CT Lung Image filtering based on Max-Tree method

Abstract

CT Lung Image formally used by radiologist to ensure the diagnosis of the patients. Manually visualization by radiologist aims to segment which part of the lung could be subject of the disease, for example lung nodules as an early suspect object of lung cancer. Computational method gives opportunities to this subject segment the curious object without much human interventions. Max-Tree constructs nodes based on the connected nodes that have similar characteristic. This method has three stages, constructs the tree, filter stage and image reconstruct phase. We applied Max-tree filters to two different well-known public medical image databases such as LIDC and Lung Time using DICOM standards. This research shows that Max-Tree methods exploit the structural features of the nodes that are similar to the images. This ability is then be used to get better visualization and eliminate the other parts that are not needed by radiologist.