3D Mesh Compression Using Projection-Based Surface Patches
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for compressing 3D meshes, such as V-PCC, are inadequate for encoding connectivity and perform poorly with sparse meshes, as they rely on dense mesh assumptions and lack efficient encoding of triangle attributes.
Innovation Solution
A projection-based approach that segments meshes into surface patches following connectivity, projects these patches to 2D, and encodes vertex positions and connectivity, allowing for efficient compression of both dense and sparse meshes by utilizing triangle patch data units and color expansion techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If V-PCC projection-based method is used for mesh compression, then compression efficiency is improved, but connectivity encoding capability is lost
Solution Approach 1:
The mesh is segmented into surface patches that follow connectivity, with each patch containing both projected vertex data and connectivity information. This segmentation allows the projection-based compression to maintain connectivity encoding by organizing vertices and triangles into coherent patch structures rather than treating them as disconnected points.
Solution Approach 2:
The connectivity information is nested within the patch structure alongside the projected vertex positions. The patch data unit contains both the 2D projected coordinates and the 3D connectivity relationships, allowing connectivity to be encoded within the same compression framework without requiring separate encoding passes.
2Ease of manufacture
If dense mesh assumption is made, then projection-based encoding works, but sparse mesh compression performance deteriorates
Solution Approach 1:
The patch generation process dynamically adapts to mesh density by adjusting the segmentation strategy. For dense meshes, larger patches can be formed, while for sparse meshes, the algorithm automatically creates smaller patches or uses alternative encoding paths, allowing the same framework to handle varying density levels without manual configuration.
Solution Approach 2:
The encoding method changes parameters such as patch size, sampling density, and triangle grouping based on the input mesh characteristics. This allows the system to adjust the projection and encoding parameters dynamically to match the sparsity or density of the input mesh, maintaining efficiency across different mesh types.
3Device complexity
If triangle attributes are not encoded, then compression is simplified, but rendering quality deteriorates
Solution Approach 1:
Different triangle attributes are encoded with varying precision and complexity depending on their importance and the local requirements of the mesh. Critical attributes like position and connectivity are encoded with high precision, while less critical attributes may use compressed or approximate representations, allowing quality optimization without uniform complexity increase.
Solution Approach 2:
The encoding includes essential triangle attributes for accurate rendering while optionally including additional attributes based on needs. The method can encode minimum required data for basic rendering or expand to include enhanced attributes for higher quality, providing flexibility between compression simplicity and rendering fidelity.
Data Source
AI summary
A method of compressing meshes using a projection-based approach, leveraging and expanding the tools and syntax generated for projection-based volumetric content compression is described. The mesh is segmented into surface patches, with the difference that the segments follow the connectivity of the mesh. The dense mesh compression utilizes 3D surface patches to represent connected triangles on a mesh surface and groups of vertices to represent triangles not captured by surface projection. Each surface patch (or 3D patch) is projected to a 2D patch, whereby for the mesh, the triangle surface sampling is similar to a common rasterization approach. For each patch, position and connectivity of the projected vertices are kept. The sampled surface resembles a point cloud and is coded with the same approach used for point cloud compression. The list of vertices and connectivity per patch is encoded, and the data is sent with the coded point cloud data.