Smart Detection of Blockage in Coronary Artery in Angiography

Abstract

In computer aided diagnosis of artery motion analysis coronary angiogram segmentation is of crucial importance. With vascular structures along with considerable variation in intensities and noise it is challenging to develop an automated and accurate vessel
segmentation algorithm. The proposed approach is an unsupervised approach with coronary angiography as the source and is used to extract the vascular centerlines and segment
the vessels and detect the blockages in the coronary artery. Initially a preprocessing step
is applied to enhance and remove the low frequency noise in the image based on a contrast
limited adaptive histogram equalization and morphological filters. The vascular structure
is extracted by using Morphological hessian based approach and region based Otsu thresholding. Two different scales are used to extract the wide and thin vessels. Then the vessel centerline is extracted. A branch detection algorithm is employed to find the bifurcation.
The blockages are detected by considering the diameter along the cross sectional area of
the vessel. The proposed system has been analyzed and the experimental results conducted on several images prove the efficiency of the proposed method producing an accuracy of 96.98%.