Reference Picture List Construction for Multiview Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies do not efficiently signal inter-view reference pictures in the reference picture set of a view component, leading to decreased decoding efficiency in multiview video coding and 3D video coding applications.
Innovation Solution
A method and system for generating and parsing a reference picture list that includes an inter-view reference picture set, allowing for efficient encoding and decoding by signaling inter-view reference pictures associated with different views within the same access unit, thereby enhancing inter-view prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If inter-view reference pictures are not explicitly signaled in the reference picture set, then the bitstream complexity is reduced, but the decoding efficiency decreases
Solution Approach 1:
The patent applies preliminary action by constructing the reference picture list in advance during encoding, where inter-view reference pictures are pre-identified and positioned in the reference picture list before decoding. This allows the decoder to efficiently access inter-view references without complex runtime determination, resolving the contradiction by preparing the reference structure beforehand.
Solution Approach 2:
The patent introduces an intermediary mechanism through the reference picture list structure, which acts as a mediator between the bitstream and the decoding process. The list explicitly signals inter-view reference picture positions, serving as an intermediary that simplifies decoder operations while maintaining coding efficiency, thus resolving the trade-off between bitstream complexity and decoding efficiency.
2Productivity
If inter-view reference pictures are explicitly included in the reference picture list, then the compression performance improves, but the bitstream size increases
Solution Approach 1:
The patent applies parameter changes by modifying the reference picture list structure to include explicit inter-view reference indicators. This changes the parameter of reference picture identification from implicit to explicit, enabling better compression performance through improved prediction accuracy while managing bitstream overhead through structured signaling.
Solution Approach 2:
The patent introduces another dimension to the reference picture organization by adding inter-view reference information to the traditionally intra-view reference picture list. This dimensional expansion allows the system to exploit inter-view correlations for better compression while the structured approach to signaling keeps the bitstream overhead controlled.
3Ease of manufacture
If the reference picture list is constructed without inter-view references, then the encoding process is simpler, but the error resilience decreases
Solution Approach 1:
The patent applies preliminary action by pre-organizing the reference picture list to include inter-view references in predetermined positions during encoding. This preliminary structuring maintains encoding simplicity while ensuring that error resilience mechanisms are already in place, as the decoder can reliably access multiple reference pictures including inter-view ones for error recovery.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A video encoder generates, based on a reference picture set of a current view component, a reference picture list for the current view component. The reference picture set includes an inter-view reference picture set. The video encoder encodes the current view component based at least in part on one or more reference pictures in the reference picture list. In addition, the video encoder generates a bitstream that includes syntax elements indicating the reference picture set of the current view component. A video decoder parses, from the bitstream, syntax elements indicating the reference picture set of the current view component. The video decoder generates, based on the reference picture set, the reference picture list for the current view component. In addition, the video decoder decodes at least a portion of the current view component based on one or more reference pictures in the reference picture list.