A novel active contour model using local and global statistics for vessel extraction.

Area of Science:

• Medical Imaging
• Computer Vision
• Biomedical Engineering

Background:

• Vessel extraction is crucial in clinical practice.
• Current deformable models often rely on image gradients or global intensity statistics.
• Global statistic models struggle with uneven intensity distributions and lack local adaptation.

Purpose of the Study:

• To propose a novel level set based active contour model for enhanced vessel extraction.
• To address the limitations of existing models in handling local intensity variations.
• To improve the accuracy and robustness of vessel segmentation in medical images.

Main Methods:

• Developed a novel level set based active contour model.
• Integrated a simple local function with global statistical information.
• Utilized curve evolution and stopping criteria based on combined local and global data.

Main Results:

• The proposed model demonstrates improved performance in vessel extraction.
• The combination of local and global information enhances adaptability to uneven intensity distributions.
• Achieved better segmentation results compared to models relying solely on global statistics.

Conclusions:

• The novel level set active contour model offers superior vessel extraction capabilities.
• Combining local and global image statistics effectively addresses limitations of traditional methods.
• This approach holds promise for advancing clinical image analysis and diagnosis.