Near-tubular fiber bundle segmentation for diffusion weighted imaging: segmentation through frame reorientation

Area of Science:

• Neuroimaging
• Medical Image Analysis
• Computational Neuroscience

Background:

• Diffusion Magnetic Resonance Imaging (DW-MRI) is crucial for mapping white matter tracts.
• Accurate segmentation of near-tubular fiber bundles remains a challenge in DW-MRI analysis.
• Existing methods may struggle with complex fiber geometries.

Purpose of the Study:

• To develop an efficient and accurate methodology for segmenting near-tubular fiber bundles from DW-MRI data.
• To simplify the segmentation process by leveraging local and global diffusion information.
• To improve the reliability of white matter tractography.

Main Methods:

• Proposed a novel segmentation methodology for near-tubular fiber bundles in DW-MRI.
• Implemented local reorientation of diffusion information guided by large-scale fiber bundle geometry.
• Employed global statistical modeling of diffusion orientation.
• Utilized a modified convex optimization formulation combining orientation statistics and spatial regularization.

Main Results:

• The proposed segmentation approach simplifies the process by reorienting diffusion information.
• Achieved segmentation through global statistical modeling of diffusion orientation.
• The method demonstrated favorable comparisons against full-brain streamline tractography segmentation.
• The convex optimization formulation effectively integrated orientation statistics and spatial regularization.

Conclusions:

• The developed methodology offers an effective approach for segmenting near-tubular fiber bundles in DW-MRI.
• Local reorientation and statistical modeling significantly simplify the segmentation task.
• This technique shows promise for enhancing the accuracy and efficiency of white matter tractography.