MISC

2008年1月25日

Electronic cleansing for CT colonography using stool tagging method based on boundary accuracy classification (医用画像)

電子情報通信学会技術研究報告. MI, 医用画像
  • Aluwee Sayed Ahmad Zikri Bin Sayed
  • ,
  • Yasutomo Motokatsu
  • ,
  • Kubo Mitsuru
  • ,
  • KAWATA Yoshiki
  • ,
  • NIKI Noboru
  • ,
  • UENO Junji
  • ,
  • NISHITANI Hiromu

107
461
開始ページ
131
終了ページ
138
記述言語
英語
掲載種別
出版者・発行元
一般社団法人電子情報通信学会

Computed tomography (CT) colonography has gained widespread multi-disciplinary interest as an evolving non-invasive colorectal evaluation screening examination, with the potential of improved patient compliance. However the disposal that makes inside the colon empty as a pretreatment preparation make patient feel very uncomfortable. CT colonography using stool tagging method known as the second major method of examination wherein bowel cleansing is needed. In order to present good electronic bowel cleansing the boundary classification must be accurate. The purpose of our study is to solved the partial volume effect that may occur in the electronic bowel cleansing based on the boundary accuracy classification by consider the four difficult condition of 3-D area as a key point in our study; (a) air, oral contrast material and colon wall are interfaced, (b) the thin colon wall is interfaced between the air and oral contrast material, (c) little oral contrast material is attached to colon wall and (d) little amount of stool has surfaced on the oral contrast material. Then we will display the cleansing result using volume rendering technique for the virtual colonoscopy display. We evaluate our method using 5 set data case on supine position scanning. Evaluation is based on four difficult condition of 3D. We get good result classification for condition (a) and condition (c), average of 91.7% condition (a) and average of 87.1% for condition (c) for 5 set data cases.

リンク情報
CiNii Articles
http://ci.nii.ac.jp/naid/110006623279
CiNii Books
http://ci.nii.ac.jp/ncid/AA11370335
URL
http://id.ndl.go.jp/bib/9378436
ID情報
  • ISSN : 0913-5685
  • CiNii Articles ID : 110006623279
  • CiNii Books ID : AA11370335

エクスポート
BibTeX RIS