Static/Dynamic Filtering for Mesh Geometry.

Area of Science:

• Computer Graphics
• Signal Processing
• Geometric Modeling

Background:

• Joint bilateral filters are used for feature-preserving signal smoothing in geometry processing.
• Existing methods use static guidance (inconsistent) or dynamic guidance (noise-sensitive).

Purpose of the Study:

• To propose a novel geometry filtering technique, the static/dynamic filter, inspired by image filtering advancements.
• To achieve state-of-the-art results by utilizing both static and dynamic guidances.

Main Methods:

• The filter is based on nonlinear optimization enforcing signal smoothness while preserving scale-specific features.
• An efficient iterative solver unifies existing static and dynamic guidance filters.
• Applied to mesh face normals and vertex positions for scale-aware, feature-preserving geometry filtering.

Main Results:

• The static/dynamic filter effectively smooths signals on mesh surfaces, including geometry and texture colors.
• Demonstrated state-of-the-art performance in mesh denoising, geometry feature enhancement, and texture color filtering.
• The unified solver efficiently handles various geometry processing tasks.

Conclusions:

• The proposed static/dynamic filter offers a robust and effective solution for geometry processing tasks.
• It outperforms existing methods by leveraging the strengths of both static and dynamic guidance.
• The technique is versatile, applicable to mesh geometry and texture data for enhanced visual quality.