AUTOMATIC LUNG NODULE SEGMENTATION FOR MEDICAL DIAGNOSIS
Keywords:
pulmonary nodule; computer-aided detection; CT image; morphology; spatialAbstract
Lungs segmentation is a vital step for computer aided diagnosis. Many lungs diseases requiring radiological
support for diagnostic purpose, including tuberculosis, emphysema and lungs cancer could be effectively diagnosed using
computer-aid. Pulmonary nodules usually act as the early stage characterization of lung cancer. It’s a crucial way to control
the disease and reduce the mortality by further examination on the diagnosis of pulmonary nodules. But the large number of
the image data and lesion analysis make difficulties to radiology expert for the correctness and effectiveness. So the
computer aided detection for the pulmonary nodules turns out to be the hot subject of medical image processing. In this
paper, we follow the classic step in detecting the nodules and an effective algorithm based on the morphology and spatial
relationship to search the lesion. This method provides an extremely effective way to locate the nodules.
