CT Image Denoising via Iterative PCA and Interpolation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing CT image denoising methods using principal component analysis often result in excessive smoothness and loss of fine tissues, compromising image quality.

Innovation Solution

A method involving generating an original CT image with a higher pixel count than the target, applying iterative denoising with algorithms like non-local means filtering, and image fusion to retain fine tissues, followed by compression using interpolation to achieve a denoised image with reduced noise and preserved details.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Object-affected harmful factors

If principal component analysis algorithm is used for denoising, then noise is reduced, but fine tissues are lost and image becomes excessively smooth

Engineering Contradiction:

Improvenoise levelVSAvoidfine tissue retention

Core Design Contradiction:

Object-affected harmful factorsVSManufacturing precision

Solution Approach 1:

The patent segments the denoising process into multiple iterations, where each iteration applies PCA with progressively adjusted parameters. This allows gradual noise reduction while preserving fine tissues that would be lost in a single aggressive denoising pass.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent dynamically adjusts the denoising strength parameter across multiple iterations, starting with stronger denoising and gradually reducing intensity. This dynamic approach allows effective noise removal while preserving fine anatomical details through adaptive parameter control.

Inventive Principle:

Principle #15Dynamics

2Loss of energy

If tube current or tube voltage is lowered, then X-ray radiation dose is reduced, but blocky and granular noises increase

Engineering Contradiction:

ImproveX-ray radiation doseVSAvoidblocky and granular noises

Core Design Contradiction:

Loss of energyVSObject-affected harmful factors

Solution Approach 1:

The patent converts the harmful noise introduced by low-dose scanning into a manageable problem by applying PCA denoising. The algorithm transforms the noisy low-dose image data into a denoised output, effectively converting the harmful noise artifact into an opportunity for enhanced image quality through computational processing.

Inventive Principle:

Principle #22Blessing in disguise (Convert harm into benefit)

Solution Approach 2:

The patent introduces PCA denoising as an intermediary processing step between low-dose image acquisition and final image output. This intermediary algorithmic processing layer mediates between the unavoidable noise from low-dose scanning and the requirement for diagnostic image quality, bridging the gap through mathematical transformation.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS10783614B2Denoising CT image

Publication Date: 2020.09.22 BEIJING NEUSOFT MEDICAL EQUIP CO LTD

AI summary

Methods, devices and a machine-readable storage medium for denoising a Computed Tomography (CT) image are provided. In one aspect, a method of denoising a CT image includes: generating an original CT image according to raw data which is obtained by scanning a subject, where a pixel count of the original CT image is greater than a preset target pixel count; obtaining a denoised image by denoising the original CT image; and obtaining a target image as a CT image of the subject by compressing the denoised image to the preset target pixel count based on an interpolation algorithm.