CT Imaging Aliasing Artifact Reduction via Deep Neural Network
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computed tomography (CT) images often suffer from aliasing artifacts due to sub-sampling, which can degrade image quality and make it difficult to preserve details, texture, and sharpness.
Innovation Solution
A method involving the use of a trained deep neural network to process three-dimensional image volumes acquired during CT scans. The network corrects aliasing artifacts by generating a corrected three-dimensional image volume, while preserving image details and quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If sub-sampling is used in CT imaging, then scan time is reduced and productivity is improved, but aliasing artifacts are introduced and image quality deteriorates
Solution Approach 1:
The patent uses a neural network to create a synthetic representation of the complete sampled image from the subsampled input. The neural network learns to copy and reconstruct the missing high-frequency information that would be present in a fully sampled image, effectively generating a virtual complete image without requiring actual complete sampling during acquisition.
Solution Approach 2:
The patent transforms the problem from the sampling domain to the neural network parameter domain. Instead of changing the physical sampling parameters (which would require longer scan times), the system changes the computational parameters by training neural network weights to reconstruct the full image from subsampled data, decoupling image quality from acquisition time.
2Loss of time
If sub-sampling is used in CT imaging, then scan time is reduced, but aliasing artifacts are introduced that degrade image quality
Solution Approach 1:
The patent converts the harmful aliasing artifacts into a training signal for the neural network. The neural network is trained on pairs of subsampled images (with artifacts) and their corresponding ground truth versions, learning to identify and remove the artifacts. The harmful effect of subsampling is thus transformed into a beneficial training opportunity.
Solution Approach 2:
The neural network creates a synthetic copy of the complete image by learning the mapping from subsampled to full-resolution images. This copied representation contains the correct high-frequency information without the artifacts present in the original subsampled input, effectively removing the harmful factor through computational synthesis.
3Device complexity
If conventional image processing is used, then computational resources are minimal, but aliasing artifacts cannot be effectively removed
Solution Approach 1:
The patent replaces traditional mechanical image processing methods (such as filter-based approaches) with a neural network-based system. The neural network uses learned transformations and feature representations to remove artifacts, substituting conventional signal processing mechanisms with data-driven computational models that achieve superior artifact removal.
Solution Approach 2:
The system changes the computational approach from parameter-based filtering to neural network parameter optimization. Instead of applying fixed filters with adjustable parameters, the system trains neural network weights to optimize artifact removal across diverse imaging conditions, achieving better performance through learned parameters rather than conventional processing.
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
The method effectively reduces aliasing artifacts in CT images, enhancing image quality by preserving details, texture, and sharpness, and providing high-resolution images without artifacts caused by sub-sampling.
Implementation Method 1
inputting the three-dimensional image volume to a trained deep neural network to generate a corrected three-dimensional image volume with a reduction in aliasing artifacts
Data Source
AI summary
Various methods and systems are provided for computed tomography imaging. In one embodiment, a method includes acquiring, with an x-ray detector and an x-ray source coupled to a gantry, a three-dimensional image volume of a subject while the subject moves through a bore of the gantry and the gantry rotates the x-ray detector and the x-ray source around the subject, inputting the three-dimensional image volume to a trained deep neural network to generate a corrected three-dimensional image volume with a reduction in aliasing artifacts present in the three-dimensional image volume, and outputting the corrected three-dimensional image volume. In this way, aliasing artifacts caused by sub-sampling may be removed from computed tomography images while preserving details, texture, and sharpness in the computed tomography images.