Mesh Motion Vector Grouping for Higher Encoding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The encoding efficiency of motion vectors in existing mesh technologies is low.
Innovation Solution
A mesh decoding device and method that includes a motion vector residual decoding unit, a motion vector prediction unit, and a motion vector calculation unit to improve encoding efficiency by generating and predicting motion vectors, with modes for setting the same mode for consecutive vertices and decoding a control signal for group size.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If motion vectors are encoded for each vertex individually using existing mesh technologies, then the motion vector can be accurately represented, but the encoding efficiency is low and the data amount is large
Solution Approach 1:
The patent segments vertices into groups where consecutive vertices share the same motion vector mode (mode 0 or mode 1). This segmentation allows the encoder to process multiple vertices together rather than individually, improving encoding efficiency while maintaining motion vector accuracy through group-based prediction and residual encoding.
Solution Approach 2:
The patent merges the encoding process for multiple consecutive vertices by applying the same mode to groups of vertices. Instead of encoding each vertex's motion vector separately, the system combines them into groups that share prediction modes, reducing the overall data amount and improving encoding efficiency.
2Measurement precision
If motion vectors are encoded with detailed prediction and residual information for each vertex, then the motion vector accuracy is maintained, but the amount of data required for encoding increases
Solution Approach 1:
By segmenting vertices into groups with shared modes, the patent reduces the number of individual motion vector encodings needed. Each group can use a single prediction mode and share residual encoding strategies, thereby maintaining accuracy while reducing the total data quantity.
Solution Approach 2:
Instead of encoding each vertex's motion vector independently with full detail, the patent inverts the approach by encoding group-level characteristics first (shared modes) and then handling individual variations through residuals. This inversion reduces redundant data while preserving accuracy.
3Ease of operation
If individual motion vector encoding is performed for each vertex, then the decoding process can be simple, but the processing speed and encoding efficiency are reduced
Solution Approach 1:
The patent segments the decoding process into group-level operations followed by individual vertex operations. By first decoding group characteristics (shared modes) and then applying them to individual vertices, the system maintains decoding simplicity while significantly improving processing speed through batch operations.
Solution Approach 2:
The patent performs preliminary decoding of group-level motion vector modes before processing individual vertices. This preliminary action establishes the prediction framework for entire groups, allowing subsequent individual vertex decoding to proceed faster while maintaining simplicity.
Data Source
AI summary
In a mesh decoding device according to the present invention, in a mode 1, the motion vector calculation unit outputs the motion vector of the vertex to be decoded by adding the motion vector residual and the motion vector predicted value, in a mode 0, the motion vector calculation unit outputs the motion vector residual as the motion vector of the vertex to be decoded, the motion vector calculation unit sets the same mode for motion vectors of N (N≥1) consecutive vertices in a decoding order, the N vertices are included in one group, and the motion vector calculation unit decodes, from a bit stream, a control signal for calculating a group size indicating a size of the group.