Affine Inter Prediction Small Subblocks Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies face challenges in efficiently compressing video data, particularly in reducing redundancy and bandwidth requirements, especially for high-resolution videos like 1080p60 4:2:0, which require significant storage space and bandwidth.
Innovation Solution
The implementation of affine inter prediction methods, including the use of small sub-blocks and advanced motion vector prediction techniques, such as affine motion compensated prediction and sub-block-based temporal motion vector prediction, to enhance compression efficiency by deriving motion vectors for smaller sub-blocks within a block, thereby reducing the data required for coding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of substance
If affine inter prediction with small subblocks is used, then compression ratio is improved, but device complexity increases
Solution Approach 1:
The current block is divided into multiple small subblocks (e.g., 4×4 subblocks within an 8×8 block). Each subblock is independently predicted using affine motion compensation, allowing for more precise local motion representation. This segmentation enables better compression by capturing fine-grained motion variations that would be lost in coarse block-based prediction.
Solution Approach 2:
Different motion models are applied to different regions of the block based on local motion characteristics. The affine motion compensation parameters (control point motion vectors) are derived separately for each subblock, allowing the prediction to adapt to local motion patterns such as rotation, scaling, or shearing that occur within specific regions of the block.
2Measurement precision
If motion vectors for small subblocks are derived, then prediction accuracy is improved, but computational complexity increases
Solution Approach 1:
Control point motion vectors are derived first from neighboring blocks or previous frames, and then used as the basis for calculating motion vectors for all subblocks within the current block. This preliminary derivation of control points reduces the overall computational burden by establishing a foundation that can be systematically applied across multiple subblocks rather than independently analyzing each subblock from scratch.
Solution Approach 2:
The patent extends traditional 2D motion compensation by introducing affine transformation parameters that operate in an extended parameter space. By using control point motion vectors to define affine transformations, the system captures complex motion patterns (rotation, scaling, shearing) that cannot be represented by simple translational motion vectors, thereby improving prediction accuracy for non-uniform motion.
Data Source
AI summary
Aspects of the disclosure provide methods and apparatuses for video encoding/decoding. In some examples, an apparatus for video decoding includes processing circuitry. The processing circuitry is configured to decode prediction information of a block in a current picture from a coded video bitstream. The prediction information indicates an affine model in an inter prediction mode. The processing circuitry is configured to determine motion vectors of control points of the block according to the affine model. The processing circuitry is configured to determine a motion vector of a sub-block of the block from the determined motion vectors of the control points, one of a width and a height of the sub-block being less than 4. Further, the processing circuitry is configured to reconstruct at least a sample of the sub-block according to the determined motion vector.