Multi-Channel MR Image Artifact Reduction Using Neural Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for reducing artifacts in magnetic resonance (MR) images, particularly channel-dependent artifacts such as streak, motion, and flow artifacts, are inaccurate and require manual intervention, compromising image quality and diagnostic accuracy.
Innovation Solution
A neural network model is trained to estimate and reduce channel-dependent artifacts by analyzing MR images from multiple channels, using computer-generated training images to recognize and mitigate these artifacts, and combining the images based on artifact indicators to produce high-quality MR images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If low-pass filtering is used to reduce artifacts in MR images, then artifact reduction is achieved, but image signals are skewed and diagnostic accuracy is compromised
Solution Approach 1:
The patent segments the artifact reduction problem by channel, processing each RF coil channel independently through the neural network. This allows channel-specific artifact estimation and removal while preserving channel-specific image signals, thereby reducing artifacts without compromising diagnostic accuracy.
Solution Approach 2:
The neural network acts as an intermediary between the raw MR images and the final artifact-reduced images. It estimates artifacts as a separate intermediate component that can be removed without directly filtering the image signals, thus preserving diagnostic information while eliminating artifacts.
2Loss of time
If traditional artifact reduction methods are used, then processing time is reduced, but image quality is compromised
Solution Approach 1:
The patent replaces traditional mechanical filtering methods (low-pass filtering) with a neural network-based computational approach. This substitution enables more precise artifact removal that preserves image quality while maintaining efficient processing through parallel computation across multiple channels.
Solution Approach 2:
The neural network learns optimal artifact estimation parameters through training on labeled data, allowing it to adapt to different imaging conditions and artifact types. This parameter optimization enables high-quality artifact reduction without requiring excessive processing time or complex post-processing.
3Device complexity
If channel-dependent artifacts are not addressed specifically, then processing complexity is reduced, but image quality suffers
Solution Approach 1:
The patent applies segmentation by processing each RF coil channel separately through the neural network, allowing channel-dependent artifact estimation and removal. This segmented approach handles the complexity of channel-specific artifacts while maintaining overall system efficiency through parallel processing.
Solution Approach 2:
The neural network applies local quality optimization by estimating and removing artifacts specific to each channel's characteristics. This localized processing ensures that each channel's unique artifact patterns are addressed appropriately, improving overall image quality without requiring overly complex global processing.
Data Source
AI summary
A computer-implemented method of reducing artifacts in multi-channel magnetic resonance (MR) images is provided. The method includes receiving a plurality of sets of MR images acquired by a radio-frequency (RF) coil assembly having a plurality of channels. Each set of MR images includes a plurality of slices of MR images acquired by one of the plurality of channels. The method also includes estimating a plurality of sets of artifacts in the plurality of sets of MR images by inputting the plurality of sets of MR images into a neural network model. Each set of artifacts corresponds to the one of the plurality of channels. The method further includes reducing artifacts in the plurality of sets of MR images based on estimated artifacts, deriving MR images of reduced artifacts by combining the MR images of reduced artifacts, and outputting the MR images of reduced artifacts.