Temporal Motion Vector Prediction Control in Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing technique for disabling temporal motion vector prediction in video coding is not robust, as it relies on a marking flag in the Picture Parameter Set, which can lead to synchronization issues between encoder and decoder if a slice is lost or contains errors.
Innovation Solution
A method and apparatus for encoding and decoding video using temporal motion vector prediction, where a flag indicates whether temporal motion vector prediction is used or not, allowing for independent enablement or disablement of this prediction for each sub-picture unit, reducing error susceptibility by eliminating the need for a marking process in the Decoder Picture Buffer.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a marking flag in Picture Parameter Set is used to disable temporal motion vector prediction, then temporal motion vector prediction can be disabled for certain slices, but synchronization between encoder and decoder is lost when a slice is lost or contains errors
Solution Approach 1:
The patent divides the video stream into independent slice units, each with its own temporal motion vector prediction flag. This segmentation allows error isolation so that errors in one slice do not propagate to other slices, maintaining synchronization independently for each slice while improving overall error robustness
Solution Approach 2:
The patent applies different temporal motion vector prediction settings to different slices based on local error conditions. Each slice can independently enable or disable temporal motion vector prediction based on its own error status, allowing localized error handling without affecting global synchronization
2Productivity
If temporal motion vector prediction is enabled for all slices, then coding efficiency is improved, but error propagation increases when reference pictures are lost or contain errors
Solution Approach 1:
The patent dynamically adjusts the temporal motion vector prediction setting for each slice based on error conditions. The encoder and decoder both adapt their behavior per-slice, enabling temporal motion vector prediction when reference pictures are reliable and disabling it when errors are detected, thus balancing coding efficiency and error robustness adaptively
Solution Approach 2:
The patent changes the operational parameters of temporal motion vector prediction on a per-slice basis. By modifying the prediction mode parameter individually for each slice according to its error status, the system optimizes coding efficiency for error-free slices while preventing error propagation in corrupted slices
Data Source
AI summary
A method of encoding a video into a coded video bitstream with temporal motion vector prediction comprises: determining a value of a flag for indicating whether temporal motion vector prediction is used or not used for the inter-picture prediction of a sub-picture unit of a picture; and writing the flag having the value into a header of the sub-picture unit or a header of the picture; wherein if the flag indicates that temporal motion vector prediction is used, the method further comprises: creating a first list of motion vector predictors comprising a plurality of motion vector predictors including at least one temporal motion vector predictor derived from at least one motion vector from a collocated reference picture; selecting a motion vector predictor out of the first list; and writing a first parameter into the coded video bitstream for indicating the selected motion vector predictor out of the first list.