A region-based approach combining marker-controlled active contour model and morphological operator for image segmentation

Abstract

This paper presents a segmentation method for medical images. The goal is to localize and parameterize the blood vessel in these images for subsequent analysis. We proposed a new segmentation strategy that combines a marker-controlled active contour model and the mathematical morphology methodology. The algorithm presented in this paper is based on mathematical morphology operator and the implicit active contour or level sets for the detection of vessel-like pattern. The initial guess for the contour is from the vessel-like pattern contour marked and controlled by the user. The contour is then moved by image driven forces to the boundaries of the blood vessel using morphological operators: erosion, dilation, opening, and closing. We focus on the robustness of the sensitivity of the morphological operator and also the initialization and the placement of the initial contour. The experiment results are evaluated visually and quantitatively for its robustness in detecting blood vessels in medical images.