Automated segmentation of the cerebellar lobules using boundary specific classification and evolution

Area of Science:

• Neuroimaging
• Computational Anatomy
• Medical Image Analysis

Background:

• The cerebellum is crucial for motor control, speech, and balance.
• Accurate cerebellar lobule delineation is essential for volumetric studies.
• Current automated methods fail to adequately delineate cerebellar lobules.

Purpose of the Study:

• To develop an automated method for accurate cerebellar lobule delineation.
• To introduce the Automatic Classification of Cerebellar Lobules Algorithm using Implicit Multi-boundary evolution (ACCLAIM).

Main Methods:

• ACCLAIM utilizes a multiple object geometric deformable model (MGDM).
• Each lobule boundary evolves under unique level set speeds derived from a trained classifier.
• The method employs implicit multi-boundary evolution for precise segmentation.

Main Results:

• ACCLAIM demonstrates superior performance compared to atlas-based and multi-atlas fusion techniques.
• The algorithm achieves accurate and consistent delineation of individual cerebellar lobules.
• Classifier-derived speeds enhance the accuracy of boundary evolution.

Conclusions:

• ACCLAIM provides a robust and accurate automated solution for cerebellar lobule segmentation.
• This method advances volumetric analysis of the cerebellum.
• ACCLAIM is a significant improvement for neuroimaging research involving cerebellar structures.