Segmentation of Lung Vessels Together With Nodules in CT Images Using Morphological Operations and Level Set

Abstract

With a fast development of computer tomography (CT) technology, CT images has become one of the most efficient examination methods of lung diseases in clinical. The appearance of vessels and nodules together with their changes over the time in CT images may provide an exact diagnosis. Segmenting blood vessels, extracting nodules together with distinguishing between vessel junctions and nodules have become important clinical challenges. Some factors usually used to distinguish between blood vessels and nodules include the structure, shape, size, color and intensity differences, i.e. bright light or shady. The precision of segmenting lung vessels and nodules plays an important role in analyzing the volumetric growth rate and the nodule status. There are many applications of image processing techniques proposed and used nowadays to give radiologists necessary information in their work such as vessel enhancement, nodule enhancement, vessel and nodule segmentation, etc. With the recognition that intensity is one of the most important factors in classifying strong and weak vessels, solid and nonsolid nodules, this paper presents a way of segmenting vessels together with nodules in CT lung images into three levels of the intensity using morphological operations and level set method. Level-1 indicates the highest intensity or bright light regions, level-3 includes the lowest intensity or shady grayish regions and level-2 is