Segmentation of Coronary Artery Using Region Based Level Set with Edge Preservation

摘要

In this paper, we present a level set method for segmentation of coronary artery. The proposed method is a region-based method with the ability of edge preservation. In the proposed method, local intensity information is introduced into the level set formulation where the bias field is estimated by a linear combination of a given set of orthogonal basis functions. The intrinsic regularization mechanism of the proposed method is used to ensure regularity of the level set function and smoothness of the zero-level contour and reduce miss-and over-segmentation caused by the noise. We evaluate the proposed method on public images from the Coronary Artery Stenoses Detection and Quantification Evaluation Framework, a satellite event of MICCAI 2012. Experimental results and comparison analysis demonstrate that segmentation results of the proposed method are close to that from three blinded observers.