Inter-layer Prediction for Spatial Scalability in Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current scalable video coding standards, such as Annex F of the H.264/AVC standard, face inefficiencies in inter-layer prediction for macroblock mode and motion vectors, especially in extended spatial scalability scenarios, leading to increased computation complexity due to unnecessary small partitions and inaccurate predictions.
Innovation Solution
The proposed system improves inter-layer prediction by using reference frame index and motion vector information for block merging, allowing blocks with different partitions to be merged if they share the same reference frame index and similar motion vectors, and utilizes multiple representative pixels to derive more accurate partition and motion vector information from multiple base layer blocks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If blocks with different partitions are merged based on reference frame index and motion vector information, then coding efficiency is improved and computation complexity is reduced, but prediction accuracy for enhancement layer macroblocks deteriorates
Solution Approach 1:
The patent merges reference blocks from the base layer that have the same reference frame index and similar motion vectors, even if they have different partitions. This merging reduces the number of separate predictions needed and decreases computation complexity while maintaining adequate prediction accuracy through the selective merging criteria.
Solution Approach 2:
The patent changes the merging criteria from requiring identical partitions to requiring matching reference frame indices and similar motion vectors. This parameter change allows more blocks to be merged, improving coding efficiency while the similarity requirement on motion vectors maintains sufficient prediction accuracy.
2Measurement precision
If multiple base layer blocks are used to derive partition and motion vector information, then prediction accuracy is improved, but computation complexity increases
Solution Approach 1:
The patent merges multiple base layer blocks that share the same reference frame index and have similar motion vectors into a single prediction unit. This reduces the number of separate processing operations needed while still utilizing information from multiple source blocks, thereby improving prediction accuracy without proportionally increasing computation complexity.
Solution Approach 2:
The patent applies different processing treatments to different groups of base layer blocks based on their reference frame indices and motion vector similarities. Blocks with matching criteria are merged and processed together, while blocks with different criteria are processed separately, optimizing the balance between prediction accuracy and computation complexity for each local group.
Data Source
Figure 1~2
Figure 3~4(a)
Figure 4(b)
AI summary
An improved system and method for providing improved inter-layer prediction for extended spatial scalability in video coding, as well as improving inter-layer prediction for motion vectors in the case of extended spatial scalability. In various embodiments, for the prediction of macroblock mode, the actual reference frame index and motion vectors from the base layer are used in determining if two blocks should be merged. Additionally, multiple representative pixels in a 4x4 block can be used to represent each 4x4 block in a virtual base layer macroblock. The partition and motion vector information for the relevant block in the virtual base layer macroblock can be derived from all of the partition information and motion vectors of those 4x4 blocks.