Scalable Video Motion Information Prediction Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In scalable video coding, the deactivation of temporal motion information predictors in the reference layer can lead to a lack of available motion information for prediction in the enhancement layer, resulting in increased storage requirements and reduced coding efficiency, especially for high-resolution videos like 4k2k or 8k4k.
Innovation Solution
A method to determine motion information predictors in the enhancement layer by correcting the position of motion information in the reference layer, using compressed motion fields to derive relevant motion information, and applying this in the motion derivation process to improve prediction efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If temporal motion information predictors are deactivated in the reference layer, then storage requirements are reduced, but motion information availability for enhancement layer prediction deteriorates
Solution Approach 1:
The patent extracts motion information from the reference layer independently of the temporal motion information predictor mechanism. By directly accessing and utilizing motion information from the reference layer through inter-layer prediction, the system obtains necessary motion data without relying on the temporal predictor storage infrastructure.
Solution Approach 2:
The patent introduces an inter-layer prediction mechanism as an intermediary between the reference layer and enhancement layer. This intermediary process allows motion information to be transferred and utilized across layers without requiring storage in the temporal motion information predictor buffers.
2Device complexity
If temporal motion information predictors are deactivated in the reference layer, then device complexity is reduced, but coding efficiency deteriorates
Solution Approach 1:
The patent segments the motion prediction process into intra-layer prediction (within enhancement layer) and inter-layer prediction (from reference layer). This segmentation allows the system to bypass the deactivated temporal motion information predictors in the reference layer while maintaining coding efficiency through direct inter-layer motion information utilization.
Solution Approach 2:
The patent transitions from temporal prediction (time dimension within a layer) to inter-layer prediction (spatial dimension across layers). By moving to another dimension of prediction, the system achieves coding efficiency without relying on the temporal motion information predictor mechanism.
3Quantity of substance
If motion information is not stored in compressed motion fields, then storage requirements are reduced, but prediction accuracy deteriorates
Solution Approach 1:
The patent creates a copy of motion information from the reference layer and utilizes it directly in the enhancement layer through inter-layer prediction. This copying mechanism allows the system to obtain accurate motion information without storing it in compressed motion fields within the temporal predictor infrastructure.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention concerns a method, device, and computer program for motion vector prediction in scalable video encoder and decoder. It concerns the process to determine motion information predictor in the enhancement layer of a scalable encoding scheme also known as motion derivation process. It comprises a correction of the position in the reference layer used to pick-up the more relevant motion information available due to the compression scheme. Accordingly, motion information prediction is improved.