Adaptive Bit-Width Compression for Varying Data Trends
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data compression methods fail to efficiently reduce data volume while preserving information integrity, particularly in systems with varying data variation trends.
Innovation Solution
A compression apparatus that decides a bit width based on data variation trends, calculates deviations within data groups, and generates frames including these deviations and bit widths, ensuring efficient data reduction without information loss.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a fixed bit width is used for all data, then the device complexity is reduced, but the data compression efficiency deteriorates when data variation trends vary
Solution Approach 1:
The patent applies dynamics by making the bit width adaptive rather than fixed. The bit width is dynamically adjusted based on the actual variation trend of the data being compressed. This allows the compression algorithm to optimize its performance for different data patterns without requiring complex manual configuration, resolving the contradiction between compression efficiency and algorithm complexity.
Solution Approach 2:
The patent changes the parameter of bit width from a static value to a dynamic value that varies according to data characteristics. By analyzing data variation trends and adjusting the bit width parameter accordingly, the system achieves better compression efficiency for diverse data types while maintaining reasonable algorithmic complexity through systematic parameter adaptation.
2Loss of information
If a larger bit width is used, then information integrity is preserved, but the data volume increases
Solution Approach 1:
The patent applies local quality by using different bit widths for different segments of data based on their individual variation characteristics. Instead of applying a uniform bit width to all data, the system analyzes local data patterns and assigns appropriate bit widths locally, thereby preserving information where needed while minimizing data volume where variations are small.
Solution Approach 2:
The patent dynamically changes the bit width parameter according to data variation trends. When data variations are large, a larger bit width is used to preserve information integrity. When variations are small, a smaller bit width suffices, reducing data volume. This adaptive parameter adjustment resolves the contradiction between information preservation and data size.
3Measurement precision
If data is divided into smaller groups, then compression precision is improved, but the processing time increases
Solution Approach 1:
The patent applies segmentation by dividing the data stream into groups and analyzing variation trends within each group. This allows for more precise compression decisions at the group level, improving compression precision. The segmentation approach balances processing time by working with manageable data chunks rather than attempting to analyze entire data sets at once.
Solution Approach 2:
The patent dynamically determines the appropriate group size and bit width based on actual data variation patterns. This dynamic approach allows the system to achieve high compression precision when needed while avoiding unnecessary processing overhead when data patterns are simple, thereby balancing precision and processing time.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
(Solution) Provided is a compression apparatus including: a decision unit which decides, based on a variation trend of data in a data string, a bit width indicating a deviation of data; a calculation unit which calculates, with the bit width, deviations of respective pieces of data in a data group obtained by dividing the data string; a generation unit which generates a frame including: the deviations of the respective pieces of data in the data group; and information indicating the bit width; and a specification unit which specifies, from the data string, at least one continuous piece of data, a deviation of which falls within the bit width, as the data group.