Image Processing Apparatus Lossless Compression via Frequency-Based Bit Conversion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies for inspecting structures like bridge beams face inefficiencies in data compression, leading to insufficient data reduction and lossless compression challenges when determining damage presence.
Innovation Solution
A system that specifies occurrence frequencies of gradation values in image data, extracts frequently occurring gradation values, generates correspondence tables for bit conversion, and encodes pixels using these tables to produce low-capacity, lossless compressed data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of substance
If conventional compression methods are applied to image data from structure inspection, then data capacity is reduced, but compression ratio is insufficient and damage information may be lost
Solution Approach 1:
The patent changes the parameter representation by converting gradation values from fixed-bit representation to variable-bit coded values based on occurrence frequency. Frequently occurring gradation values are assigned shorter coded values, while less frequent ones retain longer representations. This parameter transformation achieves both compression and preservation of damage information by adapting the encoding to the actual data distribution.
Solution Approach 2:
The patent performs preliminary analysis of the image data to determine occurrence frequencies of gradation values before compression encoding. By pre-calculating which gradation values appear most frequently and creating a correspondence table in advance, the system optimizes the compression strategy beforehand, ensuring that the most common values (including potential damage indicators) are encoded efficiently while maintaining accuracy.
2Productivity
If fixed-bit conversion is used for all pixels, then processing is simple, but data reduction efficiency is low
Solution Approach 1:
The patent segments the image data processing into distinct stages: frequency analysis, correspondence table generation, and differential encoding. It also segments the pixel population into groups based on their gradation value frequencies. This segmentation allows the system to apply different encoding strategies to different pixel groups, achieving high compression efficiency while managing complexity through structured processing steps.
Solution Approach 2:
The patent introduces dynamic encoding where the bit representation of gradation values changes based on their occurrence frequency. Instead of a static fixed-bit conversion, the system dynamically assigns coded values of varying bit lengths according to how frequently each gradation value appears in the image data. This dynamic approach maximizes compression efficiency by adapting to the actual data characteristics.
Data Source
AI summary
A computer-readable recording medium storing a program that causes a computer to execute a process, the process includes specifying occurrence frequencies of respective gradation values with regard to pixels included in image data and represented by gradation values of a predetermined bit count; extracting a predetermined number of gradation values from a gradation value having a high occurrence frequency in a descending order; generating correspondence information for performing bit conversion of the extracted gradation values into coded values of a bit count in accordance with the predetermined number; and encoding the image data by performing bit conversion of first pixels having any one of the predetermined number of gradation values among the pixels based on the correspondence information, and performing bit conversion of second pixels having any one of gradation values other than the predetermined number of gradation values among the pixels.