Subblock Temporal Motion Vector Prediction for Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution and high-quality images and videos, such as UHD images and videos of 4K or 8K, leads to higher transmission and storage costs due to increased data volume. Additionally, existing technologies face challenges in efficiently compressing and predicting motion vectors in image and video coding, particularly in immersive media applications like VR and AR.
Innovation Solution
The proposed solution involves enhancing video/image coding efficiency by deriving subblock-based temporal motion vectors and reducing complexity in the process. This is achieved by using base motion information from reference subblocks, leveraging motion information from temporal motion vector prediction, and applying a motion shift to improve prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If high-resolution and high-quality images and videos are transmitted or stored using existing mediums, then image and video quality is improved, but transmission costs and storage costs are increased
Solution Approach 1:
The current block is divided into multiple subblocks, and motion vectors are derived independently for each subblock. This segmentation allows for more precise motion compensation at the subblock level, improving prediction accuracy and enabling more efficient compression of high-resolution video content without requiring proportionally higher transmission and storage resources.
2Measurement precision
If temporal motion vector prediction technology is applied in the unit of subblock, then prediction performance is improved, but complexity in deriving base motion data is increased
Solution Approach 1:
The patent combines TMVP (temporal motion vector prediction) and sbTMVP (subblock-based temporal motion vector prediction) modes into a unified framework. The decoding apparatus determines whether to apply sbTMVP based on the block size and availability flags, merging the simplicity of TMVP with the enhanced accuracy of sbTMVP where appropriate, thereby improving prediction performance while controlling complexity through conditional application.
Solution Approach 2:
The patent applies different motion prediction strategies to different regions: full-block TMVP for larger blocks and subblock-based TMVP for smaller blocks or specific regions where higher precision is beneficial. This local quality approach ensures that computational complexity is concentrated only where it provides the most value, improving overall prediction accuracy without uniformly increasing complexity across all blocks.
3Measurement precision
If subblock-based temporal motion vectors are derived with additional base motion information, then prediction performance is improved, but computational complexity is increased
Solution Approach 1:
The patent performs preliminary checks before deriving subblock-based temporal motion vectors: it first determines if the block size meets the threshold (e.g., 64x64 or larger) and checks availability flags. These preliminary actions prevent unnecessary computational complexity from being incurred in blocks where sbTMVP is not applicable or beneficial, while still enabling enhanced prediction performance where the conditions are met.
Data Source
AI summary
According to the disclosure of this document, in subblock-based temporal motion vector prediction (sbTMVP), base motion information for a subblock for which a motion vector is not available can be efficiently derived, and through this, video/image coding efficiency can be improved.