CT lung image classification with correction for perfusion gradient

On computed tomography (CT) images of the lungs, areas affected by some small airways diseases appear as dark patches. The detection of these areas can be difficult due to the subtlety of the intensity difference between normal and diseased lung parenchyma. This paper presents an automated method for the reproducible quantification of the affected lung parenchyma in the presence of intensity distortion caused by gravity dependent perfusion gradient. The technique was validated with both phantom studies and a clinical trial of 15 patients