Dynamic Green Noise Gain for Error Diffusion Artifact Reduction

Solution Overview

Problem

Conventional error diffusion processing introduces periodic patterns or artifacts in images with intermediate density during binarization, leading to a decrease in image quality due to the feedback mechanism of binarization results from peripheral pixels.

Innovation Solution

An image processing apparatus that adjusts noise components and gain values dynamically, including an input noise offset adjusting section, a green noise gain adjusting section, and an input noise gain adjusting section, to minimize the occurrence of artifacts by optimizing the addition of noise and error values during the binarization process.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If green noise feedback from peripheral pixels is superimposed on the target pixel during binarization, then density reproductivity is improved and image uniformity is enhanced, but periodic patterns and artifacts occur in intermediate density images

Engineering Contradiction:

Improvedensity reproductivityVSAvoidperiodic patterns and artifacts

Core Design Contradiction:

Manufacturing precisionVSObject-affected harmful factors

Solution Approach 1:

The patent applies parameter changes by dynamically adjusting the gain value of the green noise filter based on the density of the target pixel. When the target pixel density is high (close to black), the gain is increased to enhance density reproductivity. When the density is low (close to white), the gain is decreased to suppress periodic patterns and artifacts. This adaptive parameter adjustment resolves the contradiction by optimizing the green noise feedback effect for different density regions.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent implements dynamics by making the green noise filter gain variable rather than fixed. The gain value changes dynamically according to the input pixel density, allowing the system to adapt its behavior to different image regions. This dynamic adjustment enables the system to maintain high density reproductivity in dark regions while minimizing artifacts in light regions.

Inventive Principle:

Principle #15Dynamics

2Productivity

If the green noise filter uses fixed weighting factors for peripheral pixels, then the processing is simple and fast, but artifacts occur when peripheral pixel arrangements coincide with filter patterns

Engineering Contradiction:

Improveprocessing speedVSAvoidartifacts from periodic patterns

Core Design Contradiction:

ProductivityVSObject-affected harmful factors

Solution Approach 1:

The patent applies dynamics by making the green noise filter gain variable rather than fixed. The gain value changes dynamically according to the input pixel density, allowing the system to adapt its behavior to different image regions. This dynamic adjustment enables the system to maintain high density reproductivity in dark regions while minimizing artifacts in light regions.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent applies parameter changes by dynamically adjusting the gain value of the green noise filter based on the density of the target pixel. When the target pixel density is high (close to black), the gain is increased to enhance density reproductivity. When the density is low (close to white), the gain is decreased to suppress periodic patterns and artifacts. This adaptive parameter adjustment resolves the contradiction by optimizing the green noise feedback effect for different density regions.

Inventive Principle:

Principle #35Parameter changes

3Manufacturing precision

If error diffusion processing is applied to binarize multi-value image data, then image quality is improved through error feedback, but black pixels become concentrated and periodic patterns emerge

Engineering Contradiction:

Improveimage qualityVSAvoidpixel distribution uniformity

Core Design Contradiction:

Manufacturing precisionVSStability of the object's composition

Solution Approach 1:

The patent applies parameter changes by dynamically adjusting the gain value of the green noise filter based on the density of the target pixel. When the target pixel density is high (close to black), the gain is increased to enhance density reproductivity. When the density is low (close to white), the gain is decreased to suppress periodic patterns and artifacts. This adaptive parameter adjustment resolves the contradiction by optimizing the green noise feedback effect for different density regions.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent applies local quality by applying different green noise gain values to different regions of the image based on local pixel density. High-density regions receive higher green noise gain to improve density reproductivity, while low-density regions receive lower gain to maintain uniformity and prevent artifacts. This localized parameter adjustment optimizes image quality while maintaining pixel distribution stability in different regions.

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS9681024B2Image processing apparatus, control method, and computer-readable recording medium configured to perform error diffusion process having and adder to add an error intergrated value diffused to the target pixel, the green noise and the predetermined noise, to the pixel value of the target pixel

Publication Date: 2017.06.13 KONICA MINOLTA INC

AI summary

An image processing apparatus receiving a pixel value of a multivalue image includes: an input noise generating section that generates a noise; a green noise generating section that generates a green noise from an output value of a binarized processed pixel; an adder section that adds an error integrated value, the green noise and the noise, to the pixel value of the target pixel; the threshold value processing section that binarizes the pixel value of the target pixel after adding the error integrated value, the green noise and the noise; a subtractor section that calculates an error value by calculating a difference between an output value of the binarized target pixel and the pixel value of the target pixel including the error integrated value and the noise; and an error integrating section that outputs the error integrated value by using the error value of the binarized processed pixel.