Affine Intra Block Copy for Geometric Screen Content Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding and decoding technologies face challenges in efficiently compressing and decompressing video sequences, particularly for screen content with complex geometric transformations, leading to inefficiencies in storage and transmission.
Innovation Solution
Implementing affine transformation-based intra block copy (IBC) prediction techniques to enhance video encoding and decoding, specifically for screen content, by using affine models with rotation, scaling, and translation to improve prediction accuracy and reduce redundancy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional intra block copy prediction is used for screen content, then encoding complexity is low, but compression efficiency is poor for content with geometric transformations
Solution Approach 1:
The patent applies affine transformation parameters (rotation angle, scaling factor, translation vector) to modify the reference block geometry. By changing the parameters of the transformation model, the system can adapt to different geometric distortions in screen content, thereby improving compression efficiency without requiring a completely new encoding framework
Solution Approach 2:
The patent introduces dynamic affine transformation that adapts to local geometric variations within different blocks. Instead of using a fixed transformation for the entire picture, the system dynamically selects and applies appropriate affine parameters for each block based on its local characteristics, improving compression efficiency while managing complexity through localized processing
2Measurement precision
If affine transformation is applied to reference blocks, then prediction accuracy improves for distorted content, but computational complexity increases
Solution Approach 1:
The patent divides the picture into multiple blocks and applies affine transformation independently to each block or group of blocks. This segmentation allows the computational complexity to be distributed and managed, while still achieving high prediction accuracy for geometrically distorted content within each segment
Solution Approach 2:
The patent applies affine transformation selectively only to blocks that exhibit geometric distortion characteristics, rather than uniformly to all blocks. By identifying and processing only the blocks that benefit from affine transformation, the system achieves high prediction accuracy where needed while minimizing unnecessary computational complexity
3Productivity
If existing IBC techniques are used, then encoding is simple, but storage and transmission efficiency are insufficient for complex screen content
Solution Approach 1:
The patent extends the intra block copy concept by copying and transforming reference blocks within the same picture using affine transformation. This allows accurate reproduction of geometrically distorted screen content while maintaining the efficiency benefits of intra-picture prediction, improving storage and transmission efficiency
Solution Approach 2:
The patent creates a unified prediction framework that can handle both standard IBC cases and affine-transformed cases through a single set of tools and algorithms. The affine IBC mechanism serves multiple functions: it handles geometric transformations, maintains intra-picture prediction efficiency, and provides a general solution for various screen content types, thereby improving storage efficiency without proportionally increasing decoding complexity
Data Source
AI summary
A decoder receives, from a bitstream, an indication of a block vector, a residual of a current block, an indication of an affine transform, and affine transform parameters. The decoder determines, based on the indication of the block vector, a reference block in a same picture as the current block for predicting the current block. The decoder applies, based on the indication of the affine transform, the affine transform parameters to the reference block to generate an affine transformation of the reference block. The decoder decodes the current block based on the residual and the affine transformation of the reference block.