Active Contour Based Lung Field Segmentation
Jun Lai, Ming Ye · 2009
Conventional methods that perform lung segmentation in CT rely on a large contrast in Hounsfield units between the lung and surrounding tissues. But the segmentation of lungs affected by high density pathologies that are connected to the lung border and discontinuities in the pixel intensities may be caused by X-ray projecting intensity changes, differing tissue reflectance and transmission properties, so the usual segmentation can't get the good results. Here we proposed a novel and more effective algorithm used for segmenting lung fields in CT images. The segmentation algorithm is based on active contour with prior shape to fit boundary of lung field. Owing to our algorithm fitting the shape profile for pulmonary area with active contour under the shape controlling, it can be used in complex shape pulmonary regions fitting, especially suitable for the segmentation of lung field with a lot of juxta-pleural pulmonary nodules. Experiments demonstrated that our algorithm could segment the lung field with pathology of different forms more precisely. From the test results, we can see that the proposed technique was successful in segmentation on lung CT image and it is found to have many advantages over the exiting methods.