Title :
Accurate Lung Segmentation For X-ray CT Images
Author :
Gao, Qixin ; Wang, Shengjun ; Zhao, Dazhe ; Liu, Jiren
Author_Institution :
Northeastern Univ., Shenyang
Abstract :
Segmentation of pulmonary X-ray computed tomography image is the first step to detect nodule of lung. This paper presents a fully automatic method for segmenting lung from three-dimensional thorax. First, the large airway is removed from lung region by anisotropic diffusion to smooth edges and region- growth. Second, we use an optimal threshold to automatically choose a threshold value, get binary images, and remove vessel of pulmonary using the optimal threshold. Third, left and right lungs are separated by detecting the anterior and posterior junctions using the largest threshold. Finally, we smooth the lung boundary along the mediastinum and lung wall by morphological smoothing. We show that our results are better than that achieved manually. Since there are no accepted criteria for defining the lung boundary near the mediastinum, we believe that our method of defining the boundary of lung near the mediastinum based on the structure of the airway tree provides a good basis for three-dimensional smoothing.
Keywords :
computerised tomography; image segmentation; medical image processing; smoothing methods; X-ray CT images; accurate lung segmentation; image segmentation; lung boundary; lung wall; morphological smoothing; pulmonary X-ray computed tomography image; three-dimensional thorax; Anisotropic magnetoresistance; Computed tomography; Image edge detection; Image segmentation; Lungs; Smoothing methods; Thorax; X-ray detection; X-ray detectors; X-ray imaging; An isotropic; Diffusion; Lung Segmentation; Region Growth;
Conference_Titel :
Natural Computation, 2007. ICNC 2007. Third International Conference on
Conference_Location :
Haikou
Print_ISBN :
978-0-7695-2875-5
DOI :
10.1109/ICNC.2007.157