Home Health Computational Anatomy Based on Whole Body Imaging: Basic Principles of Computer-Assisted Diagnosis and Therapy
Kensaku Mori, Mikio Matsuhiro, Yoshiki Kawata and Noboru Niki
Bronchus and Vessels
The tracheobronchial tree is the macroscopic framework of the respiratory system. The trachea bifurcates into two branches, the right main stem bronchus and left main stem bronchus. Many branchings continue down to the bronchioli, which allow air to diffuse into the alveoli, the site of actual gas exchange. Because the bronchial tree is filled with air, it has negative density on CT. Typical CT values of the bronchial lumen are in the range —1000 H.U. to —900 H.U. Because of the partial volume effect, the CT value becomes higher as the branches become thinner. To segment the tracheobronchial tree from chest CT examinations, the basic method involves tracing the negative-density regions from the trachea, which can be easily identified, in the direction from the center to the peripheral. There are several methods for extracting bronchial regions. These methods can be classified into two categories: (a) the region-growing-based method and (b) the machine learning-based method. In the category (a), basic region growing (e.g., one threshold value in region growing)  or adaptive region growing (e.g., changing threshold branch by branch)  is utilized. The category (b) determines whether each voxel belongs to the bronchial lumen or not based on features computed at each voxel. This classification is done by morphological operation or machine learning . Then the selected voxels are connected to portray the tracheobronchial tree.
Lo et al. (2012)  discussed the comparison of world-representing methods for bronchus extraction from 3D chest CT images. Their findings were based on results of a bronchus region extraction competition held in MICCAI in 2009.
|< Prev||CONTENTS||Next >|