Inter-Layer Prediction Clipping for Scalable Video Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In scalable video coding based on HEVC, the inefficiency in coding efficiency occurs when portions of the enhancement layer do not correspond to areas within the actual or coded frame size of the base layer, leading to unavailability of necessary information for inter-layer prediction, particularly when the base layer is coded with a different standard or has a larger frame size.
Innovation Solution
The system predicts blocks in the enhancement layer by clipping coordinate information to the range of the reference region of the base layer, using information from neighboring blocks or default syntax values when the corresponding block is outside the reference region, and allowing different smallest coding unit sizes and boundary padding sizes for the enhancement layer compared to the base layer.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the base layer is coded with a different standard or has a larger frame size, then the base layer can maintain compatibility with existing standards, but the enhancement layer cannot perform inter-layer prediction for blocks corresponding to areas outside the base layer reference region
Solution Approach 1:
The patent introduces an intermediary mechanism by mapping enhancement layer blocks to the base layer reference region through coordinate clipping. When a block in the enhancement layer corresponds to an area outside the base layer reference region, the system clips the coordinates to find the nearest available block within the reference region, enabling prediction information to be obtained through this intermediary mapping process.
2Device complexity
If the enhancement layer uses the same frame size and coding unit configuration as the base layer, then the coding structure remains simple, but the enhancement layer cannot accommodate flexible scaling requirements or different boundary padding needs
Solution Approach 1:
The patent implements dynamic adaptability by allowing the enhancement layer to have different frame sizes, smallest coding unit sizes, and boundary padding sizes compared to the base layer. The system dynamically adjusts the mapping relationship between enhancement layer blocks and base layer reference regions based on these different configurations, enabling flexible scaling while maintaining efficient prediction.
3Productivity
If inter-layer prediction is performed for all blocks in the enhancement layer, then coding efficiency is maximized, but computational complexity increases when corresponding blocks are outside the reference region
Solution Approach 1:
The patent applies local quality optimization by treating blocks differently based on their position relative to the base layer reference region. For blocks within the reference region, full inter-layer prediction is performed. For blocks outside the reference region, coordinate clipping is applied to map them to the nearest available block within the region, providing a localized solution that maintains coding efficiency while managing computational complexity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An apparatus configured to code video information includes a memory unit and a processor in communication with the memory unit. The memory unit is configured to store video information associated with an enhancement layer having a first block and a base layer having a second block, the second block in the base layer corresponding to the first block in the enhancement layer. The processor is configured to predict, by inter layer prediction, the first block in the enhancement layer based on information derived from the second block in the base layer. At least a portion of the second block is located outside of a reference region of the base layer, the reference region being available for use for the inter layer prediction of the first block. The processor may encode or decode the video information.