Research On Extraction And Segmentation Of Coronary Artery From X-Ray Angiograms

Cardiovascular disease serious harm to human health, it will take away more than12million lives each year, nearly a quarter of the total number of deaths in the world. It is often caused by coronary artery stenosis and blockage. Coronary artery play an extremely important role in the supply of blood to the heart. Coronary angiography is an important method of diagnosis of heart disease, which is recognized as the gold standard for diagnosis of the cardiovascular disease. Especially in China, X-ray coronary angiography is widely used. The vascular contrast agent makes the vessel become lighter than other organizations. However, the interpretation to the angiographic images is impaired by the effect of vessel foreshortening, overlap and noise because of the uneven distribution of the contrast agent and the environmental noise influence of image processing. In order to improve the level of clinical diagnosis, the extraction and segmentation of coronary artery from X-ray angiographic images is very necessary. In this paper, we perform three aspects of research on extraction and segmentation to the coronary angiograms:Firstly, we introduce several methods about image preprocessing, detailed image smoothing and image enhancement, we adopt the median filtering method to smooth coronary angiograms, the method is better able to suppress the noise and maintain vessel edges; we adopt Gabor filtering method to enhance vessels in images, image details can be integral remained, and have a better visualization.Second, we propose a method for segmentation of the coronary artery in angiograms, which combines Hessian matrix multi-scale filtering and region-growing. As the relationship among the eigenvalues of the Hessian matrix, Hessian matrix multi-scale filtering is used to enhance vessels in angiographic images. The region-growing step is performed on the enhanced images to extract vessels from the coronary angiograms. Using this method, both the coronary artery trees and most of smaller distal vessels could be extracted clearly.Finally, we introduce the basic theory of Markov random, compared to the several MRF models, illustrate several parameter estimation methods and image segmentation algorithms based on MRF. The MRF can make the spatial information of images be combined, hence we add MRF to the extraction and segmentation of coronary artery, and achieve the satisfying segmentation results.