Large Video Block Transform Coefficient Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression schemes face challenges in efficiently processing large video and image blocks, which require significant computing resources and storage, especially with high-resolution content, due to the processing of large transform sizes exceeding practical limits.
Innovation Solution
The method involves generating a prediction residual block, transforming its pixel values into coefficients, and if the coefficients exceed a threshold cardinality, discarding some to fit within a maximum transform block size during encoding, and then adding coefficients during decoding to restore the original size, thereby reducing computational and storage needs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If large transform sizes are used to process high-resolution video blocks, then processing accuracy and quality are improved, but computational resources and storage requirements increase significantly
Solution Approach 1:
The patent divides large transform blocks into smaller sub-blocks for processing. Instead of applying a single large transform to the entire block, the method segments the block and processes each segment separately with smaller transforms, reducing the computational complexity from O(N²) to O((N/k)² × k) where k is the number of segments, while still maintaining the ability to represent large block characteristics
Solution Approach 2:
The patent introduces a new dimension by adding synthetic coefficients that represent the relationship between sub-blocks. These additional coefficients capture inter-sub-block correlations that would otherwise require much larger transform sizes, enabling accurate representation of large blocks through a combination of local sub-block transforms and global relationship coefficients
2Measurement precision
If large transform sizes are used to process high-resolution video blocks, then processing accuracy and quality are improved, but storage requirements increase significantly
Solution Approach 1:
By segmenting large transform blocks into smaller sub-blocks, the patent reduces the number of coefficients that need to be stored and transmitted. Each sub-block requires fewer coefficients than a proportional section of a large block would require, and the total storage is reduced due to the sub-additive nature of transform coefficients
Solution Approach 2:
The patent adds a new dimensional layer by introducing coefficients that represent inter-sub-block relationships. These coefficients efficiently encode the global structure of large blocks without requiring the full coefficient set that a traditional large transform would generate, achieving compact representation through multi-scale decomposition
3Productivity
If transform block sizes are reduced to fit within practical limits, then computational and storage requirements are reduced, but processing accuracy deteriorates
Solution Approach 1:
The patent uses segmentation to process large blocks through multiple smaller transforms, achieving computational efficiency. The key innovation is that the segmentation is combined with additional relationship coefficients that compensate for the loss of global context, thereby maintaining processing accuracy that would otherwise require much larger block sizes
Solution Approach 2:
The patent changes the parameter set by introducing new coefficients that represent inter-sub-block relationships. This parameter expansion allows small sub-block transforms to collectively represent the characteristics of large blocks, effectively decoupling the transform block size from the representational capacity and enabling efficient processing without accuracy loss
Data Source
AI summary
Improved transforms are used to encode and decode large video and image blocks. During encoding, a prediction residual block having a large size (e.g., larger than 32×32) is generated. The pixel values of the prediction residual block are transformed to produce transform coefficients. After determining that the transform coefficients exceed a threshold cardinality representative of a maximum transform block size (e.g., 32×32), a number of the transform coefficients are discarded such that a remaining number of transform coefficients does not exceed the threshold cardinality. A transform block is then generated using the remaining number. During decoding, after determining that the transform coefficients exceed the threshold cardinality, a number of new coefficients are added to the transform coefficients such that a total number of transform coefficients exceeds the threshold cardinality. The transform coefficients are then inverse transformed into a prediction residual block having a large size.