V3C Mesh Compression Syntax for Decoder-Side UV Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for compressing 3D meshes in V-PCC are inefficient for sparse meshes and lack mechanisms to transmit connectivity, and existing solutions like UVAtlas require significant optimization time and are limited to point clouds.
Innovation Solution
Implementing a method that includes atlas mapping processing, connectivity processing, and vertex position processing to generate (u, v) coordinates, using a displacement sub-bitstream to carry displacement information, and reconstructing meshes based on displaced vertices and connectivity, while bypassing certain V-DMC methods and utilizing texture attribute maps as V3C components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If UVAtlas is used for atlas mapping, then texture map generation quality is improved, but optimization time increases significantly
Solution Approach 1:
The patent pre-calculates and stores mapping parameters (projection matrices, patch configurations) during encoding, so the decoder can directly generate (u,v) coordinates without performing time-consuming optimization. This shifts the computational burden to the encoding phase, enabling real-time or near-real-time decoding.
Solution Approach 2:
Instead of performing complex UVAtlas optimization during decoding, the patent uses simplified projection-based coordinate generation that replicates the essential mapping functionality. The decoder generates (u,v) coordinates through direct projection formulas rather than iterative optimization, achieving acceptable quality with minimal computation.
2Productivity
If V-PCC projection method is used, then encoding efficiency for point clouds is improved, but mechanism for transmitting mesh connectivity is lost
Solution Approach 1:
The patent combines point cloud projection methodology with mesh connectivity encoding. The same projection framework used for efficient point cloud encoding is extended to handle mesh data, allowing both geometry and connectivity to be transmitted through unified V3C syntax structures.
Solution Approach 2:
The V3C syntax structures are designed to be multi-functional, serving both point cloud and mesh compression needs. The atlas mapping and projection mechanisms work universally for both data types, while additional syntax elements enable connectivity transmission when mesh data is present.
3Adaptability or versatility
If all vertices are encoded as RAW data, then sparse mesh encoding is enabled, but encoding efficiency decreases and attribute information is lost
Solution Approach 1:
The patent applies different encoding strategies to different parts of the mesh data based on local characteristics. Vertices in sparse regions can be encoded as RAW data when needed, while other vertices use predictive or differential encoding. This localized approach maintains sparse mesh capability while improving overall efficiency.
Solution Approach 2:
The encoding method dynamically selects between RAW encoding and other strategies based on mesh density and characteristics. The system adapts the encoding approach for different vertex groups, using RAW encoding only where necessary for sparse regions while applying more efficient methods elsewhere.
4Adaptability or versatility
If V3C syntax extension is implemented, then mesh compression functionality is added, but syntax complexity increases
Solution Approach 1:
The V3C syntax extension uses existing universal structures (atlas mapping, projection parameters) that serve both point cloud and mesh compression. By reusing established syntax elements rather than creating new ones, the extension adds mesh functionality while minimizing syntax complexity increases.
Solution Approach 2:
The syntax extension is segmented into optional elements that are only present when mesh data is being encoded. Connectivity information, mesh-specific parameters, and related syntax elements are conditionally included, allowing decoders to ignore mesh-specific syntax when only processing point clouds, thus reducing effective complexity.
Data Source
AI summary
In the current implementation of V-DMC, the (u, v) coordinates are generated using Microsoft UVAtlas and they, together with the 3D positions and the topology, are carried in the base mesh sub-bitstream. High level syntax structures described herein support projection-based atlas map generation, and the means to derive the (u, v) coordinates on the decoder side using V3C syntax structure extensions. In comparison with previous implementations and in order to preserve the current V3C geometry bitstream concept, a separate sub-bitstream referred to hereby as vertex property sub-bitstream is used to carry displacement information.