Localization of Coronary Artery Thrombosis Using Coronary Angiography

Computer-aided diagnosis (CAD) systems have gained a lot of popularity in the past few decades due to their effectiveness and usefulness. A large number of such systems are proposed for a wide variety of abnormalities including those related to coronary artery disease. In this work, a CAD system is proposed for such a purpose. Specifically, the proposed system determines the location of thrombosis in x-ray coronary angiograms. The problem at hand is a challenging one as indicated by some researchers. In fact, no prior work has attempted to address this problem to the best of our knowledge. The proposed system consists of four stages: image preprocessing (which involves noise removal), vessel enhancement, segmentation (which is followed by morphological operations) and localization of thrombosis (which involves skeletonization and pruning before localization). The proposed system is tested on a rather small dataset and the results are encouraging with a 90% accuracy.