Mesh Vertex Grouping for Dynamic Mesh Position Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing mesh compression standards struggle to efficiently handle dynamic meshes with time-varying connectivity and attribute maps, particularly under real-time constraints, and do not support constant connectivity dynamic meshes effectively.
Innovation Solution
A method for grouping mesh vertices based on predetermined constants, estimating vertex positions using reference frames, and encoding prediction modes and residues using entropy, arithmetic, or context-based coding to compress vertex positions efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If existing mesh compression standards are used, then mesh data can be compressed, but compression efficiency is insufficient for dynamic meshes with time-varying connectivity and attribute maps
Solution Approach 1:
The mesh is divided into multiple vertex groups, where each group shares common prediction parameters. This segmentation allows the patent to handle dynamic meshes with time-varying connectivity by treating different regions independently, improving compression efficiency while adapting to local variations in mesh dynamics.
Solution Approach 2:
The patent implements dynamic prediction by allowing prediction parameters to vary across different vertex groups and time frames. The system adapts to time-varying connectivity by updating prediction modes and parameters based on current frame characteristics, enabling efficient compression of dynamic meshes.
2Quantity of substance
If vertex positions are compressed using reference frames, then data volume is reduced, but processing complexity increases
Solution Approach 1:
By dividing vertices into groups with shared prediction parameters, the patent reduces the overall processing complexity. Instead of computing individual predictions for all vertices, the system performs prediction once per group, significantly reducing computational load while maintaining compression effectiveness.
Solution Approach 2:
The patent changes the prediction parameters dynamically based on vertex group characteristics and frame differences. By adapting parameters such as prediction mode and reference frame selection to local conditions, the system achieves high compression ratios without requiring excessively complex processing for each vertex.
3Loss of information
If prediction modes are encoded based on estimation errors, then compression efficiency improves, but encoding complexity increases
Solution Approach 1:
The patent evaluates estimation errors at the vertex group level rather than individual vertex level. This segmentation approach reduces the number of prediction mode decisions that need to be encoded, lowering encoding complexity while still capturing the essential variations needed for high compression efficiency.
Solution Approach 2:
The system dynamically selects prediction modes based on estimated error characteristics of each vertex group. By changing parameters adaptively based on local error patterns rather than using a fixed encoding scheme, the patent achieves high compression efficiency with manageable encoding complexity.
Data Source
AI summary
A plurality of vertices corresponding to a mesh of a current frame that is positionally tracked is divided into a plurality of vertex groups based on a predetermined constant integer. The plurality of vertex groups includes a first vertex group. A first position of a current vertex in the first vertex group is estimated based on a second position of a reference vertex in a reference frame that is different from the current frame. An estimation error indicating a difference between the reference vertex in the reference frame and the current vertex in the first vertex group is determined. A prediction mode of vertices in the first vertex group is encoded based at least on the estimation error between the reference vertex and the current vertex. Prediction information of the vertices in the first vertex group is generated based on the encoded prediction mode of the vertices of the first vertex group.