LANDMARK MATCHING ON THE SPHERE USING DISTANCE FUNCTIONS.

Area of Science:

• Medical Imaging
• Computational Anatomy
• Differential Geometry

Background:

• Nonlinear registration of 3D surfaces is crucial for medical imaging applications like tracking anatomical changes and integrating functional MRI data.
• Current methods often involve transforming surfaces to simpler domains (planar or spherical) for landmark matching.

Purpose of the Study:

• To develop and study a novel diffeomorphic matching technique for landmarks on spherical surfaces.
• To improve upon existing methods for cortical surface registration and analysis.

Main Methods:

• The study adapts the level set technique for planar registration to spherical surfaces.
• It incorporates both forward and backward matching terms to ensure the invertibility of the representation.
• The method is demonstrated using the registration of lines on a sphere.

Main Results:

• The proposed technique allows for relaxed matching energy along sulcal landmarks, minimizing distortion.
• It enables the alignment of point and curve landmarks within the same framework as dense scalar fields.
• The method improves upon previous cortical matching approaches.

Conclusions:

• The developed method offers an effective approach for nonlinear registration of 3D surfaces, particularly in neuroimaging.
• It enhances the accuracy and flexibility of anatomical and functional data alignment.
• This work contributes to more robust comparative analyses of brain structures and functions.