This paper presents a novel method for extracting bronchus regions from 3D chest CT images. Several research groups have proposed methods for extracting bronchus regions by using region growing algorithm. However, these methods do not work well in the case that tumors divide the bronchus region into several parts. Therefore, this paper proposes a method that can extract the bronchus region in such cases. The proposed method uses CT value distribution features for extracting the bronchus region. The bronchus region shows relatively lower CT value in comparison with that of the bronchial wall. Also, cross-section of the bronchus can be observed as circular or near-circular region. The proposed method tries to extract the bronchus region candidate by detecting voxels which have these bronchial features. We applied the proposed method to one CT image. Experimental result shows that the proposed method could extract regions of thin bronchial branches.     «

This paper presents a novel method for extracting bronchus regions from 3D chest CT images. Several research groups have proposed methods for extracting bronchus regions by using region growing algorithm. However, these methods do not work well in the case that tumors divide the bronchus region into several parts. Therefore, this paper proposes a method that can extract the bronchus region in such cases. The proposed method uses CT value distribution features for extracting the bronchus region....     »