Prediction Error Conversion in Lossless Compression Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Lossless compression algorithms, such as JPEG-LS, face inefficiencies in achieving high compression ratios due to large prediction errors when pixel correlation is weak, leading to increased data amounts and reduced compression effectiveness.
Innovation Solution
A lossless compression-encoding device that employs differential coding, prediction error conversion by reversing the sign bit to decrease absolute values, and variable-length coding to generate shorter codes as prediction error absolute values decrease, thereby reducing overall code length.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If differential coding and variable-length coding are used in lossless compression, then the compression algorithm can reconstruct original data completely, but the prediction error absolute value becomes large when pixel correlation is weak, resulting in large code length and insufficient compression ratio
Solution Approach 1:
The patent changes the parameter of prediction error by introducing sign bit reversal operation. When the sign bit is reversed, the prediction error value transforms (e.g., from a large positive value to a smaller negative value or vice versa), which directly reduces the absolute value of prediction error. This parameter transformation allows the variable-length coding to generate shorter codes while maintaining complete data reconstructability, thus resolving the contradiction between reliability and data quantity.
2Measurement precision
If the absolute value of prediction error is large, then the variable-length code length increases, but this leads to reduced compression effectiveness and larger total data amount
Solution Approach 1:
The patent applies inversion by reversing the sign bit of the prediction error. This inversion operation transforms the prediction error value in a way that reduces its absolute value. For example, a prediction error of +127 becomes -127 or a smaller magnitude value after sign bit reversal depending on the specific implementation. This inverted transformation directly addresses the issue of large prediction error magnitudes, enabling more efficient variable-length coding and improving compression ratio without sacrificing prediction accuracy.
Data Source
AI summary
In a lossless compression-encoding device, a differential coding part calculates a prediction value of data to be compressed, and calculates a prediction error that is a difference between the prediction value and an actual value of the data to be compressed. A prediction error conversion part performs reversal of a sign bit of the prediction error output by the differential coding part in case that the reversal of the sign bit of the prediction error decreases an absolute value of the prediction error. A variable-length coding part performs variable-length coding on the prediction error that has been processed by the prediction error conversion part to generate a variable-length code representing the prediction error, and outputs the variable-length code as compressed data in such a manner that a code length of the compressed data decreases as the absolute value of the prediction error decreases.