| Popis: |
Purpose: This paper introduces a highly adaptive and fast approach for generating a finite element (FE) discretization mesh for a given multi-compartment human head model obtained through a magnetic resonance imaging (MRI) dataset. The goal is to create accurate deep brain structures for electroencephalographic (EEG) source localization and applications. Methods: We present a recursive application of solid angle labeling for surface segmentation with an adaptive scheme, i.e., a set of smoothing, inflation, and optimization tasks to enhance the mesh quality. Results: The results indicate that our approach can produce a FE mesh with an accuracy greater than 1.0 millimeters, a significant milestone for 3D structure discretization and EEG source localization estimation. Conclusions: Our method, implemented in the Matlab-based Zeffiro Interface, can manage the labeling aspect remarkably well to achieve human head FE meshes with complex deep brain structures using a time-effective parallel computing system. |
