Image Processing Apparatus Synthesis Patches Noise Reduction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional patch-based noise reduction methods in image processing are ineffective in ensuring consistent noise reduction across varying image content due to the uneven distribution of similar patches, leading to poor overall image quality.

Innovation Solution

An image processing apparatus that sets a patch of interest, searches for similar patches, generates synthesis patches, and uses these patches along with the original and similar patches to perform noise reduction, thereby stabilizing noise reduction performance across the image.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If conventional patch-based noise reduction method is used, then noise reduction processing can be performed, but the noise reduction effect varies from area to area due to uneven distribution of similar patches

Engineering Contradiction:

Improvenoise reduction effect consistencyVSAvoidadaptability to varying image content

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent generates synthesis patches by copying and modifying existing patches. Specifically, it creates synthetic patches that mimic the texture characteristics of real patches without using actual similar patches from the image. This allows the system to maintain consistent noise reduction performance across different image regions by using generated synthetic patches instead of relying solely on the availability of similar real patches.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent changes the parameter of patch source from real image patches to synthesized patches. By generating synthetic patches through parameter transformation and combination of existing patches, the system ensures sufficient patch availability regardless of the actual image content distribution, thereby achieving consistent noise reduction effects across varying image textures and frequencies.

Inventive Principle:

Principle #35Parameter changes

2Manufacturing precision

If sufficient similar patches are required for effective noise reduction, then noise reduction performance improves, but this cannot be ensured in images with uneven texture distribution

Engineering Contradiction:

Improvenoise reduction qualityVSAvoidavailability of similar patches

Core Design Contradiction:

Manufacturing precisionVSReliability

Solution Approach 1:

The patent creates synthetic patches by copying and combining existing patches in new configurations. The synthesis unit generates patches that replicate the texture characteristics needed for noise reduction, ensuring sufficient patch availability even when similar real patches are scarce in images with uneven texture distribution.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent performs preliminary patch generation before noise reduction processing. By pre-generating synthesis patches that simulate various texture patterns, the system ensures that sufficient patches are available when needed for noise reduction, regardless of the actual image content distribution.

Inventive Principle:

Principle #10Preliminary action

3Productivity

If patch-based noise reduction is applied to images with uneven texture, then processing can be performed, but regional unevenness in noise reduction effects occurs

Engineering Contradiction:

Improvenoise reduction processing capabilityVSAvoiduniformity of noise reduction effect

Core Design Contradiction:

ProductivityVSStability of the object's composition

Solution Approach 1:

The patent changes the patch source parameter from real image patches to synthesized patches. The synthesis unit generates patches with controlled texture characteristics that ensure uniform noise reduction performance across different image regions, eliminating regional unevenness while maintaining processing capability.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent applies different synthesis strategies for different patch types based on local image characteristics. The synthesis unit generates appropriate synthetic patches tailored to local texture conditions, ensuring that each region receives appropriate noise reduction treatment while maintaining overall uniformity.

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS11393074B2Image processing apparatus, image processing method, and storage medium

Publication Date: 2022.07.19 CANON KK

AI summary

There is provided an image processing apparatus that performs noise reduction processing by using a patch obtained by dividing an image into small areas. The image processing apparatus includes a setting unit configured to set, for the image, a patch of interest, a search unit configured to search the image for a similar patch based on the patch of interest, a generation unit configured to generate one or more synthesis patches different from a patch included in the image, based on at least one of the patch of interest and the similar patch, and a noise reduction unit configured to reduce noise in the patch of interest by using a patch group including the patch of interest, the similar patch, and the one or more synthesis patches.