複数形状に対応した楕円体検出フィルタを用いた3次元腹部CT像からのリンパ節検出手法

IR HANDLE Web Site 9 References Open Access

Bibliographic Information

Other Title
  • フクスウ ケイジョウ ニ タイオウ シタ ダエンタイ ケンシュツ フィルタ オ モチイタ 3ジゲン フクブ CTゾウ カラ ノ リンパセツ ケンシュツ シュホウ
  • Development of a lymph node detection method from 3D abdominal CT images using multi shape ellipsoidal structure detection filter

Search this article

Abstract

This paper proposes a lymph node detection method which utilizing a multi-shape and multi-scale ellipsoidal structure detection filter. Detection of the enlarged lymph nodes is required in diagnosis of metastatic cancer using a computer aided diagnosis system. Automated detection of the enlarged lymph nodes improves efficiency of cancer diagnosis. This paper describes a method for detecting gastric lymph nodes from abdominal CT images. In CT images, the lymph node has higher CT values than the surrounding regions. Shapes of the lymph nodes are sphere or ellipsoid. We utilized a multi-shape and multi-scale ellipsoidal structure detection filter based on local intensity structure analysis to detect the lymph nodes. A Hessian matrix is calculated from a polynomial which approximates a local CT value structure. We utilize three eigenvalues of the Hessian matrix in the detection filter. In a ellipsoidal structure, values of the eigenvalues correspond lengths of the long and short axes of the ellipsoidal structure. The detection filter outputs high values if the eigenvalues take particular values of the ellipsoidal structure. Experimental results using 6 cases of CT images showed that our method can detect 69.2% of lymph nodes.

IEICE Technical Report;MI2011-123

Journal

References(9)*help

See more

Keywords

Details 詳細情報について

Report a problem

Back to top