Takayuki Kitasaka1*, Kensaku Mori2, Jun-ichi Hasegawa3 and Jun-ichiro Toriwaki1
1Department of Information Engineering, Graduate School of Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan
2Research Center for Advanced Waste and Emission Management, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan
3School of Computer and Cognitive Sciences, Chukyo University, Toyota 470-0393, Japan
*E-mail address: kitasaka@toriwaki.nuie.nagoya-u.ac.jp
(Received November 11, 2002; Accepted December 12, 2002)
Keywords: Bronchus, Chest X-ray CT Image, Volume of Interest, Sharpening Operation
Abstract. A method for extraction of bronchus regions from 3D chest X-ray CT images—by analyzing structural features of the bronchus-has been developed. This method enhances the bronchial walls by applying a sharpening operation and segments each bronchial branch by recognizing the tree structure from the trachea. It uses the volume of interest (VOI) to segment each branch. That is, the VOI is placed so as to envelope the branch currently being processed. The region growing is performed only inside the VOI so that each branch is extracted by a suitable threshold value. The final bronchus region is obtained by unifying the extracted branches. The new method was applied to three cases of 3D chest X-ray CT images. The experimental results showed that the method improved the extraction accuracy significantly compared to the previous method; that is, the average accuracy of segmentation was 82% for 4th-order, 49% for 5th-order, and 20% for 6th-order bronchi, compared to 45%, 16% and 3% by the previous method using the region-growing method with the constant threshold value.