R-centipedeモデルを用いたトモグラフィー電子顕微鏡像からの輪郭抽出の高速化手法

概要

論文の詳細を見る
Extracting structures of interest and their parts from medical and Electron-Microscope (EM) tomography slice images for 3D visualization is a time-consuming process. The problem is to extract the boundaries of the structures of interest from each slice image as fast as possible and preferably with no or minimal user interface. This article presents a class of region-based deformable contour models called restructuring centipedes, R-centipedes. On the contrary to the topologically flexible edge-based snakes and the implicit level-set snakes, R-centipedes and their two parallel versions which are the main focus of this article, 1) do not rely on any image preprocessing such as the isotropic Gaussian smoothing and/or threshold setting needed for removing spurious and weak edges, and 2) don't get stuck with textural artifacts which are abundantly found in EM tomography images. The presented R-centipedes have several other desirable features, including geometrical flexibility, high speed, accuracy and insensitivity to their initial positioning with respect to the structures of interest.