Geometry Image Chart Linking for 3D Scene Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing geometry image representations, such as multi-chart geometry images, face limitations in linking charts together seamlessly, preventing arbitrary movement and reconstruction of 3D scenes due to independent chart representations and distortion artifacts.
Innovation Solution
A method and device for processing geometry images that compute a reference direction from mesh coordinates, compare it with candidate directions, and select a corresponding pixel on the border of adjacent charts to establish connections between charts, allowing for seamless reconstruction of the 3D scene by determining the most parallel candidate direction and adjusting tangents for consistent orientation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If multi-chart geometry images are used to limit distortions, then manufacturing precision is improved, but device complexity increases due to multiple independent charts requiring border zippering algorithms
Solution Approach 1:
The patent merges multiple independent charts into a unified continuous structure by establishing connectivity relationships between adjacent charts. Border pixels of neighboring charts are linked through computed directions and tangent matching, allowing seamless traversal across chart boundaries while maintaining low distortion benefits of multi-chart representation.
Solution Approach 2:
The patent introduces border pixels as intermediary elements that mediate between adjacent charts. These border pixels contain directional information and tangent data that enable continuous navigation across chart boundaries, acting as connectors that resolve the independence issue while preserving the multi-chart distortion advantages.
2Device complexity
If charts are represented independently to simplify data structure, then device complexity is reduced, but adaptability deteriorates because arbitrary movement between charts is impossible
Solution Approach 1:
The patent introduces dynamic connectivity to the previously static independent chart structure. By computing reference directions from mesh coordinates and comparing them with candidate directions at border pixels, the system enables dynamic navigation and arbitrary movement between charts while maintaining the simplicity of the underlying data structure.
Solution Approach 2:
The patent changes the parameter representation at border pixels by adding directional information and tangent data. This parameter enrichment allows the system to determine connectivity and enable arbitrary movement between charts without fundamentally altering the overall data structure organization.
3Measurement precision
If border zippering algorithm is applied to ensure continuity, then measurement precision is improved, but productivity decreases due to additional processing steps required
Solution Approach 1:
The patent performs preliminary computation of reference directions and candidate directions at border pixels during the geometry image creation process. By pre-establishing these directional relationships and tangent matching information, the system achieves continuity accuracy without requiring additional processing steps during runtime or reconstruction phases.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Method and device for processing a geometry image generated from a mesh associated with a 3D scene and comprising a plurality of charts that each represents a part of the 3D scene, the method comprising computing a reference direction (21) from mesh coordinates associated with a current pixel (203) and mesh coordinates associated with a first pixel (213), the first pixel (213) corresponding to a pixel of the border of the first chart located along a determined direction (20) having as origin the current pixel (203); comparing the reference direction (21) with a set of candidate directions (22, 23, 24, 25, 26) being each computed from the mesh coordinates associated with the at least one current pixel (203) and indirection mesh coordinates associated with one pixel (222 to 226; 212 to 214) of the neighbourhood of the first pixel (213), the indirection mesh coordinates corresponding to mesh coordinates of a pixel of the border a second chart corresponding to the pixel of the neighbourhood of the first pixel; and selecting a pixel of the border of the second chart according to the comparison result.