Pixel-Adaptive Edge-Preserving Smoothing for Uneven Luminance Noise

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional noise removal techniques using edge-preserving smoothing filters struggle to effectively remove noise from images with varying luminance distributions, leading to incomplete noise removal or blurring of edges.

Innovation Solution

An image processing method that calculates a relationship between luminance values and statistical values, sets parameters for an edge-preserving smoothing filter based on these values for each pixel, and applies the filter to the image, optimizing noise removal across the entire image.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If a fixed parameter edge-preserving smoothing filter is used, then the filter structure is simple, but noise removal effectiveness deteriorates in images with varying luminance distributions

Engineering Contradiction:
Improvefilter implementation simplicityVSAvoidnoise removal effectiveness
Core Design Contradiction:
Ease of manufactureVSManufacturing precision

Solution Approach 1:

The patent applies dynamics by transitioning from fixed parameters to dynamic parameters that adapt to local image characteristics. Specifically, the standard deviation σ is calculated locally for each pixel based on the luminance values in its neighborhood, allowing the filter parameters to change dynamically according to the local luminance distribution. This resolves the contradiction by making the filter adaptive to varying luminance regions while maintaining a relatively simple implementation framework.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent implements local quality by setting different filter parameters for different regions of the image based on local luminance statistics. Each pixel receives a customized standard deviation value calculated from its local neighborhood, ensuring that noise removal is optimized for each specific region's characteristics. This allows the filter to handle diverse luminance distributions across the image effectively.

Inventive Principle:
Principle #3Local quality

2Manufacturing precision

If parameters are adjusted manually for different regions, then noise removal effectiveness improves, but operation complexity increases

Engineering Contradiction:
Improvenoise removal effectivenessVSAvoidparameter adjustment complexity
Core Design Contradiction:
Manufacturing precisionVSEase of operation

Solution Approach 1:

The patent applies self-service by enabling the filter to automatically determine its own parameters based on the image content itself. The standard deviation for each pixel is calculated autonomously from the luminance values in its local neighborhood, without requiring external manual intervention. This resolves the contradiction by making the system self-adaptive, eliminating the need for manual parameter adjustment while achieving region-specific noise removal optimization.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent implements parameter changes by dynamically adjusting the standard deviation parameter based on local luminance statistics. Instead of using fixed or manually set parameters, the system automatically computes appropriate parameter values for each pixel based on the local image characteristics. This automated parameter adaptation improves noise removal effectiveness while maintaining ease of operation.

Inventive Principle:
Principle #35Parameter changes

3Manufacturing precision

If strong filtering is applied to remove noise, then noise removal effectiveness improves, but edge preservation deteriorates

Engineering Contradiction:
Improvenoise removal effectivenessVSAvoidedge sharpness
Core Design Contradiction:
Manufacturing precisionVSShape

Solution Approach 1:

The patent applies local quality by using different filter strengths for different regions based on local luminance characteristics. In regions with high luminance variation (likely containing edges), the locally calculated standard deviation naturally results in gentler filtering. In uniform regions, stronger filtering can be applied. This spatially adaptive approach resolves the contradiction by applying appropriate filtering strength locally, removing noise effectively while preserving edges.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP4679360A1Image processing method, image processing device, and image processing program
Publication Date: 2026.01.14 HAMAMATSU PHOTONICS KK
  • EP4679360A1 patent drawingFigure 1
  • EP4679360A1 patent drawingFigure 2
  • EP4679360A1 patent drawingFigure 3

AI summary

An image acquisition device 1 includes a relationship calculation unit 202 that calculates relationship data between pixel values in an image and a statistical value of the pixel values, and a parameter setting unit 204 and a filter processing unit 205 that process an image to be processed. The parameter setting unit 204 and the filter processing unit 205 specify the statistical value for each pixel of the image on the basis of a pixel value for each pixel of the image and the relationship data, set parameters of an edge-preserving smoothing filter for each pixel of the image on the basis of the specified statistical value, and process the image using the edge-preserving smoothing filter in which the parameters are set for each pixel of the image.