditional region growing. Then, based on the principle of symmetrical region growingmethod, the vessel3D segmentation algorithm is designed and achieved, and itdescribes the segmentation process, method and related technical points in more detail.Finally, applying the algorithm with the actual CTA images, By the segmentationexperiments, it analyzes the disadvantages of symmetric region growing in theapplication of vessel3D segmentation.(2) According to the drawbacks of symmetric region growing in the application ofvessel3D segmentation, the vessel3D segmentation algorithm combined with LBFmodel and symmetric region growing is put forward. The improved process of thealgorithm includes two key technology points: the first is makes the growing andjudgment process of symmetric region growing become simple, and the second isimproving the performance of the algorithm by boundary extraction technology of LBFmodel. By studying the LBF model algorithm basic principle and function and combining with the segmentation problem of the paper, it demonstrates the feasibility ofthe LBF model improving symmetric region growing, put forward the specificimprovement scheme, and describes the improved3D segmentation process of vesselsin detail. In the final experiment results and analysis, the original algorithm andimproved algorithm are compared in a variety of aspects, and it proves that improvedalgorithm has better performance in vessel3D segmentation applications.