Filter-based Image Processing of Coronary Angiography Imaging Data of the Left Coronary Artery

Abstract

Ischemic heart disease (IHS) is the most common type of heart disease in the world and is characterized by plaque buildup in the walls of the arteries that supply blood to the heart (called coronary arteries). In clinical practice X-ray coronary angiography is dominant imaging diagnostic procedure, even though it produces lower quality images due to high noise, imaging artifacts, vessel overlap, etc. In this work, we propose an X-ray coronary angiography image processing methodology that combines different filters in order to improve quality of images and separate blood vessels of left coronary artery from the background. In order to suppress the noise in the image and enhance edges of the left coronary artery, angiography image is processed with CLAHE, the Wiener filter and gamma correction. Image processed in this manner is then segmented using global Otsu’s thresholding technique, that binarizes the image into the region of interest (left coronary artery) and background. Otsu’s thresholding successfully separated blood vessel, but also it wrongly labeled some of the background as blood vessel. These non-vessel areas have been removed using connected components analysis.