Depth Lookup Table Signaling in 3D Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing 3D-HEVC standard for multi-view video coding faces issues with DLT signaling, including invalid predictions, bit-depth inconsistencies, and redundancy in signaling depth lookup tables across views.
Innovation Solution
The proposed method addresses these issues by ensuring valid DLT predictions, consistent bit-depth information, and reducing redundancy by only signaling DLT in picture parameter sets (PPS) for depth data, using existing DLTs when possible, and scaling bit depths to match sequence-level indications, and eliminating redundant signaling for texture components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If DLT is signaled in PPS for all views including texture components, then completeness of parameter signaling is improved, but signaling redundancy increases
Solution Approach 1:
The patent applies local quality by differentiating the signaling requirements for different components (depth vs. texture) and different views. DLT is signaled only in the first view's PPS and only for depth components, rather than uniformly for all views and components. This localized approach eliminates redundant signaling while ensuring necessary parameters are present where needed.
2Productivity
If inter-view DLT prediction is enabled for all views, then coding efficiency is improved, but prediction validity deteriorates when reference DLT does not exist
Solution Approach 1:
The patent introduces dynamic control of inter-view prediction through the inter_view_dlt_pred_enable_flag, which is view-dependent. This allows the system to adaptively enable or disable prediction based on whether a valid reference DLT exists, rather than applying a static prediction rule to all views. The flag dynamically adjusts prediction behavior to maintain validity while preserving efficiency where applicable.
Solution Approach 2:
The patent prevents invalid predictions by checking the existence of reference DLT before enabling inter-view prediction. The conditionals in the syntax structure ensure that prediction is only attempted when a valid reference exists, preemptively avoiding the harmful effect of invalid predictions on coding reliability.
3Adaptability or versatility
If separate DLT signaling is used for each view, then view-specific optimization is improved, but overall signaling efficiency deteriorates
Solution Approach 1:
The patent merges the DLT signaling for multiple views into a single PPS structure. Instead of repeating DLT parameters in each view's PPS, the DLT is signaled once in the first view's PPS and then referenced by subsequent views through the inter_view_dlt_pred_enable_flag. This combining approach improves overall signaling efficiency while still allowing view-specific prediction behavior.
4Adaptability or versatility
If DLT bit depth is independently signaled in PPS, then flexibility in depth representation is improved, but consistency with sequence level parameters deteriorates
Solution Approach 1:
The patent implements feedback control by checking whether inter_view_dlt_pred_enable_flag is set before allowing independent DLT bit depth signaling. When prediction is enabled, the system ensures bit depth consistency with the reference DLT. This feedback mechanism maintains parameter consistency while still allowing flexibility when prediction is not used, resolving the contradiction between adaptability and stability.
Data Source
AI summary
A method and apparatus for depth lookup table (DLT) signaling in a three-dimensional and multi-view coding system. The method identifies one or more pictures to be processed. If one or more pictures contain depth data, then the method determines the DLT associated with said one or more pictures, applies predictive coding to the DLT based on the previous DLT, includes syntax related to the DLT in the PPS, and includes first bit-depth information related to first depth samples of the DLT in the PPS. The first bit-depth information is consistent with second bit depth information signaled in a sequence level. The method further signals the PPS in a video bitstream for a sequence including said one or more pictures. A circuit is also provided that embodies circuitry configured to carry out the operations specified above.