Study 0n Vessel Recognition In Coronary Angiography Image

Coronary angiography image is the golden standard for diagnoses of cardiovascular diseases. With the improvement of living standards and life expectancy increases, cardiovascular disease in China is rising year by year, and it has become the main life-threatening diseases. According to coronary angiography images, the disease is not only timely diagnosed and aggressively treated, but also can be prevented from the development of the disease, and the pain of the patients can be eased. The image noises are generated in the processes which include the X-ray irradiation and attenuation pass through the body and image enhancement and optical focusing and optical switching and amplification and analog-digital conversion and other transformations. In order to get the accurate diagnosis and clinical treatment, this paper studies on the following three aspects in coronary angiography image:Firstly, coronary angiography image were preprocessed and smooth, sharpen and enhance. In addition, the Gabor filters and rotating one-dimensional Gaussian function are used to enhance the image, and suppress most of the noises, and highlight the vascular information. The computation burden is decreased comparing with the common two–dimensional Gaussian function method.Second, aiming at the goal of vessel path extraction, this paper presents the square template gray criteria to obtain the quasi-vascular center line. The concept of information measure is introduced. The computational method is obtained by combining together the three kinds of information measure of the neighborhood consistency of image with the structure of the image and direction of the image. The vessel edge line is detected based on the quasi-vascular center line.Finally, we give a method of the diagnosis of vascular stenosis. In order to find the vessel diameter, this paper uses the least squares to fit the curve. According to the slope of line and the coordinates of edge points, the vessel diameter can be calculated. The vessel diameter and stenosis can be used as an evaluation basis of clinical disease diagnosis. According to the normal diameter of proximal, vessel stenosis was compensated, and it can be used as reference data of medical diagnosis.The concept of information measure is introduced to process the angiography images and the template gray criterion is improved. The accuracy of vascular information processing has been increased comparing with the traditional method. This paper provides a new deal and method. A complete and practical medical processing program is developed.