X-Ray Shadow Pixel Segmentation for Metal Artifact Reduction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing algorithms for metal artifact reduction in X-ray images fail when high-contrast objects lie partially or completely outside the reconstructable volume during 3D reconstruction, as they cannot accurately segment shadow pixels.

Innovation Solution

A method for segmenting high-contrast objects in X-ray projection images by determining the ratio of scattered radiation to primary radiation, using threshold values to classify pixels as shadow pixels, and applying morphological operations to enhance segmentation accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If conventional metal artifact reduction algorithms are used, then metal artifacts inside the reconstructable volume can be reduced, but metal objects outside the reconstructable volume cannot be accurately segmented and processed

Engineering Contradiction:

Improvesegmentation reliabilityVSAvoidhandling of metal objects outside reconstructable volume

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent segments the image into different regions based on the scatter-to-primary radiation ratio, identifying shadow pixels caused by metal objects regardless of whether they are within or outside the reconstructable volume. This allows separate processing of different image regions, enabling accurate segmentation of metal objects even when partially outside the volume of interest.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent uses the scatter-to-primary radiation ratio as an intermediary parameter to indirectly identify metal objects and shadow pixels. Instead of directly detecting metal objects or relying solely on geometric reconstruction, the method uses this radiological parameter as a mediator to locate and process affected regions accurately.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Measurement precision

If shadow pixels are identified using conventional methods, then metal artifacts can be reduced, but pixels outside the reconstructable volume are incorrectly processed or ignored

Engineering Contradiction:

Improveshadow pixel identification accuracyVSAvoidinformation loss from outside volume pixels

Core Design Contradiction:

Measurement precisionVSLoss of information

Solution Approach 1:

The patent applies a universal criterion (scatter-to-primary radiation ratio thresholding) that works for both pixels inside and outside the reconstructable volume. This multi-functional approach allows the same processing method to accurately identify shadow pixels throughout the entire image, preventing information loss from regions outside the volume of interest.

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Manufacturing precision

If high-contrast objects are segmented using brightness thresholding in 3D space, then metal objects can be identified, but segmentation fails when metal objects are outside the reconstructable volume

Engineering Contradiction:

Improvemetal object segmentation accuracyVSAvoidcomplexity of segmentation algorithm

Core Design Contradiction:

Manufacturing precisionVSDevice complexity

Solution Approach 1:

Instead of segmenting metal objects in the reconstructed 3D space using brightness thresholds (conventional approach), the patent inverts the approach by segmenting shadow pixels in the 2D projection image using scatter-to-primary radiation ratio. This inversion allows accurate identification of metal object effects even when the metal objects themselves are outside the reconstructable volume.

Inventive Principle:

Principle #13The other way round (Inversion)

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

Improves the reliability of high-contrast object segmentation and reduces image errors in 3D reconstructed X-ray images by accurately identifying shadow pixels, thereby enhancing image quality and reducing metal artifacts.

Implementation Method 1

determining a ratio of scattered radiation to be assigned to the pixel to primary radiation to be assigned to the pixel

Methodology Applied

Scientific EffectScattered radiation: Scattering

Data Source

PatentUS12536667B2Method and facility for segmentation of high-contrast objects in X-ray images

Publication Date: 2026.01.27 SIEMENS HEALTHINEERS AG

AI summary

A method for segmentation of high-contrast objects in X-ray images includes receiving an X-ray projection image through an interface. The X-ray projection image includes a plurality of pixels. An image processing processor determines, for one or more pixels of the received X-ray projection image, whether the respective pixel is a shadow pixel representing a projection mapping of a high-contrast object. High-contrast object segmentation of one or more high-contrast objects is generated for the X-ray projection image based on the shadow pixels. The determination of shadow pixels includes, for the respective pixel of the received X-ray projection image, determining a ratio of the scattered radiation intensity to be assigned to the primary radiation intensity to be assigned to the pixel, comparing the ratio with a scatter-primary threshold value, and classifying the pixel, wherein the pixel is classified as a shadow pixel if the ratio is greater than the scatter-primary threshold value.