A Volumetric Approach to Point Cloud Compression, Part II: Geometry Compression.

Area of Science:

• Computer Vision
• 3D Data Processing
• Signal Processing

Background:

• Traditional point cloud compression focuses on discrete points and their attributes.
• Existing methods like MPEG Point Cloud Compression (G-PCC) have limitations in representing complex geometry.
• Continuous function representations offer potential for more efficient data encoding.

Purpose of the Study:

• To introduce and evaluate a novel point cloud compression method using volumetric functions.
• To demonstrate improved geometry compression performance compared to existing standards.
• To represent both geometry and attributes of point clouds using continuous volumetric functions.

Main Methods:

• Utilizing volumetric functions, specifically B-spline wavelets, to represent point cloud data.
• Implicitly representing geometry as the level set of a volumetric function (e.g., signed distance function).
• Coding both geometry and attributes using the derived volumetric function representations.

Main Results:

• Volumetric function-based geometry compression shows superior performance compared to the emerging MPEG G-PCC standard.
• The proposed method effectively encodes both geometric and attribute information.
• Continuous function interpolation allows for accurate representation of point cloud data.

Conclusions:

• Volumetric functions provide a powerful and efficient alternative for point cloud compression.
• This approach advances the state-of-the-art in 3D data compression, particularly for geometry.
• The method holds promise for future standards and applications requiring high-fidelity 3D data representation.