Vascular active contour for vessel tree segmentation.

Area of Science:

• Medical Imaging
• Computer Vision
• Biomedical Engineering

Background:

• Accurate vessel tree segmentation is crucial for diagnosing various cardiovascular and pulmonary diseases.
• Existing segmentation methods often struggle with thin, weak, or complex vascular structures.

Purpose of the Study:

• To develop a novel active contour model for robust and accurate vessel tree segmentation.
• To improve the extraction of both thick and thin vessels from volumetric medical images.

Main Methods:

• A region competition-based active contour model using Gaussian Mixture Models for thick vessels.
• A vascular vector field derived from Hessian matrix eigenanalysis for thin vessel centerline evolution.
• A dual curvature strategy for vessel surface smoothing without shape distortion.

Main Results:

• The proposed model successfully segmented liver, lung, and coronary artery trees from high-resolution CT images.
• Demonstrated superior accuracy and robustness compared to classical active contour models and manual segmentation.
• Achieved effective extraction of both thick and thin vessel segments.

Conclusions:

• The novel active contour model offers a more accurate and robust solution for automatic vessel tree segmentation.
• This method holds significant potential for clinical applications in cardiovascular and pulmonary imaging analysis.