Radiological Image Denoising With Spatial Blur Maps

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image processing methods fail to effectively remove noise considering changes in spatial blur, leading to insufficient noise removal effects.

Innovation Solution

An image processing method that includes generating a spatial blur map to indicate the distribution of spatial blur of noise, and inputting this map along with the image into a trained model constructed through machine learning to execute noise removal.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If noise removal is performed based on luminance value and standard deviation of pixel values, then noise removal is achieved considering noise spread in luminance direction, but sufficient noise removal effect cannot be obtained when spatial blur of noise changes in the image

Engineering Contradiction:

Improvenoise removal effectVSAvoidadaptability to spatial blur changes

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent introduces a spatial blur map as a new dimension of information to complement the existing luminance-based noise evaluation. By adding the spatial domain (horizontal and vertical blur directions) to the luminance domain, the system achieves comprehensive noise characterization that adapts to varying spatial blur conditions while maintaining reliable noise removal performance.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

2Adaptability or versatility

If a spatial blur map is generated to capture spatial blur distribution, then adaptability to spatial blur changes is improved, but device complexity increases

Engineering Contradiction:

Improveadaptability to spatial blur changesVSAvoidprocessing complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The spatial blur map is segmented into distinct horizontal and vertical blur components, allowing independent calculation and processing. This segmentation enables the system to handle spatial blur variations through separate, manageable operations rather than a single complex operation, reducing overall processing complexity while maintaining adaptability.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The spatial blur map is generated as a preliminary step before the main noise removal processing. By pre-calculating and storing the spatial blur characteristics in a separate map structure, the system prepares the necessary information in advance, which simplifies the subsequent noise removal process and reduces real-time computational complexity.

Inventive Principle:

Principle #10Preliminary action

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

Enables effective noise removal by addressing changes in spatial blur, enhancing the noise removal process.

Implementation Method 1

acquiring an image obtained by irradiating a subject with an energy beam and capturing an image of the energy beam transmitted through the subject

Methodology Applied

Scientific EffectX-ray transmission: X-Ray

Data Source

PatentEP4623827A1Image processing method, training method, trained model, radiological image processing module, radiological image processing program, and radiological image processing system

Publication Date: 2025.10.01 HAMAMATSU PHOTONICS KK

Figure 1
Figure 2
Figure 3

AI summary

An image processing method includes an image acquisition step of acquiring an image obtained by irradiating a subject F with an energy beam and capturing an image of the energy beam transmitted through the subject F, a spatial blur map generation step of generating a spatial blur map indicating a distribution of spatial blur of noise based on the image, and a processing step of inputting the image and the spatial blur map into a trained model 207 constructed in advance through machine learning and executing image processing for removing noise from the image.