作者: Yan Xia , Aamer Aziz , QingMao Hu , Wieslaw L. Nowinski
DOI: 10.1007/978-3-540-39903-2_78
关键词:
摘要: This paper describes a rapid and automated method that extracts the fourth ventricle from MR brain images in normal pathological subjects. Anatomical knowledge of has been incorporated into to define region interest (ROI), determine intensity thresholds histogram ROI, locate starting point for 3D growing, extract all parts aqueduct (even if they are disconnected on as well avoid inclusion non-ventricular cerebrospinal fluid regions). The was validated qualitatively 30 MRI studies with variable parameters. best overlap metric between radiology expert 0.9988 worst 0.9621 10 quantitative studies. mean standard deviation were 0.9806 0.0105.