Progressive Mesh Compression Using Superset LOD Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing mesh compression techniques, such as Draco compression, do not efficiently utilize all available information during compression and decompression of progressive meshes, leading to inefficiencies in storing and transmitting computer-generated three-dimensional models.
Innovation Solution
The proposed method involves encoding meshes using progressive mesh compression, where subsequent meshes are encoded based on a superset relationship with previous meshes, applying operators such as 'add quantization bits,' 'add vertex,' and 'move vertex' to produce a progressive mesh, and using differential mesh compression to encode differences in vertex coordinates, enabling more efficient use of information during transmission or storage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If single-mesh compression methods like Draco are used, then compression is achieved for individual meshes, but information from lower-quality LODs cannot be utilized when decompressing higher-quality LODs
Solution Approach 1:
The patent applies preliminary action by pre-processing meshes to identify and encode superset relationships between consecutive LODs before compression. The system analyzes the hierarchical structure in advance, determining which vertices and triangles in higher-quality LODs are supersets of those in lower-quality LODs, and encodes this relationship information into the compressed stream, enabling efficient incremental reconstruction during decompression.
Solution Approach 2:
The patent segments the compression process into distinct phases: identifying superset relationships between consecutive LODs, encoding the base mesh, and encoding differential updates for each subsequent LOD. This segmentation allows the system to handle each LOD transition independently, utilizing information from previous LODs while maintaining modular compression operations.
2Loss of information
If progressive mesh compression is implemented to utilize information from lower-quality LODs, then information efficiency improves, but existing methods focus on mathematical compression rather than bit-encoding, leading to inefficiencies
Solution Approach 1:
The patent applies parameter changes by transforming the representation of mesh data from absolute coordinates to differential updates based on superset relationships. The system changes the encoding parameters to represent only the differences between consecutive LODs, using techniques like vertex displacement vectors and triangle insertion flags, which significantly reduces the bit-stream size while maintaining geometric accuracy.
Solution Approach 2:
The patent introduces an intermediary encoding layer that sits between the mathematical mesh representation and the final bit-stream. This intermediary layer includes data structures for storing superset relationship information, vertex displacement data, and triangle insertion markers, which mediate the transformation from geometric models to compressed streams, enabling efficient information utilization.
3Measurement precision
If vertex coordinates are encoded with high precision, then rendering quality improves, but data size increases
Solution Approach 1:
The patent applies local quality by encoding vertex coordinates with varying precision based on their position and importance in the mesh hierarchy. Critical vertices that define mesh topology are encoded with higher precision, while less significant vertices use lower precision. The system also applies different precision levels to different LODs, using full precision only for the base mesh and reduced precision for differential updates in higher-quality LODs.
Solution Approach 2:
The patent applies partial action by encoding only the necessary precision for each vertex coordinate based on the required rendering quality and the vertex's role in the mesh. Instead of uniformly encoding all vertices with maximum precision, the system encodes only the minimal required precision for each vertex, reducing overall data size while maintaining adequate rendering quality.
Data Source
AI summary
The systems and methods described herein may improve the rendering of computer-generated three-dimensional models using progressive mesh compression. In various implementations, an initial mesh may be obtained and encoded into a data stream. Subsequent meshes may then be encoded based on a superset relationship between consecutive meshes. If the vertices of the mesh are not a superset of a prior mesh, the mesh may be encoded within an intermediate symbol stream using a non-incremental mesh compression technique. If the vertices of the mesh are a superset of a prior mesh, a sequence of per-triangle operators may be applied to the mesh to produce a progressive mesh. The mesh may then be encoded by encoding the operators applied to the mesh in sequence. When encoding the mesh, coordinates of vertices may be defined based on the difference between the coordinates and predicted values generated using a prediction function.