Active Geodesics: Region-based Active Contour Segmentation with a Global Edge-based Constraint.

Area of Science:

• Computer Vision
• Image Processing
• Computational Geometry

Background:

• Traditional geodesic active contour models apply geodesic constraints only at the final state.
• Balancing edge-based and region-based energy terms often requires arbitrary weight adjustments.

Purpose of the Study:

• To present a novel active geodesic contour model that enforces geodesic properties throughout the entire evolution.
• To develop a region-based energy minimization model that naturally incorporates edge information without arbitrary weighting.

Main Methods:

• Constraining the active contour to be a geodesic with respect to a weighted edge-based energy during its entire evolution.
• Optimizing a purely region-based energy function using edge information as a geodesic constraint.

Main Results:

• The proposed model achieves local optimality inherently due to the continuous geodesic constraint.
• A new class of active contours is generated, exhibiting both local and global behaviors responsive to user interaction.
• Demonstrated a clear relationship between the new model and traditional minimal path methods.

Conclusions:

• The active geodesic contour model offers an improved approach to image segmentation by naturally integrating region and edge information.
• This method simplifies model construction and enhances responsiveness to user input compared to traditional active contour models.