Coronary artery tree segmentation in computed tomography angiography using Otsu method

Abstract

Analysis of coronary artery in medical images has been regarded as crucial step in the assessment of cardiovascular diseases during clinical diagnosis and surgical planning. Segmentation of coronaries in medical images is the most important step in such analysis. In this paper, an automatic segmentation method is presented for segmenting coronary artery lumen in Computed Tomography Angiography (CTA) datasets. Our proposed segmentation approach consists of two main steps. The first one is an enhancement step to enhance coronary vessels. The second is the coronary segmentation step done by applying Otsu method. Otsu thresholding is applied with different number of thresholds to check the proper number of thresholds that can better identify coronary vessel regions. Experiments are carried out on 18 real CTA datasets. Segmentation results are quantitatively evaluated using three different evaluation metrics and the results at each number of thresholds are compared.