Recursive Prediction Unit in Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies face challenges in efficiently processing and compressing video data, particularly in managing prediction modes and tree structures within coding units, which affects bandwidth usage and compression efficiency.
Innovation Solution
The method involves determining the application of a prediction splitting tree to a prediction tree unit, selecting a prediction mode based on whether the leaf prediction unit is split from the prediction tree unit, and performing conversions between visual media data and bitstreams, while disallowing certain prediction modes and techniques under specific conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a prediction splitting tree is applied to a prediction tree unit with multiple prediction modes, then prediction accuracy is improved, but device complexity and processing overhead increase
Solution Approach 1:
The prediction tree unit is segmented into multiple leaf prediction units through a prediction splitting tree structure. Each leaf PU can be independently processed with specific prediction modes, allowing precise prediction while managing complexity through hierarchical decomposition. The segmentation enables selective application of complex prediction modes only where needed in the tree structure.
Solution Approach 2:
Different prediction modes are applied to different leaf prediction units based on local characteristics. The patent allows selective disallowing of certain prediction modes (intra mode, combined inter-intra prediction, template-matching-based inter mode, local illumination compensation) for leaf PUs split from PTU, while allowing other modes. This local quality approach optimizes prediction accuracy for each region while controlling overall device complexity.
2Adaptability or versatility
If more prediction modes are allowed for leaf prediction units, then prediction flexibility and accuracy improve, but bandwidth usage and processing complexity increase
Solution Approach 1:
The patent applies local quality by allowing different prediction modes for different leaf prediction units based on their specific characteristics and position in the prediction tree. Certain modes are disallowed for leaf PUs split from PTU, while other modes remain available, creating a localized optimization strategy that balances flexibility and bandwidth efficiency for each region.
Solution Approach 2:
The patent changes the parameter of prediction mode availability based on the leaf PU's relationship to the prediction tree unit. By dynamically adjusting which prediction modes are permitted (disallowing intra mode, combined inter-intra prediction, template-matching-based inter mode, and local illumination compensation for split leaf PUs), the system optimizes the balance between prediction flexibility and bandwidth consumption.
3Device complexity
If prediction modes are restricted for leaf prediction units split from prediction tree units, then device complexity is reduced, but prediction accuracy may deteriorate
Solution Approach 1:
The prediction splitting tree segments the prediction tree unit into multiple leaf PUs, each of which can use a subset of prediction modes. This segmentation allows the system to reduce device complexity by limiting the modes available to split leaf PUs while maintaining overall prediction accuracy through the hierarchical structure and selective mode application in different tree regions.
Data Source
AI summary
A mechanism for processing video data implemented by a video coding apparatus is disclosed. The mechanism determines to apply a prediction splitting tree to a prediction tree unit (PTU). The prediction splitting tree includes a leaf prediction unit (PU). A prediction mode is selected for the leaf PU based on whether the leaf PU is split from the PTU. The method performs a conversion between a visual media data and a bitstream based on the leaf PU.