Merge-Predicted Video Blocks With Triangular Inter-Intra Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards struggle to meet the increasing demand for higher quality video reproduction as playback equipment improves, particularly in scenarios where low to medium quality reconstruction is insufficient.
Innovation Solution
Implement a combination of inter-prediction and intra-prediction for triangular partitions of video blocks, using weighted averages and additional candidates in a look-up table to enhance prediction accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional video coding standards are used, then transmission payload is reduced, but video reproduction quality is insufficient for high-quality playback equipment
Solution Approach 1:
The video block is divided into two triangular partitions by a diagonal line, allowing different prediction modes to be applied to each partition. This segmentation enables more precise local prediction, improving overall video quality while maintaining compression efficiency.
Solution Approach 2:
The patent dynamically selects between inter-prediction and intra-prediction modes for each triangular partition based on local video characteristics. This dynamic adaptation allows the encoding system to optimize prediction accuracy for each region, improving reproduction quality without excessive bitrate increase.
2Measurement precision
If triangular partitioning with combination prediction is implemented, then prediction accuracy is improved, but encoding complexity increases
Solution Approach 1:
By dividing the block into two triangular partitions, the patent applies prediction techniques more locally and accurately. Each partition can use the most appropriate prediction mode (inter or intra), improving accuracy while keeping the complexity manageable through structured division.
Solution Approach 2:
The patent changes the prediction parameters (mode selection, weighting factors) based on the triangular partition structure and local video content. This allows adaptive optimization of prediction accuracy without requiring fundamentally new encoding algorithms, thus managing complexity.
3Productivity
If more prediction candidates are added to look-up table, then prediction efficiency is enhanced, but data transmission overhead increases
Solution Approach 1:
The patent dynamically selects from multiple prediction candidates in the look-up table based on local video characteristics and triangular partition properties. By having multiple candidates available but only transmitting the selected one (or a compact representation), prediction efficiency is improved without proportionally increasing data transmission.
Solution Approach 2:
The patent uses pre-defined prediction candidates stored in both encoder and decoder look-up tables. These candidates represent common prediction patterns that can be referenced without transmitting full prediction data, improving efficiency while minimizing overhead through compact indexing.
Data Source
AI summary
Decoding of a merge-predicted block is carried out, for triangular partitions of the block, using a combination of inter-prediction and intra-prediction. The combination of inter-prediction and intra-prediction to be used is signalled on a received bitstream. Combination may be algorithmic, for example arithmetic.