Image Deformation Dispersion for Artifact Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image deformation processing methods in electrophotographic processes often result in deteriorated image quality when multiple deformation processes, such as enlargement, rotation, and skew correction, are overlapped, due to regular pixel insertion patterns causing texture artifacts and visible steps in the image.
Innovation Solution
An image deformation processing device and method that calculates and sums shift amounts from multiple deformation processes, using a dispersion threshold table to adjust pixel positions and correct cumulative histogram non-linearity, ensuring image quality is maintained even with overlapping processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If pixel insertion positions are arranged on identical lines during enlargement, then the enlargement process is simple, but texture artifacts appear and image quality deteriorates
Solution Approach 1:
The patent applies dispersion in the main scanning direction (horizontal dimension) to complement the existing sub-scanning direction (vertical dimension) pixel insertion dispersion. By introducing dispersion along another dimension, the patent effectively distributes pixel insertion positions across both dimensions, reducing regularity patterns while maintaining processing feasibility.
Solution Approach 2:
The patent introduces asymmetric dispersion patterns where the dispersion amounts in the main scanning direction differ from those in the sub-scanning direction. This asymmetric approach creates irregular pixel distribution that breaks the symmetry of regular grid patterns, thereby reducing texture artifacts while maintaining processing simplicity.
2Adaptability or versatility
If multiple deformation processes are applied separately with individual dispersion, then each process can be optimized independently, but cumulative steps and image quality deterioration occur
Solution Approach 1:
The patent merges the dispersion operations of multiple deformation processes into a single unified dispersion step. Instead of applying dispersion separately after each deformation process, the patent accumulates all deformation transformations first, then applies a single dispersion operation based on the total transformation matrix. This combining approach eliminates cumulative steps while preserving the versatility of individual process optimizations.
Solution Approach 2:
The patent performs preliminary accumulation of all deformation transformations before applying dispersion. By calculating the total transformation matrix that combines all individual deformations first, the system prepares a unified basis for subsequent dispersion, preventing the accumulation of intermediate steps that would degrade image quality.
3Manufacturing precision
If pixel insertion positions are dispersed to improve image quality, then texture artifacts are reduced, but the processing complexity increases
Solution Approach 1:
The patent changes the parameters of the transformation matrix to incorporate dispersion information directly into the deformation calculation. By integrating dispersion amounts into the transformation matrix parameters, the patent eliminates the need for separate dispersion processing steps, thereby reducing overall processing complexity while maintaining image quality improvements.
Data Source
AI summary
Disclosed is an image deformation processing device including: a shift amount calculating unit configured to calculate the difference between the coordinate of the input image which is mapped to the position of the specified pixel by the deformation process and the coordinate of the specified pixel as a shift amount, for each specified pixel in the output image; and an output unit configured to obtain the threshold corresponding to the coordinate of the specified pixel in the second direction from the dispersion threshold table, to calculate a pixel shift amount by comparing the obtained threshold with a decimal component of the shift amount for the specified pixel and by rounding the decimal component of the shift amount, and to output the pixel in the input image, which has a coordinate shifted from the coordinate of the specified pixel by the pixel shift amount, as the specified pixel.