Tomosynthesis Noise Reduction via Filter-Adaptive Parameter Adjustment

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing tomography techniques face challenges in reducing noise in images reconstructed from limited projection angles, leading to redundant smoothing effects and loss of physical significance in CT values, especially in low-dose CT imaging.

Innovation Solution

An image processing apparatus that performs noise reduction by selecting appropriate image processing parameters based on the frequency characteristics of the reconstruction filter, using multiresolution decomposition and weighted coefficients to effectively suppress noise in high-frequency components, allowing for noise reduction regardless of the reconstruction filter type, even in tomosynthesis with incomplete projection sets.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If noise reduction processing is performed on projected images before reconstruction, then noise is reduced in the final image, but redundant smoothing effects occur when combined with certain reconstruction filters

Engineering Contradiction:
Improvenoise reductionVSAvoidimage smoothing effect
Core Design Contradiction:
Measurement precisionVSManufacturing precision

Solution Approach 1:

The patent changes the parameter of noise reduction strength based on the type of reconstruction filter selected. When a smoothing filter is chosen, the noise reduction processing applies weaker smoothing to avoid redundant effects. When a sharp filter is selected, stronger noise reduction is applied. This dynamic parameter adjustment resolves the contradiction by adapting the noise reduction intensity to match the reconstruction filter's characteristics.

Inventive Principle:
Principle #35Parameter changes

2Loss of time

If reconstruction is performed using limited projection angles (tomosynthesis), then imaging time and radiation exposure are reduced, but noise reduction becomes difficult due to loss of CT value physical significance

Engineering Contradiction:
Improveimaging timeVSAvoidnoise estimation accuracy
Core Design Contradiction:
Loss of timeVSMeasurement precision

Solution Approach 1:

The patent changes the approach to noise estimation by not relying on CT value physical significance. Instead, it uses statistical methods that estimate noise based on the projection data characteristics and reconstruction process, which remains valid even with limited angles. This allows noise reduction to be performed effectively in tomosynthesis while maintaining the benefits of reduced imaging time.

Inventive Principle:
Principle #35Parameter changes

3Object-affected harmful factors

If radiation exposure amount is reduced, then patient dose is minimized, but noise in the captured images increases

Engineering Contradiction:
Improveradiation exposureVSAvoidimage noise level
Core Design Contradiction:
Object-affected harmful factorsVSMeasurement precision

Solution Approach 1:

The patent applies noise reduction processing to the projected images before reconstruction is performed. This preliminary action removes noise from the input data, allowing subsequent reconstruction to produce high-quality images even when the original projection images were captured with low radiation exposure. This enables low-dose imaging while maintaining diagnostic image quality.

Inventive Principle:
Principle #10Preliminary action

4Measurement precision

If strong noise reduction processing is applied, then noise is effectively reduced, but image detail and diagnostic information may be lost

Engineering Contradiction:
Improvenoise reduction effectivenessVSAvoidimage detail
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent dynamically adjusts the noise reduction parameter based on the selected reconstruction filter type. Smoothing filters that preserve edges and details allow for stronger noise reduction without significant loss of diagnostic information. Sharp filters use more conservative noise reduction to maintain image detail. This adaptive parameter selection resolves the contradiction between noise reduction effectiveness and preservation of image detail.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP2824638B1Image processing apparatus and image processing method
Publication Date: 2019.11.06 CANON KK
  • EP2824638B1 patent drawingFigure 1
  • EP2824638B1 patent drawingFigure 2
  • EP2824638B1 patent drawingFigure 3

AI summary

An image processing apparatus of the present invention performs reconstruction processing on a projected image of an object obtained by using a radiation source and a detector in order to obtain a tomographic image in which noise has been reduced regardless of the type of reconstruction filter that is used. The image processing apparatus includes a selection means that selects a reconstruction filter in accordance with an imaging technique; a setting means that sets an image processing parameter for noise reduction according to the type of the selected reconstruction filter; an image processing means that performs image processing on the projected image using the image processing parameter; and a reconstruction means that uses the selected reconstruction filter to perform reconstruction processing on the projected image resulting from the image processing and obtain a tomographic image of the object.