Segmentation of coronary arteries
Sabeen Basit, Shoab Ahmed Khan, Muhammad Usman Akram · 2014
Coronary artery disease is a one of the leading cause of death and it is most common heart disease. Diagnosis of coronary artery diseases requires segmentation of coronary arteries it is also a precondition in many of the medical imaging procedures. In this paper we present multiscale model for 2D segmentation of coronary arteries combining the directional information and hessian eigenvalue analysis. First eight directional images are calculated using directional filter bank of same size as the input image, each represents the enhanced vessels in respective direction. Then the eigenvalues and eigenvectors of hessian matrix are calculated at each directional image, based on that values enhancement filter is applied to each directional image. Enhancement filter enhances the vessels in corresponding direction. Finally the output image is obtained by recombining the enhanced directional image. The system applies post processing to remove aorta and extra noise. The evaluation is performed on locally gathered images.