Deep-learning based Significant Stenosis detection from Multi-planar Reformatted Images of traced Intracranial arteries
Ke Han · 2020
Purpose To automatically detect significant intracranial artery stenosis in Computed Tomography angiography (CTA) images. Materials and Methods Centerline Extraction We adopted a skull-striping method[1] to extract brain regions from CTA images and trace arteries of interest using iCafe[2]. To reduce spurious branches or fragments, we designed an optimizing algorithm based on