Temporal Derived Bi-Directional Motion Vector Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards face inefficiencies in motion vector prediction, particularly in reducing data usage for bi-directional motion vectors, as they rely on spatial or temporal neighboring blocks without effectively leveraging additional temporal information for improved coding efficiency.
Innovation Solution
Deriving a temporal derived motion vector based on a given motion vector to form bi-directional motion vector predictors, which involves identifying reference blocks in different frames and scaling motion vectors accordingly, and using these derived vectors as candidates in merge or advanced motion vector prediction modes to enhance coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If spatial or temporal neighboring blocks are used for motion vector prediction, then the coding complexity is reduced, but the coding efficiency is insufficient due to lack of additional temporal information
Solution Approach 1:
The patent extends motion vector prediction from traditional spatial/temporal neighboring blocks to multi-dimensional temporal derivation by scaling motion vectors across multiple reference frames with different time distances. This adds a temporal dimensionality to the prediction process, deriving motion vectors not only from immediate neighbors but from blocks across multiple reference frames scaled by their respective time distances, thereby improving coding efficiency without proportionally increasing data usage.
Solution Approach 2:
The patent changes the parameter of time distance by introducing scaling factors based on the temporal distance between current block and reference blocks. Motion vectors from different reference frames are scaled according to their time distances, allowing the system to adaptively adjust prediction accuracy based on temporal separation. This parameter change enables more effective utilization of temporal information while maintaining reasonable data usage.
2Measurement precision
If bi-directional prediction with two motion vectors is used, then the prediction accuracy is improved, but the data amount for signaling motion vectors increases
Solution Approach 1:
The patent uses copying by deriving the second motion vector through temporal scaling of the first motion vector from reference frames. Instead of independently signaling two full motion vectors for bi-directional prediction, the system copies and scales motion information from previously coded blocks and reference frames, maintaining prediction accuracy while reducing the data amount required for motion vector signaling.
Solution Approach 2:
The patent introduces temporal scaling factors as intermediaries between the current block and reference frames. These scaling factors mediate the relationship between motion vectors at different time distances, allowing the system to derive accurate bi-directional motion vectors through intermediate temporal relationships rather than directly signaling all motion information, thus reducing data overhead.
3Productivity
If motion vectors from multiple reference frames are derived and scaled, then the temporal information utilization is improved, but the computational complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing motion vectors from spatial and temporal neighboring blocks before the actual prediction process. These pre-derived motion vectors serve as candidates that can be quickly scaled and applied without requiring complex real-time calculations during encoding/decoding, thus improving temporal information utilization while controlling computational complexity.
Solution Approach 2:
The patent applies local quality by selectively deriving and scaling motion vectors from specific reference frames based on local temporal characteristics. Rather than uniformly processing all reference frames with equal complexity, the system adapts the derivation process to local temporal distances and motion characteristics, optimizing the balance between temporal information utilization and computational complexity for different regions.
Data Source
Figure 1
Figure 2
AI summary
A method and apparatus of deriving a temporal derived motion vector in a second direction based on a given motion vector in a first direction for motion vector prediction are disclosed. According to the present invention, a given motion vector for a current block is determined, where the given motion vector points from the current block in a first direction. A reference motion vector associated with a first reference block in a first reference frame is identified. Then, based on the reference motion vector and the given motion vector, a temporal derived motion vector is derived. The temporal derived motion vector points from the current block to a second reference block in a second reference frame in a second direction different from the first direction. The temporal derived motion vector is then used as one predictor for encoding or decoding of the motion vector of the current block.