Temporal Motion Vector Prediction Error Robustness
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In video coding systems, errors in the bitstream can lead to incorrect parsing of temporal Motion Vector Prediction (MVP) syntax elements, causing decoding artifacts that affect multiple frames and propagate errors, especially in low-bitrate applications where motion vector transmission is significant.
Innovation Solution
A method and apparatus for deriving a motion vector predictor (MVP) that selectively disables or enables temporal MVP candidates using a flag, allowing only spatial MVP candidates for prediction when errors are anticipated, and ensuring the availability of temporal MVPs to prevent parsing errors, while maintaining coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If temporal MVP candidates are used to improve coding efficiency, then compression performance is improved, but parsing errors can propagate across multiple frames
Solution Approach 1:
The patent introduces a dynamic flag (temporal_mvp_candidates_not_used) that allows the system to switch between using temporal MVP candidates and spatial MVP candidates based on error conditions. This dynamic control mechanism enables the system to adapt its prediction strategy in real-time, disabling temporal MVP when errors are detected while maintaining spatial MVP functionality, thus resolving the contradiction between coding efficiency and error robustness
Solution Approach 2:
The patent changes the operational parameter of MVP candidate selection by introducing a control flag that modifies the availability of temporal MVP candidates. When the flag is set, temporal MVP candidates are excluded from consideration, fundamentally changing the prediction process from relying on temporal redundancy to relying solely on spatial redundancy, thereby preventing error propagation while maintaining acceptable coding performance
2Reliability
If temporal MVP candidates are disabled to prevent error propagation, then parsing error robustness is improved, but coding efficiency deteriorates
Solution Approach 1:
The patent applies local quality control by selectively disabling temporal MVP candidates only in specific regions or conditions where error propagation is a concern, rather than globally disabling them. The flag can be applied at different granularities (slice level, picture level, or sequence level), allowing temporal MVP to remain enabled in safe regions while being disabled in problematic areas, thus maintaining overall coding efficiency while improving error robustness where needed
3Reliability
If spatial MVP candidates are used exclusively to ensure reliability, then error robustness is improved, but bitrate increases
Solution Approach 1:
The patent implements partial action by using spatial MVP candidates exclusively only when temporal MVP candidates are disabled due to error conditions. In normal operation, the system continues to use temporal MVP candidates to maintain optimal bitrate efficiency. This partial application of spatial-only MVP strategy minimizes the bitrate penalty while providing error robustness when needed
Data Source
AI summary
A method and apparatus for deriving a motion vector predictor (MVP) are disclosed. The MVP is selected from spatial MVP and temporalone or more MVP candidates. The method determines a value of a flag in a video bitstream, where the flag is utilized for selectively disabling use of one or more temporal MVP candidates for motion vector prediction. The method selects, based on an index derived from the video bitstream, the MVP from one or more non-temporal MVP candidates responsive to the flag indicating that said one or more temporal MVP candidates are not to be utilized for motion vector prediction. Further, the method provides the MVP for the current block.