Method and apparatus for motion artifact correction using artificial neural networks

By using artificial neural networks to train and simulate motion artifacts, the challenge of motion artifacts in magnetic resonance imaging is solved, and efficient artifact removal is achieved in dynamic scanning objects, improving image quality and analysis accuracy.

CN114926366BActive Publication Date: 2025-09-23SHANGHAI UNITED IMAGING INTELLIGENCE CO LTD
2 Cites 0 Cited by

Patent Information

Application Number
CN202210642037.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2021-07-16
Filing Date
2022-06-07
Publication Date
2025-09-23
Estimated Expiration
2042-06-07

AI Technical Summary

Technical Problem

During magnetic resonance imaging, especially during dynamic scanning of objects, the removal of motion artifacts remains a challenging task. The existing technology lacks an effective deep learning-based image processing system, and insufficient training data makes it difficult to use multiple interrelated images for motion artifact correction.

Method used

An artificial neural network (ANN) was used to learn the parameters associated with motion artifact removal through a training process. A training dataset of multiple paired MR images containing different motion artifacts was used to simulate motion artifacts and minimize the differences. The network was trained to remove artifacts from actual MR images. No-reference learning and correlation information between multiple images were used for training.

Benefits of technology

It achieves effective removal of motion artifacts without the need for controlled motion image data, improves the quality of MR images, reduces the impact of motion artifacts, and enhances the accuracy and efficiency of image analysis.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

Neural network-based systems, methods, and apparatus can be used to remove motion artifacts from magnetic resonance (MR) images. Such neural network-based systems can be trained to perform motion artifact removal tasks without a reference (e.g., without using pairs of motion-contaminated and motion-free MR images). Various training techniques are described herein, including techniques that present pairs of MR images with varying levels of motion contamination to a neural network and force the neural network to learn to correct for the motion contamination by transforming the first image of the contaminated pair into the second image of the contaminated pair. Other neural network training techniques are also described that aim to reduce reliance on difficult-to-obtain training data.
Need to check novelty before this filing date? Find Prior Art