3D Video Coding SAO Parameter Sharing Across Views
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current 3D video coding techniques are inefficient in reducing storage and transmission bandwidth for multi-view video, as they do not effectively exploit the correlation between different views, leading to high data requirements.
Innovation Solution
Applying the Sample Adaptive Offset (SAO) process to dependent-view images and depth maps in 3D video coding systems, allowing SAO parameters to be shared or predicted between regions, and using disparity vectors to determine correspondence between target and source regions for improved coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If conventional video coding techniques are applied to each single-view video sequence independently, then the coding process is simple and straightforward, but the coding efficiency is very poor and storage/transmission bandwidth requirements are high
Solution Approach 1:
The patent merges the coding processes of multiple views by applying SAO parameters derived from one view to corresponding regions in other views. This combines the independent processing simplicity with inter-view correlation exploitation, achieving both ease of implementation and improved coding efficiency by reducing redundant SAO parameter encoding across multiple views
Solution Approach 2:
The patent copies SAO parameters from a source view to target views for corresponding regions. Instead of independently encoding SAO parameters for each view, the system copies parameters across views where correlation exists, significantly reducing the total number of parameters that need to be encoded while maintaining coding quality
2Ease of operation
If SAO parameters are independently encoded for each view, then each view can be processed independently, but the amount of side information increases and storage/transmission bandwidth is high
Solution Approach 1:
The patent merges SAO parameter encoding across multiple views by identifying corresponding regions and reusing parameters from one view in another. This reduces the total quantity of side information while preserving the ability to process views independently where correlation is not exploited
Solution Approach 2:
The patent applies local quality by selectively sharing SAO parameters only in regions where inter-view correlation exists, while maintaining independent parameter encoding in regions where views differ significantly. This is achieved through region matching and correspondence detection between views
3Adaptability or versatility
If more cameras are used to generate multi-view video with more views, then the viewing experience and realism are improved, but the storage space and transmission bandwidth requirements increase significantly
Solution Approach 1:
The patent changes the parameter representation by using disparity vectors to describe the relationship between views instead of independently encoding all parameters. This parameter transformation exploits the geometric correlation between views, allowing high-quality multi-view video with reduced data requirements through compact parameter representation
Data Source
AI summary
A method and apparatus for 3D video coding system are disclosed. Embodiments according to the present invention apply SAO process (sample adaptive offset process) to at least one dependent-view image of the processed multi-view images if processed multi-view images are received. Also embodiments according to the present invention apply the SAO process to at least one dependent-view image of the processed multi-view images or at least one depth map of the processed multi-view depth maps if both processed multi-view images and the processed multi-view depth maps are received. The SAO can be applied to each color component of the processed multi-view images or the processed multi-view depth maps. The SAO parameters associated with a target region in one dependent-view image or in one depth map corresponding to one view may share or may be predicted by second SAO parameters associated with a source region corresponding to another view.