Voxel-Wise 3D Mesh Texture Patch Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for generating texture maps for 3D meshes require complex 2D mapping operations, which are computationally intensive and burdensome for high-resolution scene images, leading to inefficiencies in hardware operation and rendering quality.
Innovation Solution
A method and apparatus for generating voxel-wise 3D mesh texture patches using machine learning, which involves generating a truncated signed distance function (TSDF) volume from a multi-view image, allocating texture patches to each voxel, and optimizing these patches through joint learning with a decoder network.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If 2D mapping operation is performed to generate texture map, then texture mapping can be achieved, but hardware burden and computational complexity increase significantly
Solution Approach 1:
The patent divides the 3D mesh surface into multiple texture patches, each representing a local region. Instead of performing complex global 2D mapping, the system segments the texture mapping task into independent patch-level operations, significantly reducing hardware burden while maintaining texture mapping capability
Solution Approach 2:
The patent transitions from traditional 2D UV mapping to a 3D voxel-based approach where texture patches are directly associated with volumetric elements. This dimensional shift eliminates the need for complex 2D parameterization while preserving texture information through spatial correspondence in 3D space
2Manufacturing precision
If 2D mapping operation is performed on high-resolution scene images, then complete texture coverage is achieved, but processing time and computational resources increase
Solution Approach 1:
By segmenting the scene into multiple view images and corresponding texture patches, the system processes only relevant regions from each view. This avoids processing the entire high-resolution scene globally, reducing processing time while maintaining complete texture coverage through aggregated patch information
Solution Approach 2:
The patent performs preliminary processing by generating texture patches from multiple views before final composition. This pre-processing step organizes texture information in advance, enabling faster final assembly and reducing overall processing time for high-resolution scenes
Data Source
AI summary
The present disclosure relates to a method and device for generating a voxel-by-voxel three-dimensional mesh texture patch. A method for generating a voxel-wise texture patch according to an embodiment of the present disclosure may comprise: generating a truncated signed distance function (TSDF) volume based on a multi-view image; generating and allocating a texture patch for each voxel in the TSDF volume; and performing optimization of the texture patch based on joint learning for the texture patch and a specific decoder to which the texture patch is input.