Research On Coronary Angiography Image Enhancement And Its Application Based On DDFB

Coronary arteriography is a highly effective means for the treatment of coronary heart disease. During the imaging process of coronary angiogram, noises are caused due to impacts of heartbeat, imaging device, contrast agent, attenuation and other factors. Achieved images are unsatisfactory. The traditional vessel enhancement algorithm always increase the noise while enhance the contrast of vessel, which is unacceptable for the higher quality requirement medical images.In order to achieve more satisfactory images and more precise diagnosis, it is necessary to carry out noise-reducing, vessel-enhancing and other treatment on preliminarily achieved images. Taking coronary angiogram images as study objects, the paper analyzes characteristics of coronary angiogram images and enhances vessels. Major study contents cover the followings.(1)Coronary angiogram images were enhanced with the method combining decimation-free directional filter banks and Hessian Matrix. Based on the directionality of vessels in coronary angiogram images, an image enhancement algorithm based on decimation-free directional filter bank was presented in this paper to effectively extract more intact vessel information in angiogram images. Firstly, decimation-free directional filter bank was adopted to extract directional features of image vessels and to achieve directional images; secondly, Hessian Matrix-based image enhancement was performed in all directional images separately; lastly, all enhanced directional images were reconstructed to achieve final enhanced images. Compared with traditional method of achieving direction composition by directional additions, the reconstruction method was approved in this paper to a direction gradient-based reconstruction, which enhancing weights of principal direction to a more sufficient extent by using the characteristic of vessel directionality. It was suggested by the experiment that this method effectively improved vessel information in coronary angiogram images.(2)A method was presented to calculate vessel diameter rapidly and estimate vessel stenosis. Firstly, enhanced images achieved with the method in this paper were further processes to achieve vessel edges and center lines; secondly, based on pixel characteristics in binary images, draw lines perpendicular to center lines and find intersections by logic and operation to decrease calculation amount and improve speed for computation of vessel widths; lastly, average vessel sectional areas were calculated for different vessel segments to estimate the existence of stenosis.