|
KMID : 1147120120180010032
|
|
Journal of the Korean Society of Imaging Informatics in Medicine
2012 Volume.18 No. 1 p.32 ~ p.38
|
|
|
Parameter Optimization for Automated Airway and Left-Right LungSegmentation of Low-Dose Three-Dimension Lung CT: Towards theConstruction of Korean Standard Database
|
|
Heo Chang-Yong
Kim Jong-Hyo
|
|
|
Abstract
|
|
|
|
Segmentation of airways and left-right lungs is a basic and essential step in the computer analysisof chest CT. This study presents a parameter optimization method for automated airway and left-rightlung segmentation of low-dose three-dimension lung CT. Our method consisted of following threesteps. First, a seed point of airway area was selected using the anatomical information of the airway,followed by selection of the optimal threshold value for airway segmentation based on the intervalchange of temporarily segmented airway volume according to the incremental threshold change.Second, seed points of left and right lung area were determined using the morphological informationand signal intensity of the lung, followed by applying a region growing technique resulting in segmentedlung. Finally, adjoined part of the left and right lungs were examined, and in case of fused lung, amorphological operation was applied to separate the lungs. The performance of the developedmethod was verified with a set of low-dose lung screening 3D CT exams of 465 cases. In addition, weobtained a set of basic image-derived data for quantitative lung healthy indices such as lung volumeand emphysema scores from the segmented lung and airway data.
|
|
|
KEYWORD
|
|
|
Low-dose Computed Tomography (CT), Airway segmentation, Lung segmentation, Parameter optimization, Lung volume, Emphysema score
|
|
|
FullTexts / Linksout information
|
|
|
|
|
|
Listed journal information
|
|
|
|