Related Experiment Video
Updated: Jul 10, 2025
15:26
3D Modeling of the Lateral Ventricles and Histological Characterization of Periventricular Tissue in Humans and Mouse
Published on: May 19, 2015
14.1K
RAPID BRAIN MENINGES SURFACE RECONSTRUCTION WITH LAYER TOPOLOGY GUARANTEE
Peiyu Duan1,2, Yuan Xue3, Shuo Han1
1Department of Biomedical Engineering, Johns Hopkins School of Medicine, USA.
Proceedings. IEEE International Symposium on Biomedical Imaging
|November 22, 2023
Summary
Convolutional neural networks (CNNs) reconstruct brain meningeal layers from MR images. This method accurately measures intracranial volume (ICV) and subarachnoid space (SAS), aiding neurodegenerative disease research.
Area of Science:
- Neuroimaging
- Medical Image Analysis
- Computational Anatomy
Background:
- The meninges (pia, arachnoid, dura) are crucial brain layers.
- Reconstructing meningeal layers aids in studying neurodegenerative diseases and aging.
Purpose of the Study:
- To develop a novel method using CNNs for accurate meningeal layer reconstruction from MR images.
- To enable precise computation of intracranial volume (ICV) and subarachnoid space (SAS).
Main Methods:
- Convolutional Neural Networks (CNNs) predicted signed distance functions (SDFs) for meningeal surfaces.
- Marching cubes algorithm generated continuous surface representations.
- Volumetric analysis was performed on healthy and multiple sclerosis (MS) subjects.
Main Results:
- The CNN-based method produced smoother, more accurate surfaces than deformable models, with reduced computation time.
- ICV and SAS volumes showed significant correlations with sex and age in healthy and MS subjects.
Conclusions:
- CNNs offer an efficient and accurate approach for meningeal surface reconstruction in neuroimaging.
- This method facilitates volumetric analysis for understanding brain changes in aging and disease.
