ベクトル集中度フィルタを用いたMRA画像における脳動脈瘤の検出法

概要

論文の詳細を見る
The detection of unruptured aneurysms is major subject in magnetic resonance angiography(MRA). However accurate detection of aneurysms is difficult because of the overlap between the aneurysms and the adjacent vessels on maximum intensity projection images. Therefore, the purpose of this study was to develop a computer-aided diagnosis (CAD)scheme for the detection of unruptured aneurysms. First, we segmented the vessels regions using gray-level thresholding and region growing techniques. The gradient concentration(GC)filter was then applied to the segmented vessel regions for the enhancement of aneurysms. The initial candidates were determined in the GC image with a gray-level threshold. For removal of false positives(FPs), we determined 4 features, i.e., volume, sphericity, mean value of GC image, and maximum value of GC image. The rule-based schemes and a quadratic discriminant analysis with 4 features were employed for distinguishing aneurysms and FPs. The sensitivity for detection of aneurysms was 93.8%(15/16)with 4.06 FPs per patient. Our CAD scheme will be useful in assisting radiologists and/or neurosurgeons for identifying aneurysms in MRA images.
医用画像情報学会の論文
2007-10-01