Video Block Merge Processing for Parallel Motion Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding/decoding technologies face inefficiencies in performing inter prediction and deriving merge candidates for blocks of varying shapes and sizes, particularly in high-resolution video processing, leading to increased data transmission and storage costs.
Innovation Solution
A method and apparatus for deriving spatial merge candidates for current blocks, generating a merge candidate list, obtaining motion information, and performing motion compensation based on predefined shapes or sizes, allowing parallel processing of merge operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional video encoding/decoding techniques are used for high-resolution images, then image quality is improved, but transmission and storage costs increase due to increasing data amounts
Solution Approach 1:
The patent segments the current block into multiple sub-blocks and performs separate merge candidate derivation for each sub-block. This segmentation allows for more precise local motion compensation while reducing the overall computational burden by processing smaller, independent regions rather than the entire block at once.
Solution Approach 2:
The patent introduces a new dimension of processing by considering both spatial and temporal dimensions for merge candidate derivation. It derives merge candidates from neighboring blocks in the spatial dimension and from reference pictures in the temporal dimension, creating a multi-dimensional approach to motion compensation that improves compression efficiency.
2Adaptability or versatility
If inter prediction is performed for blocks of varying shapes and sizes, then adaptability is improved, but processing complexity increases
Solution Approach 1:
The patent applies local quality by deriving merge candidates specifically for each sub-block based on its local characteristics and position. Different sub-blocks can have different merge candidate lists tailored to their local motion patterns, allowing the system to adapt to varying block shapes and sizes while maintaining manageable processing complexity through localized processing.
Solution Approach 2:
The patent introduces dynamics by allowing the merge candidate derivation process to adapt dynamically to different block configurations. The system can handle blocks of varying shapes and sizes by dynamically adjusting the sub-block segmentation and merge candidate selection based on the specific block geometry and motion characteristics.
3Measurement precision
If merge operations are performed sequentially for each block, then processing accuracy is maintained, but processing speed decreases
Solution Approach 1:
The patent segments the merge operation into independent sub-block operations that can be performed in parallel. By dividing the current block into multiple sub-blocks and deriving merge candidates for each independently, the system enables parallel processing while maintaining the accuracy of individual merge operations, thus improving processing speed without sacrificing precision.
Solution Approach 2:
The patent performs preliminary actions by pre-deriving merge candidates from neighboring blocks and reference pictures before the actual motion compensation. This preliminary derivation of merge candidate lists allows subsequent parallel processing to proceed efficiently, as the candidate selection work is completed in advance for each sub-block.
Data Source
AI summary
A method for decoding a video according to the present invention may comprise: deriving a spatial merge candidate for a current block, generating a merge candidate list for the current block based on the spatial merge candidate, obtaining motion information for the current block based on the merge candidate list, and performing motion compensation for the current block using the motion information. Herein, if the current block does not have a pre-defined shape or does not have a size equal to or greater than a pre-defined size, the spatial merge candidate of the current block is derived based on a block having the pre-defined shape or having a size equal to or greater than the pre-defined size, the block comprising the current block.