Picture Scalable Coding With 4:2:0 Base and 4:4:4 Enhancement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies using a 4:2:0 sampling format struggle with low prediction coding efficiency for enhancement layers when dealing with high-sampling-rate chrominance components, such as 4:4:4 format, leading to suboptimal picture quality and transmission efficiency.
Innovation Solution
A picture scalable coding method that involves downsampling a 4:4:4 sampling format to a 4:2:0 format for the base layer and encoding differential chrominance components separately to create a base and enhancement layer bitstreams, ensuring efficient transmission and high-frequency component preservation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a 4:2:0 sampling format is used for coding, then coding compatibility with existing standard video codecs is improved, but prediction coding efficiency of enhancement layer deteriorates when dealing with high-sampling-rate chrominance components
Solution Approach 1:
The patent segments the chrominance components into base layer (4:2:0) and enhancement layer (4:4:4) separately. The base layer contains low-frequency chrominance information while the enhancement layer contains high-frequency chrominance details, allowing each layer to be optimized for its specific requirements and resolving the contradiction between compatibility and coding efficiency.
Solution Approach 2:
The patent introduces a sampling rate dimension by creating two different sampling formats (4:2:0 and 4:4:4) for different layers. This dimensional approach allows the system to maintain compatibility at the base layer while achieving high coding efficiency at the enhancement layer through separate optimization in the sampling rate dimension.
2Productivity
If high-sampling-rate chrominance components are coded through existing standard video codecs, then picture transmission efficiency is improved, but picture quality deteriorates due to low prediction coding efficiency
Solution Approach 1:
The patent segments the chrominance information into base layer (low-frequency) and enhancement layer (high-frequency) components. This segmentation allows efficient transmission at the base layer while preserving picture quality through the enhancement layer that contains detailed high-frequency chrominance information.
Solution Approach 2:
The patent changes the sampling rate parameter between layers - using 4:2:0 sampling for the base layer to maximize transmission efficiency and 4:4:4 sampling for the enhancement layer to preserve picture quality. This parameter change allows both efficiency and quality requirements to be met simultaneously.
Data Source
AI summary
A method includes obtaining a current picture block, using a 4:4:4 sampling format for the current picture block. The current picture block includes three 2W×2H sample arrays, one 2W×2H sample array represents a first chrominance component, one 2W×2H sample array represents a second chrominance component, and one 2W×2H sample array represents a luminance component. The method further includes separately downsampling the first chrominance component and the second chrominance component of the current picture block to obtain a base layer picture block, where a 4:2:0 sampling format is used for the base layer picture block, obtaining a differential chrominance picture block based on the current picture block, where the differential chrominance picture block includes at least part of chrominance samples of the current picture block, encoding the base layer picture block to obtain a base layer bitstream, and encoding the differential chrominance picture block to obtain an enhancement layer bitstream.