Adaptive Kalman snake for semi-autonomous 3D vessel tracking

Area of Science:

• Medical Imaging
• Image Processing
• Computational Anatomy

Background:

• Accurate 3D vessel segmentation and tracking are crucial for medical diagnosis.
• Existing methods may lack robustness or efficiency in analyzing complex vascular structures.

Purpose of the Study:

• To develop a robust semi-autonomous algorithm for 3D vessel segmentation and tracking.
• To improve the accuracy and efficiency of analyzing computed tomography angiography (CTA) data.

Main Methods:

• Utilizes an active contour model for vessel boundary segmentation on each CTA slice.
• Employs a Kalman filter for tracking vessel boundary variations between slices.
• Incorporates an adaptive control spacing algorithm and block search for refined segmentation and initial contour estimation.

Main Results:

• The proposed algorithm demonstrated superior performance in segmentation and tracking compared to the Chan-Vese (CV) model.
• Receiver operating characteristic analysis confirmed the model's time efficiency and tracking robustness on synthetic and real CTA images.
• Achieved an approximate 20% reduction in processing time.

Conclusions:

• The developed algorithm offers a robust and efficient solution for 3D vessel segmentation and tracking in CTA.
• The method shows significant improvements in accuracy and computational speed, benefiting clinical applications.