Video Coding Sub-picture Segmentation and Adaptive Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies face challenges in efficiently managing bandwidth demand due to the increasing number of connected devices, particularly in digital video transmission, where existing methods do not effectively optimize video processing for sub-picture based coding and decoding.
Innovation Solution
The proposed techniques involve methods for video processing that include conversion between video blocks and bitstreams, utilizing sub-picture based coding, palette coding, and dynamic resolution conversion, while disabling certain filtering processes to improve efficiency and adaptability in video coding standards like HEVC and VVC.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If sub-picture based coding is performed, then coding efficiency and bandwidth usage are improved, but device complexity increases
Solution Approach 1:
The picture is divided into multiple sub-pictures, each with independent coding parameters and processing constraints. This segmentation allows optimized coding for each sub-picture region while maintaining overall system manageability through structured organization of coding units.
Solution Approach 2:
Different coding parameters such as loop filtering disable/enable status, deblocking filtering constraints, and motion estimation region sizes are changed based on sub-picture boundaries. This enables adaptive parameter adjustment to improve coding efficiency without requiring complete system redesign.
2Loss of time
If loop filtering is disabled across sub-picture boundaries, then processing time is reduced, but video quality deteriorates
Solution Approach 1:
Loop filtering is selectively enabled or disabled based on local characteristics of different sub-pictures and boundary conditions. This local adaptation allows quality preservation in regions where it benefits most while reducing processing time in regions where computational resources are constrained.
Solution Approach 2:
Instead of applying loop filtering uniformly across all boundaries, the method applies filtering selectively to specific sub-picture boundaries based on configured constraints. This partial application reduces overall processing time while maintaining adequate quality through strategic filtering where most needed.
3Device complexity
If deblocking filtering is disabled at sub-picture boundaries, then processing complexity is reduced, but artifacts increase
Solution Approach 1:
The deblocking filtering process is dynamically adjusted based on boundary characteristics and sub-picture configurations. Filtering strength and application are modified adaptively rather than applied uniformly, reducing complexity where boundaries are less critical while maintaining artifact reduction where needed.
Solution Approach 2:
Deblocking filtering parameters such as filter strength, application threshold, and processing scope are changed based on sub-picture boundary conditions. This enables the system to reduce processing complexity by disabling or weakening filtering at boundaries where artifacts are less noticeable while maintaining quality where artifacts would be more prominent.
4Measurement precision
If merge estimation region size is increased, then motion prediction accuracy is improved, but processing time increases
Solution Approach 1:
The motion estimation process is segmented into different regions with different MER sizes based on sub-picture boundaries and motion characteristics. This allows high-precision motion prediction in regions where it is most important while using smaller, faster-to-process MERs in regions where motion is less variable.
Solution Approach 2:
The size and characteristics of merge estimation regions are locally adapted based on sub-picture content and boundary conditions. This enables optimized motion prediction accuracy for each local region while controlling overall processing time through selective application of computationally intensive operations only where necessary.
Data Source
AI summary
A method of video processing is provided that includes performing a conversion between a block of a video and a bitstream of the video. The bitstream conforms to a formatting rule specifying that a size of a merge estimation region (MER) is indicated in the bitstream and the size of the MER is based on a dimension of a video unit. The MER comprises a region used for deriving a motion candidate for the conversion.