Image Processing Multi-Valued Tone Consistency Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems that reduce the gradient number of input image data using dither matrices often result in unnatural output images, particularly when converting gradation images, as they fail to maintain the tone consistency across all regions, leading to a feeling of granulated material inconsistency.
Innovation Solution
The proposed image processing system employs a correction mechanism to determine whether the calculated value α is an integer, adjusting it to ensure it is not, and using this adjusted value to calculate output image data using the formula O(x,y) = int[α + d(xd,yd) / (gmax + 1)], thereby maintaining tone consistency and avoiding unnatural output.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multi-valued processing is performed using the formula O(x,y)=int[α+d(xd,yd)/(gmax+1)] with α=(n-1)/(m-1)*I(x,y), then processing speed is improved and memory capacity is reduced, but unnatural regions appear in the output image where tone consistency is lost
Solution Approach 1:
The patent modifies the parameter α by adding a correction term δ(x,y) to the original formula α=(n-1)/(m-1)*I(x,y), transforming it into α'=(n-1)/(m-1)*I(x,y)+δ(x,y). This parameter change ensures that α' is never an integer, preventing the formation of unnatural single-gradation regions while maintaining the efficiency of the single-comparison multi-valued processing method.
2Manufacturing precision
If multiple dither matrices are used for multi-valued processing as proposed in U.S. Pat. No. 5,640,249, then tone consistency is maintained, but processing speed decreases and memory capacity increases
Solution Approach 1:
The patent merges the functionality of multiple dither matrices into a single dither matrix by incorporating the correction term δ(x,y) directly into the calculation of α'. This allows the system to achieve the tone consistency benefits of multiple matrices while maintaining the processing speed and memory efficiency of a single matrix approach.
Data Source
AI summary
An image processing system receives an input gradient number m, input image data I(x,y) whose input gradient number is m, and an output gradient number n and calculates output image data O(x,y) having an output gradient number of n on the basis of the following formula (1).O(x,y)=int[α+d(xd,yd)/(gmax+1)] (1)When α is an integer, the output image data O(x,y) is calculated on the basis of a formula obtained by substituting the calculated α′ for α in the above formula (1), whereas when α is not an integer, the output image data O(x,y) is calculated on the basis of the above formula (1).