単純X 線CT 画像における肝臓血管の自動抽出法の初期検討

概要

論文の詳細を見る
We propose a new method for liver vessel extraction from 3-D non-contrast X-ray CT images and evaluate the performance of the method. First, the X-ray CT images are smoothed to reduce the noise using a median filter following by the extraction of liver region, and then the voxels in the range of pre-estimated liver-vessel grayscale levels are enhanced. Moreover, the line components in the liver are enhanced using multi-scale filtering. Finally, the liver vessels are extracted by using binalization. Twenty cases including two cases of contrast-enhanced CT images were employed to validate the proposed method by visual assessment between the vessel-enhanced result from non-contrast CT images and the result of vessels from contrast-enhanced CT images. It was found from our preliminary study that our proposed method may be useful to enhance liver vessels in non-contrast CT images.
医用画像情報学会の論文