Adaptive Ridge Point Refinement for Seeds Detection in X-Ray Coronary Angiogram

Area of Science:

• Medical Imaging
• Computer Vision
• Biomedical Engineering

Background:

• Tracking-based methods for extracting vascular structures from angiograms require a seed point.
• Accurate seed point detection is crucial for reliable coronary artery segmentation and centerline extraction.

Purpose of the Study:

• To propose a novel, automated seed point detection method for coronary artery segmentation.
• To enhance the accuracy and robustness of seed point detection in angiographic images.

Main Methods:

• Vessel enhancement using Hessian eigenvalue distribution in multiscale space.
• Extraction and mathematical refinement of candidate seed points from enhanced vessel centerlines.
• Robustness improvement using a self-adaptive threshold for ridge point validation.

Main Results:

• The proposed algorithm successfully detects a large number of true seed points on coronary artery branches.
• Achieved higher precision and a greater number of detected seed points compared to traditional methods.
• Demonstrated theoretical feasibility and robustness in clinical angiogram evaluations.

Conclusions:

• The novel method provides accurate, automated seed point detection for coronary artery segmentation.
• It offers significant improvements over traditional algorithms in terms of precision and detection rate.
• Enables clinical applications like vessel segmentation, centerline extraction, and topological identification without human interaction.