Design And Implementation Of Internal Carotid Artery Extraction System

Cerebrovascular segmentation is a key step for cerebral aneurysm detection, pathology quantifICAtion and treatment planning. Existing segmentation methods either requires laborious manually tuning of various parameters to optimize segmentation results or fail to achieve accurate results. Especially,the Internal Carotid Artery.This paper describes the threshold segmentation with adaptive contrast enhancement and new framework base on active contour model. The two frameworks can automatically segment the Internal Carotid Artery in the Circle of Willis with a high accuracy rate and little manual configurations. The two key contribution of proposed framework are: 1) Auto detection of the Internal Carotid Artery region. 2) Local Contrast Enhancement with Guidence improves contrast in challenging regions, 3)The improvement of the framework base on active contour.An evaluation on 6 clinical CT datasets demonstrates that can achieve 89% average accuracy respect to the ground-truth, which is 20%~30% greater than other state-of-the-art methods with manually optimized parameters.