An Effective Enhancement and Segmentation of Coronary Arteries in 2D Angiograms

Vessel enhancement in two-dimensional angiogram images is an essential pre-requisite step towards the isolation of coronary arteries. Hessian-based filters are the most commonly used vessel enhancement filters; however, these filters are more sensitive to noise and suppress the bifurcation regions. Suppression of bifurcation regions results in disconnected vessels. In this study, we present a technique that enhances the arteries of the heart in 2D angiograms and also refines the noisy vesselness obtained through Frangi's method by using guided filter which produces more enhanced image that can be used as an effective pre-processing step for binarization of the Frangi vessel response having less discontinuities and joint suppression. The proposed approach makes use of the guided filter which smooths the edges, and at the same time preserves the edges as well for the enhancement of vessels. Following this filter, an Adaptive thresholding is applied to segment the coronary arteries from the angiogram. The proposed method has been tested on real angiography images and the efficiency of the method has been shown qualitatively as well as quantitatively.