講演抄録/キーワード |
講演名 |
2016-01-20 09:32
Dual-energy CT画像からの特徴抽出に基づく肝内胆管領域自動抽出 ○陳 鵬飛・田中 寛・小田昌宏・林 雄一郎・伊神 剛・梛野正人・森 健策(名大) MI2015-112 |
抄録 |
(和) |
(まだ登録されていません) |
(英) |
In this paper, we present our feasibility study on the intrahepatic bile duct (IHBD) extraction using dual-energy CT (DECT) volumes. We used the biliary dilation cases for this study. Compared with single energy CT (SECT), information contained in the volumes, i.e. CT values, increases in DECT. We expect that IHBD extraction performances improve by using the DECT volumes. We present a preliminary study of our IHBD extraction by using DECT volumes. As for the extraction process, we used the SVM classifier that utilizes CT values and local intensity structure analysis results as feature values. With lower-energy (100 kVp) volume being denoted by A and higher-energy (140 kVp) volume by B, the feature values we used include the CT values of A and B, and the eigenvalues of multiscale Hessian matrix from the CT volumes of A and B. The difference and effectiveness of these features and their combinations, were tested independently through our experiments. The extraction results were good despite the simplicity of the method, therefore, we believed the utilization of DECT for biliary tract extraction is promising. |
キーワード |
(和) |
/ / / / / / / |
(英) |
Intrahepatic biliary duct extraction / Dual-energy CT / Abdominal contrast-enhanced / Machine learning / SVM / / / |
文献情報 |
信学技報, vol. 115, no. 401, MI2015-112, pp. 187-192, 2016年1月. |
資料番号 |
MI2015-112 |
発行日 |
2016-01-12 (MI) |
ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します.(許諾番号:10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
PDFダウンロード |
MI2015-112 |