Tracking pylorus in ultrasonic image sequences with edge-based optical flow.

Area of Science:

• Medical imaging
• Biomedical engineering
• Gastroenterology

Background:

• Accurate tracking of the pylorus in ultrasound sequences is crucial for analyzing duodenogastric reflux (DGR).
• Existing methods may lack precision in capturing complex pyloric dynamics.

Purpose of the Study:

• To develop and evaluate a novel joint prediction and segmentation (JPS) method for improved pyloric tracking.
• To enhance the accuracy of pyloric tracking by integrating edge connectivity and spatio-temporal information.

Main Methods:

• The proposed JPS method combines optical flow with active contour models.
• It leverages both edge point connectivity and spatio-temporal information from consecutive frames.
• Performance was evaluated against four other tracking methods using synthetic and real ultrasound data.

Main Results:

• JPS demonstrated superior performance, achieving minimum Hausdorff distance (HD), average distance (AD), and mean edge distance (MED).
• The method also resulted in a smaller edge curvature (EC), indicating smoother tracking.
• JPS showed better agreement with gold standard curves compared to other methods.

Conclusions:

• The JPS method offers enhanced accuracy and smoothness for pyloric tracking in ultrasound images.
• This improved tracking facilitates more precise analysis of duodenogastric reflux.
• JPS represents a significant advancement over existing pyloric tracking techniques.