Adaptive Reduced Secondary Transform for High-Resolution Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution and high-quality images/videos, particularly in immersive media formats like VR and AR, necessitates a more efficient image/video compression technique to reduce transmission and storage costs.
Innovation Solution
An image coding method and apparatus utilizing a reduced secondary transform (RST) with optimized transformation kernel matrices and adaptive transform coefficient arrays based on intra prediction modes to enhance coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional image coding techniques are used, then existing compression is achieved, but transmission cost and storage cost increase significantly for high-resolution and high-quality images/videos
Solution Approach 1:
The patent applies parameter changes by optimizing the transform kernel matrix parameters and adapting the secondary transform based on intra prediction modes. The transform kernel matrix is adjusted according to different prediction modes (e.g., angular, planar, DC) to maximize compression efficiency for specific image characteristics, thereby reducing transmission and storage costs while maintaining high resolution and quality
Solution Approach 2:
The patent implements dynamics by making the secondary transform adaptive rather than fixed. The transform is dynamically selected based on the intra prediction mode of the current block, allowing the system to adapt to different image content and prediction characteristics. This dynamic adaptation enables optimized compression for each specific coding situation, addressing the contradiction between compression efficiency and transmission/storage cost
2Productivity
If transform efficiency is increased through optimized kernel matrices, then coding efficiency improves, but device complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the transform process into distinct components: primary transform and secondary transform. The secondary transform is further segmented by applying different transform kernels based on prediction modes. This segmentation allows each component to be optimized independently, improving overall coding efficiency while managing complexity through modular design
Solution Approach 2:
The patent implements local quality by applying different transform kernel matrices tailored to specific local conditions (intra prediction modes). Instead of using a single universal transform, the system selects optimal kernels for different prediction modes (angular, planar, DC), thereby improving coding efficiency for each local coding situation while keeping the overall system manageable through rule-based selection
3Productivity
If a reduced secondary transform is applied, then compression efficiency increases, but measurement precision and manufacturing precision requirements become more stringent
Solution Approach 1:
The patent applies parameter changes by using a reduced set of transform kernel parameters compared to conventional transforms. The optimized kernel matrices are designed to achieve high compression efficiency with fewer parameters, which reduces the precision requirements for implementation while maintaining or improving compression performance. This parameter reduction directly addresses the contradiction between compression efficiency and implementation precision
Data Source
AI summary
A video decoding method according to the present document is characterized by comprising: a step for deriving transform coefficients through inverse quantization on the basis of quantized transform coefficients for a target block; a step for deriving modified transform coefficients on the basis of an inverse reduced secondary transform (RST) of the transform coefficients; and a step for generating a reconstructed picture on the basis of residual samples for the target block on the basis of an inverse primary transform of the modified transform coefficients, wherein the inverse RST using a transform kernel matrix is performed on transform coefficients of the upper-left 4×4 region of an 8×8 region of the target block, and the modified transform coefficients of the upper-left 4×4 region, upper-right 4×4 region, and lower-left 4×4 region of the 8×8 region are derived through the inverse RST.