CT Image Denoising via Synthetic Sinogram Noise Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current CT scanner technologies face challenges in producing high-quality images at lower radiation doses, as iterative reconstruction systems are expensive and not compatible with all legacy systems, posing difficulties in upgrading and increasing cancer risk due to higher radiation exposure.
Innovation Solution
A method and apparatus for denoising CT images by generating a synthetic sinogram, extracting a noise component, and applying filtered back-projection to reduce noise, allowing for high-quality image production comparable to regular-dose CT images without the need for expensive upgrades.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If iterative reconstruction CT systems are used to reduce noise, then image quality is improved, but device cost and complexity increase
Solution Approach 1:
The patent creates a synthetic sinogram that copies the structure and noise characteristics of a real sinogram from a low-dose CT image. This synthetic sinogram is then processed through filtered back-projection to generate a noise component image, which is subtracted from the original low-dose image to produce a denoised image. This approach replicates the effect of iterative reconstruction without requiring the complex iterative reconstruction hardware and software systems.
2Measurement precision
If regular-dose CT scans are performed to reduce noise, then image quality is improved, but radiation exposure increases
Solution Approach 1:
The patent takes the harmful noise present in low-dose CT images and converts it into a useful component. By generating a synthetic sinogram from the low-dose image and processing it through filtered back-projection, the noise is extracted and separated into a noise component image. This noise component is then subtracted from the original image, transforming the harmful noise into a benefit by revealing the underlying clean image structure.
3Measurement precision
If legacy CT systems are upgraded to iterative reconstruction systems, then image quality is improved, but compatibility and ease of manufacture decrease
Solution Approach 1:
The patent develops a denoising method that can be applied universally to legacy CT systems without requiring system-specific modifications. The approach uses standard filtered back-projection technology that is already present in conventional CT systems, making it compatible with existing hardware. The method processes the raw projection data through synthetic sinogram generation and noise component extraction, achieving iterative-reconstruction-quality images on conventional systems.
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
This approach enables the production of high-quality denoised CT images from low-dose images, reducing noise effectively and minimizing radiation exposure, while being applicable to a wider range of CT systems.
Implementation Method 1
Computed tomography (CT) scan produces cross-sectional images of the body lying in a cylinder-shaped large machine equipped with an X-ray generator
Implementation Method 2
generating a noise component CT image by applying the filtered back-projection operation to the noise component synthetic sinogram
Data Source
AI summary
Provided is a denoising method including: generating a synthetic sinogram from an input original CT image; acquiring a noise component from the generated synthetic sinogram; generating a noise component CT image based on the noise component; and reducing noise in the original CT image based on the noise component CT image.