Constrained Intra-Prediction for Video Coding Pipeline Integrity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies face challenges in maintaining the integrity of the S×S pipeline structure due to intra-prediction reference sample dependencies, which affects compression efficiency and user experience, especially with limited bandwidth and increasing video quality demands.
Innovation Solution
The method involves splitting coding tree units into coding units using specific split modes (such as triple tree, vertical binary tree, or horizontal binary tree) and marking certain units not to be predicted by intra-prediction based on threshold conditions, thereby preventing intra-prediction under certain conditions to maintain the S×S pipeline structure, allowing for improved compression and decoding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If intra-prediction is applied to all coding units, then compression efficiency is improved, but the S×S pipeline structure integrity is compromised due to reference sample dependencies
Solution Approach 1:
The patent applies different prediction strategies to different regions of the video data based on the S×S pipeline structure. Specifically, coding units that would disrupt the pipeline structure are identified and excluded from intra-prediction, while other units continue to use intra-prediction for compression. This localized application of prediction constraints maintains overall system integrity while preserving compression efficiency where possible.
Solution Approach 2:
The patent segments the video data processing into distinct phases that respect the S×S pipeline structure. By dividing the coding units into those that can undergo intra-prediction and those that cannot (based on their position and size relative to the S×S blocks), the system maintains pipeline integrity while still applying compression techniques where appropriate.
2Reliability
If the S×S pipeline structure is maintained strictly, then processing reliability is improved, but compression flexibility is reduced
Solution Approach 1:
The patent introduces dynamic decision-making in the encoding process by evaluating each coding unit's characteristics (size, position, split mode) against the S×S pipeline constraints. This allows the system to adaptively determine whether intra-prediction can be applied to each unit, providing flexibility within the structured pipeline framework rather than imposing rigid constraints on all units.
Solution Approach 2:
The patent changes the parameters of coding units that are marked for intra-prediction exclusion, specifically setting their prediction mode to inter-prediction or other modes that do not require reference samples from adjacent S×S blocks. This parameter adjustment maintains pipeline reliability while accommodating the need for compression flexibility in permissible regions.
3Stability of the object's composition
If intra-prediction is prevented in certain conditions, then the S×S pipeline structure is maintained, but compression ratio is reduced
Solution Approach 1:
The patent converts the potential harm of reduced compression ratio into a benefit by ensuring the S×S pipeline structure integrity, which enables reliable parallel processing and prevents decoding errors. The constraint that initially appears to reduce compression efficiency actually ensures proper pipeline operation, which is essential for the overall system functionality and long-term compression performance.
Solution Approach 2:
The patent applies intra-prediction selectively rather than universally, marking only those coding units that would disrupt the S×S pipeline structure for exclusion. This partial application of intra-prediction maintains the pipeline integrity while still achieving compression benefits from the majority of coding units where intra-prediction is permissible.
Data Source
AI summary
A method of decoding a coded video bitstream implemented by a video decoder. The method includes determining a split mode for a coding tree unit (CTU) in an encoded bitstream, where the split mode is one of triple tree (TT), vertical binary tree (BT), and horizontal BT; marking a coding unit (CU) from the CTU as not to be predicted by intra prediction when one of three conditions is met; parsing the encoded bitstream to obtain a prediction unit and residual information for the CU from the CTU; and reconstructing the CU based on the prediction unit and the residual information when the CU has been marked.