Probabilistic vessel axis tracing and its application to vessel segmentation with stream surfaces and minimum cost...

Area of Science:

• Medical Imaging
• Computational Anatomy
• Image Processing

Background:

• Accurate vascular segmentation is crucial for diagnosing and treating various medical conditions.
• Existing methods for 3D vascular segmentation often struggle with complex vessel structures like bifurcations and diseased regions.

Purpose of the Study:

• To develop a novel framework for segmenting vessels by analyzing their cross-sections.
• To improve the accuracy and robustness of 3D vascular segmentation, especially in challenging anatomical areas.

Main Methods:

• Probabilistic vessel axis tracing in 3D angiograms.
• Vessel boundary delineation on derived cross-sections using minimum cost path on a graph.
• User guidance integration for continuous tracing through complex regions.
• Tiling of contours to form a vessel boundary surface, followed by deformation and voxelization.

Main Results:

• The proposed method achieves continuous and less jittered vessel axes compared to existing trace-based algorithms.
• The segmentation is robust to noise and accurately delineates vessel boundaries, comparable to manual segmentation.
• The framework successfully handles complex vascular structures including bifurcations, diseased portions, and closely spaced vessels.

Conclusions:

• The novel cross-section-based framework provides accurate and robust 3D vascular segmentation.
• The method's ability to integrate user guidance enhances its applicability in clinical settings.
• This approach offers a significant advancement in automated and semi-automated vascular segmentation techniques.