Virtual Boundary Signaling for 360-Degree Video Artifact Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In video processing, existing technologies face challenges in efficiently disabling in-loop filtering operations across virtual boundaries in 360-degree video and gradual decoding refresh applications, leading to face seam artifacts and increased transmission latency.
Innovation Solution
The method involves signaling virtual boundaries in the bitstream to disable in-loop filtering operations across these boundaries, using Sequence Parameter Sets (SPS) and Picture Headers (PH) to define the virtual boundary positions, and employing wrap-around motion compensation to improve compression performance and reduce artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If in-loop filtering operations are applied across virtual boundaries, then filtering quality is improved, but face seam artifacts occur and transmission latency increases
Solution Approach 1:
The video sequence is segmented into multiple picture groups with virtual boundaries, allowing independent filtering control in each segment. This segmentation enables the system to apply filtering within segments while preventing cross-segment filtering that causes face seam artifacts, thus resolving the contradiction between filtering quality and artifact prevention.
Solution Approach 2:
The patent applies different filtering strategies to different spatial regions by introducing virtual boundaries. In-loop filtering is enabled within each picture group but disabled across virtual boundaries, creating local quality variations that prevent face seam artifacts while maintaining filtering benefits in appropriate regions.
2Manufacturing precision
If in-loop filtering operations are applied across virtual boundaries, then filtering quality is improved, but transmission latency increases
Solution Approach 1:
By segmenting the video sequence into picture groups with virtual boundaries, the system can process and transmit each segment independently with appropriate filtering applied only where beneficial. This reduces overall transmission latency compared to applying filtering across the entire sequence, while still maintaining filtering quality within each segment.
3Measurement precision
If virtual boundaries are signaled at picture level, then filtering control precision is improved, but device complexity increases
Solution Approach 1:
The patent introduces a new dimension for boundary signaling by using virtual boundaries at the picture group level rather than only at the picture level. This dimensional change allows for coarser-grained control that reduces bitstream complexity while still providing sufficient filtering control precision for 360-degree video applications.
4Productivity
If wrap-around motion compensation is employed, then compression performance is improved, but computational complexity increases
Solution Approach 1:
Wrap-around motion compensation is applied locally at virtual boundaries in 360-degree video content, where it provides the most benefit for compression performance. By limiting the application to specific boundary regions rather than the entire video sequence, the computational complexity is reduced while maintaining compression advantages where they are most needed.
Data Source
AI summary
The present disclosure provides methods for picture processing. The method can include: receiving a bitstream comprising a set of pictures; determining, according to the received bitstream, whether a virtual boundary is signaled at a sequence level for the set of pictures; in response to the virtual boundary being signaled at the sequence level, determining a position of the virtual boundary for the set of pictures, the position being bounded by a range signaled in the received bitstream; and disabling in-loop filtering operations across the virtual boundary.