A monocentric centerline extraction method for ring-like blood vessels.

Area of Science:

• Medical Imaging
• Computational Biology
• Biomedical Engineering

Background:

• Accurate blood vessel centerline extraction is crucial for analyzing vascular diseases.
• Existing methods struggle with complex structures like ring-like vessels and multi-voxel widths.

Purpose of the Study:

• To develop a novel monocentric centerline extraction method for ring-like blood vessels.
• To address limitations of previous methods, specifically failure on ring-like structures and multi-voxel width.

Main Methods:

• Generating multiple centerlines from randomly sprinkled seed points.
• Employing multi-centerline fusion and local maximum distance from boundary to repair and supplement centerlines.
• Utilizing monocentric processing to ensure a single-voxel width for the extracted vascular centerline.

Main Results:

• The proposed monocentric method demonstrated superior performance compared to Wan et al.'s method and topological thinning.
• Visual inspection and quantitative assessments confirmed the method's effectiveness on synthesized and real MR brain image data.
• The method successfully handled ring-like structures and eliminated multi-voxel width issues.

Conclusions:

• The proposed monocentric centerline extraction method is effective for complex, ring-like blood vessels.
• This advancement improves the accuracy and reliability of quantitative analysis in vascular disease research.
• The method offers a significant improvement over existing techniques for vascular centerline extraction.