Pixel Compression Circuit Selection for Adaptive Lossless Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data compression techniques for computing devices face challenges in achieving optimal compression ratios and power efficiency, particularly in lossless and lossy compression methods, which affect performance and circuit area, and there is a need for adaptive compression strategies that can switch between lossless and lossy compression based on target output size.
Innovation Solution
The implementation of a compression circuitry that dynamically selects between lossless and lossy compression techniques by using multiple predictors to determine the best compression method for a block of pixels, allowing for different delta widths in different regions and considering bias and decorrelation to achieve the smallest compression size, and falling back to lossy compression if lossless compression cannot meet a target output size.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If lossless compression techniques are used to achieve original data fidelity, then data accuracy is improved, but compression ratio and power efficiency deteriorate
Solution Approach 1:
The compression system dynamically switches between lossless and lossy compression modes based on target output size requirements. The circuitry evaluates multiple predictors and compression techniques in real-time, adjusting the compression strategy adaptively rather than using a fixed approach, thereby optimizing power efficiency while maintaining data accuracy when needed.
Solution Approach 2:
The system changes compression parameters dynamically by selecting different predictors (e.g., gradient predictors, neighbor predictors) and adjusting delta widths for different regions. When target output size is achieved through lossless methods, the system maintains high data accuracy; when target size cannot be met, it transitions to lossy modes with adjusted parameters to improve power efficiency.
2Quantity of substance
If multiple predictors are used to determine the best compression method, then compression ratio is improved, but device complexity increases
Solution Approach 1:
The pixel block is divided into multiple regions, and different predictors are applied to different regions based on their characteristics. The circuitry evaluates multiple predictors (gradient, neighbor, etc.) for each region and selects the best one, achieving high compression ratios without requiring all predictors to be fully implemented throughout the entire block, thus managing circuit area effectively.
Solution Approach 2:
The system implements a subset of predictors and evaluation logic only when and where needed to achieve the target output size. Rather than always using all available predictors, the circuitry performs partial evaluation and selects sufficient predictors to meet compression goals, reducing overall device complexity while maintaining effective compression ratios.
3Productivity
If adaptive compression strategies are implemented to switch between lossless and lossy modes, then productivity is improved, but device complexity increases
Solution Approach 1:
The system performs preliminary evaluation of multiple predictors and compression techniques before final compression. The circuitry quickly assesses which predictors will achieve the target output size and pre-determines the compression strategy, enabling fast switching between lossless and lossy modes without extensive computation during actual compression, thus improving productivity while managing complexity.
Solution Approach 2:
The compression circuitry is designed to perform multiple functions: evaluating gradient predictors, neighbor predictors, determining delta widths, selecting compression modes, and achieving target output sizes. This multi-functional design consolidates what could be separate complex circuits into a unified compression system, improving productivity without proportionally increasing device complexity.
Data Source
AI summary
Techniques are disclosed relating to data compression. In some embodiments, compression circuitry determines, at least partially in parallel for multiple different lossless compression techniques, a number of bits needed to represent a least compressible pixel, using that technique, in a set of pixels being compressed. The compression techniques may include neighbor, origin, and gradient techniques, for example. The compression circuitry may select one of the compression techniques based on the determined numbers of bits for the multiple compression techniques and corresponding header sizes. In some embodiments, the compression circuitry determines, for multiple regions of pixels in the set of pixels, for ones of the compression techniques, a region number of bits needed to represent a least compressible pixel, using that technique. The selection of a compression technique may be further based on region numbers of bits.