An approach for detecting cerebral artery stenosis based on MRA images

摘要

Cerebral artery stenosis is an important cause and risk factor to ischemic cerebrovascular disease.This paper describes a simple and effective method for the detection of cerebral artery stenosis in magnetic resonance angiography (MRA).Because of the complex structure of the brain blood vessels,the fast and accurate segmentation method is needed.Here,we used the information of the maximum intensity projection (MIP) image to get 3D vascular structure.As skeleton was one of useful measures for charactering region-based shape features,we extracted 3D-Skeleton of blood vessels by using fast marching method (FMM).Finally,the accurate cross-sectional areas based on cerebrovascular skeleton were computed to determine the location and extent of vascular stenosis.The results showed that our method could effectively detect the cerebral artery stenosis.