Dynamic Mesh Encoding via Base Mesh Segmentation and Displacement Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for encoding and decoding dynamic meshes face challenges in compression efficiency due to the large number of vertices, leading to deteriorated mesh restoration quality when using static compression codecs like AVC, HEVC, or VVC.
Innovation Solution
The method involves creating a base mesh through mesh decimation, subdividing it, and encoding displacement information, with adaptive mesh subdivision types and in-loop filtering to improve data efficiency and quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If static compression codecs (AVC, HEVC, VVC) are used to compress dynamic mesh, then compression is achieved, but mesh restoration quality deteriorates due to the large number of vertices
Solution Approach 1:
The dynamic mesh is segmented into a base mesh (low-frequency component) and displacement information (high-frequency component). The base mesh is obtained through mesh decimation, and displacement information represents the difference between the original mesh and the base mesh. This segmentation allows independent encoding of different frequency components, improving both compression efficiency and restoration quality.
Solution Approach 2:
The patent transforms the mesh representation from direct vertex coordinates to a two-dimensional representation in the frequency domain (base mesh + displacement). This dimensional transformation enables more efficient compression by separating the mesh into structurally important base information and detailed displacement information, which can be encoded with different precision levels.
2Quantity of substance
If mesh decimation is applied to reduce vertex count, then data size is reduced, but mesh detail information is lost
Solution Approach 1:
The patent extracts the essential structural information of the mesh into the base mesh through decimation, and separately extracts the detailed geometric information into displacement information. By taking out and independently encoding the displacement field, the patent preserves mesh details that would otherwise be lost in simple decimation, allowing high-quality reconstruction from the compressed representation.
3Manufacturing precision
If multiple mesh subdivision types are used for different iterations, then restoration quality is improved, but encoding complexity increases
Solution Approach 1:
The patent introduces dynamic adaptability in the encoding process by allowing the encoder to select different mesh subdivision types (e.g., 4-vertex, 6-vertex, or 8-vertex subdivision) for different refinement iterations based on local mesh characteristics. This dynamic selection is controlled through syntax elements that indicate the subdivision type for each iteration, enabling quality optimization without requiring complex manual configuration.
Solution Approach 2:
The patent changes the subdivision parameters (number of vertices per face, refinement pattern) dynamically during the encoding process. Different subdivision types are applied at different refinement stages to optimize the balance between compression ratio and restoration quality. The encoder can adjust these parameters based on the specific characteristics of the mesh being encoded, such as curvature and feature density.
Data Source
AI summary
A method of encoding a dynamic mesh includes creating a base mesh through mesh decimation, subdividing the base mesh, extracting displacement information for the subdivided mesh, and encoding the base mesh and the displacement information. In this instance, mesh subdivision information for subdivision of the base mesh is encoded and signaled.