Blood Vessels Detection Using Morphological Operations and Hough Transform in CT Images

The detection and segmentation of the blood vessels are often essential prerequisite steps in the identification of anatomy and pathology. In this study, an approach for blood vessels extraction using mathematical morphology and Hough Transform is presented. The main steps of the processing are: enhancement operation is applied to the original Computer Tomography (CT) image in order to remove the noise based on homomorphic filtration and morphological operation such as hit-miss or directly skeletonization is employed to extract blood vessels on the medial axes. For more accurate vessel tracking a Probabilistic Hough Line Transform (PHT) is applied to binary image of the skeleton modification or of selected region of interest (ROI).