Proposal of Atlas-guided Eigen-organ Method for Location Detection of Multi-Organs in Three Dimensional Medical Images(Joint Session 2)

概要

論文の詳細を見る
This paper proposes an approach for the location detection of multi-organs in volumetric medical images. Local eigen-organ spaces are constructed according to different types of training organs. Global eigen-space, which describes the spatial relationship among organs, is also constructed. In this method, each volume of interest in the abdomen CT image is projected into the local eigen-organ spaces and a number of candidate locations are selected, then final decision is made by the global eigen-space. We use the probabilistic atlas of organs to skip locations with low probability and to guide the selection of candidate locations. The proposed method produces an average coincidence ratio of 71.94% for 11 organs which was estimated by leave-one-out method using 10 volumetric abdomen CT images.
社団法人電子情報通信学会の論文
2005-09-15