Inter-view Motion Vector Prediction for 3D Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current 3D video coding techniques lack accurate motion vector prediction for depth views, which hinders efficient compression and transmission of three-dimensional video content.
Innovation Solution
The proposed solution involves deriving a disparity vector from neighboring depth pixels in a dependent depth view and using it to generate inter-view motion vector candidates from a base depth view, improving the accuracy of motion vector prediction by leveraging more motion vector candidates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional motion vector prediction is used for depth views, then coding complexity is reduced, but prediction accuracy deteriorates
Solution Approach 1:
The patent introduces an intermediary process that derives disparity vectors from neighboring depth pixels and uses them to generate inter-view motion vector candidates. This intermediary step bridges the base depth view and dependent depth view, improving prediction accuracy by leveraging depth information and disparity relationships without requiring completely new prediction mechanisms.
Solution Approach 2:
The patent extends motion vector prediction from a single-view temporal dimension to a multi-view spatial-temporal dimension by incorporating inter-view prediction. It uses disparity vectors derived from depth values to access motion information from corresponding blocks in base views, adding a new dimensional perspective for prediction that improves accuracy beyond conventional intra-view methods.
2Measurement precision
If more motion vector candidates are generated from base depth view, then prediction accuracy improves, but computational load increases
Solution Approach 1:
The patent performs preliminary derivation of disparity vectors from neighboring depth pixels before generating motion vector candidates. By pre-computing disparity information from already-available depth data, the system prepares prediction candidates in advance without requiring intensive real-time computation during the prediction process itself, thus improving accuracy while managing computational load.
Solution Approach 2:
The patent generates motion vector candidates by copying and adapting motion information from corresponding blocks in the base depth view. Instead of independently computing motion vectors for each block in the dependent view, the system copies motion vectors from the base view and adjusts them using disparity vectors, significantly reducing computational requirements while maintaining prediction accuracy.
Data Source
AI summary
An example video coding device is configured to compare an inter-view predicted motion vector candidate (IPMVC) to a motion vector inheritance (MVI) candidate, where the IPMVC and the MVI candidate are each associated with a block of video data in a dependent depth view, and where the IPMVC is generated from a corresponding block of video data in a base depth view. The video coding device may be further configured to perform one of adding the IPMVC to a merge candidate list based on the IPMVC being different from the MVI candidate, or omitting the IPMVC from the merge candidate list based on the IPMVC being identical to the MVI candidate.