Scaled Intra Reference Pictures for Multi-Resolution Block Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently compressing video data with varying spatial resolutions, particularly in handling scaled and full-scale versions of video pictures, leading to inefficiencies in reconstruction and prediction processes.
Innovation Solution
The proposed solution involves reconstructing scaled and full-scale video pictures using partitioning information and intra prediction techniques, where samples in the scaled version are used to predict and blend with up-sampled reconstructed samples, and residues are predicted based on residue distributions, with flags and intra prediction modes guiding the reconstruction process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of substance
If video data is compressed using traditional intra prediction techniques, then compression efficiency is improved, but reconstruction accuracy for scaled and full-scale pictures deteriorates
Solution Approach 1:
The picture is divided into multiple blocks, and each block is processed separately using appropriate prediction modes. Scaled picture blocks use scaling-based prediction while full-scale blocks use traditional intra prediction, allowing optimized handling of different resolution regions
Solution Approach 2:
A scaling process is introduced as an intermediary step between the compressed bitstream and the final reconstructed picture. The scaled picture serves as an intermediate reference that bridges the gap between compressed data and high-quality reconstruction across different resolutions
2Manufacturing precision
If separate processing is applied to scaled and full-scale pictures, then reconstruction quality is improved, but computational complexity increases
Solution Approach 1:
Different processing strategies are applied to different regions: scaled picture blocks use scaling-based prediction while full-scale blocks use traditional intra prediction. This local differentiation optimizes quality where needed while reducing unnecessary computation in other regions
Solution Approach 2:
The patent applies full scaling processing only to blocks that require it (scaled picture blocks), while using simpler traditional prediction for full-scale blocks. This partial application of the complex scaling process reduces overall computational burden while maintaining quality where necessary
3Productivity
If traditional intra prediction is used for all blocks, then processing speed is maintained, but prediction accuracy for scaled regions deteriorates
Solution Approach 1:
The prediction method dynamically adapts to the block type: scaled picture blocks use scaling-based prediction while full-scale blocks use traditional intra prediction. This dynamic selection maintains processing efficiency by using simpler methods where appropriate while improving accuracy where needed
Data Source
AI summary
Aspects of the disclosure include methods and apparatuses for video coding. The apparatuses include processing circuitry that receives a bitstream including a first sub-bitstream that corresponds to a scaled version of a current picture and a second sub-bitstream that corresponds to a current picture in a full scale. The processing circuitry reconstructs the scaled version of the current picture from the first sub-bitstream and reconstructs a second block in the current picture in the full scale based on (i) partitioning information of one or more first blocks in the scaled version of the current picture or (ii) intra prediction information of the one or more first blocks.