Depth potential function for folding pattern representation, registration and analysis.

Area of Science:

• Computational geometry
• Medical imaging analysis
• Surface analysis

Background:

• Analyzing folding patterns on curved surfaces, such as the human brain's cortical surface, is challenging due to the lack of a natural projection plane.
• Existing methods often rely on scalar functions and struggle with complex geometries.

Purpose of the Study:

• To extend the concept of depth maps for analyzing folding patterns on curved surfaces without a natural projection plane.
• To introduce a novel, computationally efficient method for shape matching and variability analysis of surface folds.

Main Methods:

• Developed a 'depth potential function' that integrates surface curvature information into a point-of-view invariant representation.
• Computed the depth potential function by solving a time-independent Poisson equation.
• Validated the method on synthetic and real human brain cortical surfaces using magnetic resonance imaging data.

Main Results:

• The depth potential function provides a computationally advantageous surface representation, orders of magnitude faster than other methods.
• Shape matching accuracy improved by an average of 11% when using the depth potential function.
• The method effectively analyzes folding patterns on complex curved surfaces.

Conclusions:

• The depth potential function offers a robust and efficient solution for analyzing and matching folding patterns on curved surfaces.
• This approach has significant implications for fields requiring detailed surface analysis, including neuroimaging and computational geometry.