Octree grid topology preserving geometric deformable model for three-dimensional medical image segmentation

Area of Science:

• Medical image analysis
• Computational geometry
• Digital topology

Background:

• Topology-preserving geometric deformable models (TGDMs) are crucial for segmenting objects with known topology.
• Current TGDMs face accuracy limitations due to computational grid resolution.
• High-resolution grids increase computational cost and surface mesh size.

Purpose of the Study:

• To introduce a novel framework, Octree-based Topology-Preserving Geometric Deformable Models (OTGDMs), for efficient and accurate segmentation.
• To extend digital topology concepts to balanced octree grids (BOGs) for TGDMs.
• To address implementation challenges for OTGDMs on BOGs.

Main Methods:

• Developed OTGDMs utilizing balanced octree grids (BOGs).
• Extended digital topology definitions and concepts to BOGs for simple point characterization.
• Addressed critical implementation aspects of the OTGDMs framework.

Main Results:

• Demonstrated the performance of OTGDMs using mathematical phantoms.
• Validated the effectiveness of OTGDMs on real medical images.
• Achieved topology-preserving segmentation with improved computational efficiency and manageable surface mesh size.

Conclusions:

• OTGDMs provide an efficient and accurate solution for topology-preserving segmentation.
• The framework effectively overcomes the limitations of traditional TGDMs related to grid resolution.
• The method shows promise for applications in medical image analysis and other fields requiring precise geometric segmentation.