OBJECTIVE: The goal of this paper is to automatically segment perivascular spaces (PVSs) in brain from high-resolution 7T magnetic resonance (MR) images. METHODS: We propose a structured-learning-based segmentation framework to extract the PVSs from high-resolution 7T MR images. Specifically, we integrate three types of vascular filter responses into a structured random forest for classifying voxels into two categories, i.e., PVS and background. In addition, we propose a novel entropy-based sampling strategy to extract informative samples in the background for training an explicit classification model. Since the vascular filters can extract various vascular features, even thin and low-contrast structures can be effectively extracted from noisy backgrounds. Moreover, continuous and smooth segmentation results can be obtained by utilizing patch-based structured labels. RESULTS: The performance of our proposed method is evaluated on 19 subjects with 7T MR images, with the Dice similarity coefficient reaching 66%. CONCLUSION: The joint use of entropy-based sampling strategy, vascular features, and structured learning can improve the segmentation accuracy. SIGNIFICANCE: Instead of manual annotation, our method provides an automatic way for PVS segmentation. Moreover, our method can be potentially used for other vascular structure segmentation because of its data-driven property.
OBJECTIVE: The goal of this paper is to automatically segment perivascular spaces (PVSs) in brain from high-resolution 7T magnetic resonance (MR) images. METHODS: We propose a structured-learning-based segmentation framework to extract the PVSs from high-resolution 7T MR images. Specifically, we integrate three types of vascular filter responses into a structured random forest for classifying voxels into two categories, i.e., PVS and background. In addition, we propose a novel entropy-based sampling strategy to extract informative samples in the background for training an explicit classification model. Since the vascular filters can extract various vascular features, even thin and low-contrast structures can be effectively extracted from noisy backgrounds. Moreover, continuous and smooth segmentation results can be obtained by utilizing patch-based structured labels. RESULTS: The performance of our proposed method is evaluated on 19 subjects with 7T MR images, with the Dice similarity coefficient reaching 66%. CONCLUSION: The joint use of entropy-based sampling strategy, vascular features, and structured learning can improve the segmentation accuracy. SIGNIFICANCE: Instead of manual annotation, our method provides an automatic way for PVS segmentation. Moreover, our method can be potentially used for other vascular structure segmentation because of its data-driven property.
Authors: Diego Marin; Arturo Aquino; Manuel Emilio Gegundez-Arias; José Manuel Bravo Journal: IEEE Trans Med Imaging Date: 2010-08-09 Impact factor: 10.048
Authors: Courtney Berezuk; Joel Ramirez; Fuqiang Gao; Christopher J M Scott; Menal Huroy; Richard H Swartz; Brian J Murray; Sandra E Black; Mark I Boulos Journal: Sleep Date: 2015-06-01 Impact factor: 5.849
Authors: Mark Jenkinson; Christian F Beckmann; Timothy E J Behrens; Mark W Woolrich; Stephen M Smith Journal: Neuroimage Date: 2011-09-16 Impact factor: 6.556
Authors: Daniel L Schwartz; Erin L Boespflug; David L Lahna; Jeffrey Pollock; Natalie E Roese; Lisa C Silbert Journal: Neuroimage Date: 2019-08-25 Impact factor: 6.556
Authors: Mackenzie T Langan; Derek A Smith; Gaurav Verma; Oleksandr Khegai; Sera Saju; Shams Rashid; Daniel Ranti; Matthew Markowitz; Puneet Belani; Nathalie Jette; Brian Mathew; Jonathan Goldstein; Claudia F E Kirsch; Laurel S Morris; Jacqueline H Becker; Bradley N Delman; Priti Balchandani Journal: Front Neurol Date: 2022-04-01 Impact factor: 4.086