Brain MRI Segmentation with Multiphase Minimal Partitioning
Author Information
Author(s): Elsa D. Angelini, Ting Song, Brett D. Mensh, Andew F. Laine
Primary Institution: Ecole Nationale Supérieure des Télécommunications, Groupe des Ecoles des Télécommunications, CNRS UMR 5141, Paris, France
Hypothesis
Can a multiphase three-dimensional deformable model improve the segmentation of brain MRIs compared to traditional methods?
Conclusion
The multiphase three-dimensional level set method showed very high quality and stability in segmenting brain MRIs, outperforming traditional methods.
Supporting Evidence
- The multiphase level set method achieved high accuracy in segmenting white matter, gray matter, and cerebrospinal fluid.
- Statistical comparisons showed significant differences in performance between the multiphase method and traditional methods.
- Random initialization improved the robustness of the segmentation process.
Takeaway
This study created a new way to look at brain scans that helps doctors see different parts of the brain more clearly, using a smart computer program.
Methodology
The study compared four segmentation methods on ten clinical T1-weighted MRIs, using a multiphase level set framework for the new method.
Potential Biases
The segmentation methods may be influenced by the inherent inhomogeneities in MRI data.
Limitations
The study focused only on cortical structures and did not include subcortical structures.
Participant Demographics
Healthy young volunteers were used for the MRI data.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website