| 電子情報通信学会 研究会発表申込システム 講演論文 詳細 |
技報閲覧サービス [ログイン] 技報アーカイブ |
| トップに戻る | 前のページに戻る | [Japanese] / [English] |
| 講演抄録／キーワード | ||
| 講演名 | 2009-01-19 13:30 ［ポスター講演］Multiscale Polyp Segmentation Algorithm in CT Colonography ○June-Goo Lee・Jong Hyo Kim（Seoul National Univ.）・Se Hyung Kim（Seoul National Univ. Hospital）  MI2008-101 |
|
| 抄録 | （和） | An automatic polyp segmentation algorithm in CT colonography is presented. This algorithm can provide quantitative measurement of polyp size which is used as the biomarker of malignancy. Hessian matrix is the square matrix of second partial derivatives of a scalar-valued function and is well known for object recognition in computer vision and medical shape analysis. We applied this multiscale Hessian classification method to polyp segmentation in CTC. The overall segmentation scheme is performed by following procedures. First, multiscale blob-likeness filter output images were produced. Second, the voxel which has the local maximum of blob-likeness value was selected as initial seed point in the following region growing process. Third, the region growing algorithm was applied to make initial polyp region. And the output region was dilated to include outside partial volume effect region. After the polyp region was extracted, the principle component analysis (PCA) was performed to find principle axis. According to this principle axis the sizes of polyps were measured automatically. For evaluation 6 anthropomorphic pig phantoms were prepared. There were 82 artificial polyps with different sizes and shapes. The results of computerized measurement were compared to that of manual ex-vivo measurement with vernier calipers. Linear regression was performed, and the result was 。ーy=2.82+0.79x。ア. Two groups were statistically correlated (p<0.001). |
| （英） | An automatic polyp segmentation algorithm in CT colonography is presented. This algorithm can provide quantitative measurement of polyp size which is used as the biomarker of malignancy. Hessian matrix is the square matrix of second partial derivatives of a scalar-valued function and is well known for object recognition in computer vision and medical shape analysis. We applied this multiscale Hessian classification method to polyp segmentation in CTC. The overall segmentation scheme is performed by following procedures. First, multiscale blob-likeness filter output images were produced. Second, the voxel which has the local maximum of blob-likeness value was selected as initial seed point in the following region growing process. Third, the region growing algorithm was applied to make initial polyp region. And the output region was dilated to include outside partial volume effect region. After the polyp region was extracted, the principle component analysis (PCA) was performed to find principle axis. According to this principle axis the sizes of polyps were measured automatically. For evaluation 6 anthropomorphic pig phantoms were prepared. There were 82 artificial polyps with different sizes and shapes. The results of computerized measurement were compared to that of manual ex-vivo measurement with vernier calipers. Linear regression was performed, and the result was 。ーy=2.82+0.79x。ア. Two groups were statistically correlated (p<0.001). | |
| キーワード | （和） | Polyp segmentation / quantitative measurement / CTC CAD / Hessian matrix / / / / |
| （英） | Polyp segmentation / quantitative measurement / CTC CAD / Hessian matrix / / / / | |
| 文献情報 | 信学技報, vol. 108, no. 385, MI2008-101, pp. 197-201, 2009年1月. | |
| 資料番号 | MI2008-101 | |
| 発行日 | 2009-01-12 (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 | MI2008-101 |
|
| 研究会情報 | |
| 研究会 | MI |
| 開催期間 | 2009-01-19 - 2009-01-21 |
| 開催地（和） | 国立台湾大学(台北市) |
| 開催地（英） | National Taiwan University |
| テーマ（和） | 医用画像一般 |
| テーマ（英） | Medical Imaging |
| 講演論文情報の詳細 | |
| 申込み研究会 | MI |
| 会議コード | 2009-01-MI |
| 本文の言語 | 英語 |
| タイトル（和） | |
| サブタイトル（和） | |
| タイトル（英） | Multiscale Polyp Segmentation Algorithm in CT Colonography |
| サブタイトル（英） | |
| キーワード(1)（和/英） | Polyp segmentation / Polyp segmentation |
| キーワード(2)（和/英） | quantitative measurement / quantitative measurement |
| キーワード(3)（和/英） | CTC CAD / CTC CAD |
| キーワード(4)（和/英） | Hessian matrix / Hessian matrix |
| キーワード(5)（和/英） | / |
| キーワード(6)（和/英） | / |
| キーワード(7)（和/英） | / |
| キーワード(8)（和/英） | / |
| 第1著者 氏名（和/英/ヨミ） | June-Goo Lee / June-Goo Lee / |
| 第1著者 所属（和/英） | Seoul National University, College of Medicine (略称： Seoul National Univ.) Seoul National University, College of Medicine (略称： Seoul National Univ.) |
| 第2著者 氏名（和/英/ヨミ） | Jong Hyo Kim / Jong Hyo Kim / |
| 第2著者 所属（和/英） | Seoul National University, College of Medicine (略称： Seoul National Univ.) Seoul National University, College of Medicine (略称： Seoul National Univ.) |
| 第3著者 氏名（和/英/ヨミ） | Se Hyung Kim / Se Hyung Kim / |
| 第3著者 所属（和/英） | Seoul National University Hospital (略称： Seoul National Univ. Hospital) Seoul National University Hospital (略称： Seoul National Univ. Hospital) |
| 講演者 | 第1著者 |
| 発表日時 | 2009-01-19 13:30:00 |
| 発表時間 | 210分 |
| 申込先研究会 | MI |
| 資料番号 | MI2008-101 |
| 巻番号（vol） | vol.108 |
| 号番号（no） | no.385 |
| ページ範囲 | pp.197-201 |
| ページ数 | 5 |
| 発行日 | 2009-01-12 (MI) |
[研究会発表申込システムのトップページに戻る]