Unified Disparity Vector Derivation for 3D Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current 3D video coding systems face inefficiencies in disparity vector derivation, leading to potential mismatches between encoders and decoders due to the use of default disparity vectors pointing to non-existent reference pictures, especially when inter-view reference pictures are not included in the reference picture lists.
Innovation Solution
A unified disparity vector derivation method is introduced, where if a derived disparity vector is unavailable, a default disparity vector is set to point to the inter-view reference picture with the minimum view index or nearest view index, and if no inter-view reference pictures are found, the default view index is set to -1 to disable certain coding tools, ensuring valid reference pictures are used.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a default disparity vector is used when derived DV is unavailable, then coding efficiency is maintained, but mismatches between encoders and decoders occur when reference pictures are not included in reference picture lists
Solution Approach 1:
The patent applies preliminary action by checking the availability of inter-view reference pictures in the reference picture lists before using a default disparity vector. The encoder and decoder both perform this check in advance to ensure that the default DV points to a valid reference picture, preventing mismatches while maintaining coding efficiency.
Solution Approach 2:
The patent implements feedback by having the decoder verify whether the reference picture indicated by the default disparity vector is actually available in the reference picture lists. If the reference picture is not available, the decoder can detect the mismatch and handle it appropriately, ensuring encoder-decoder consistency.
2Device complexity
If disparity vector derivation is simplified, then processing complexity is reduced, but accuracy of view synthesis prediction deteriorates
Solution Approach 1:
The patent applies dynamics by making the disparity vector derivation adaptive: when a derived disparity vector is available from neighboring blocks, it is used; when unavailable, a default DV is used with validation. This dynamic approach balances processing complexity and accuracy based on available information.
Solution Approach 2:
The patent changes the parameter of disparity vector selection based on availability conditions. The system switches between using derived DVs (when available) and default DVs (when unavailable), with an additional validation parameter to ensure the default DV points to a valid reference picture, thus maintaining accuracy without excessive complexity.
3Quantity of substance
If inter-view reference pictures are excluded from reference picture lists, then memory usage is reduced, but validity of default disparity vectors is compromised
Solution Approach 1:
The patent performs a preliminary check to verify that the reference picture indicated by the default disparity vector is available in the reference picture lists before using it. This preliminary validation ensures that even when inter-view reference pictures are excluded from the lists to save memory, the system only uses valid reference pictures, maintaining reliability.
Solution Approach 2:
The patent uses a validation mechanism that is computationally inexpensive to check the validity of the default disparity vector. This lightweight validation ensures that memory is saved by excluding unnecessary reference pictures while still maintaining reliability through quick validity checks when needed.
Data Source
AI summary
A method and apparatus for a three-dimensional or multi-view video encoding or decoding system utilizing unified disparity vector derivation is disclosed. When a three-dimensional coding tool using a derived disparity vector (DV) is selected, embodiments according to the present invention will first obtain the derived DV from one or more neighboring blocks. If the derived DV is available, the selected three-dimensional coding tool is applied to the current block using the derived DV. If the derived DV is not available, the selected three-dimensional coding tool is applied to the current block using a default DV, where the default DV is set to point to an inter-view reference picture in a reference picture list of the current block.


