Mesh Patch Sub-Division for 3D Geometry Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for compressing 3D meshes are not suitable for transmitting the connectivity of points, which is essential for efficient mesh compression, and existing solutions for 3D mesh encoding are inefficient and lack support for attributes of triangle faces.
Innovation Solution
The proposed solution involves post-processing decoded meshes by modifying received triangles per patch to improve mesh geometry. This is achieved through methods such as triangle centroid implementation, edge midpoint implementation, vertex split implementation, and marching cubes, which generate new triangles aligned with the original surface using high-resolution depth maps.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If V-PCC projection method is used to compress 3D point clouds, then compression efficiency is improved, but the method cannot transmit connectivity information required for 3D mesh compression
Solution Approach 1:
The patent segments the mesh into multiple patches, encoding each patch independently with its own connectivity information. This allows connectivity to be preserved within each patch while maintaining compression efficiency, resolving the contradiction between compression and connectivity transmission.
Solution Approach 2:
The patent adds a connectivity dimension to the existing V-PCC projection framework by incorporating face lists and vertex lists that describe mesh topology. This transforms the 2D projection approach into a 3D-aware representation that preserves both compression efficiency and connectivity information.
2Adaptability or versatility
If RAW patch data is used to encode vertices position, then sparse point clouds can be encoded, but encoding efficiency decreases and triangle face attributes are missing
Solution Approach 1:
The patent merges the advantages of RAW patch encoding (sparse point cloud support) with structured mesh encoding (efficiency and attribute preservation). By combining face lists, vertex lists, and position data in a unified format, it achieves both sparse cloud handling and efficient encoding with complete attribute information.
3Productivity
If mesh is decoded without post-processing, then decoding speed is maintained, but mesh geometry quality is insufficient
Solution Approach 1:
The patent performs preliminary mesh reconstruction to generate an initial mesh, then applies post-processing operations (subdivision, smoothing, refinement) to improve geometry quality. This two-stage approach maintains decoding speed by preparing data in advance while achieving high geometric precision through targeted post-processing.
Data Source
AI summary
Ways to post-process a decoded mesh and modify received triangles per patch to improve the mesh geometry are described herein. Since the transmitted geometry contains the high-resolution surface information, the information is able to be applied to the mesh reconstruction operation to generate triangles that are aligned with the original surface. Methods include generating new triangles by splitting the received triangles' edges according to their size, by inserting new vertices at the triangle's centroids, by splitting the vertices, and by performing marching cubes in surfaces defined by the geometry images.