Related Experiment Video
Updated: Jan 25, 2026
A Virtual Machine Platform for Non-Computer Professionals for Using Deep Learning to Classify Biological Sequences of Metagenomic Data
Published on: September 25, 2021
Retrospective correction of motion-affected MR images using deep learning frameworks
Thomas Küstner1,2,3, Karim Armanious1,2, Jiahuan Yang1
1Department for Signal Processing and System Theory, University of Stuttgart, Stuttgart, Germany.
Purpose:
Motion is 1 extrinsic source for imaging artifacts in MRI that can strongly deteriorate image quality and, thus, impair diagnostic accuracy. In addition to involuntary physiological motion such as respiration and cardiac motion, intended and accidental patient movements can occur. Any impairment by motion artifacts can reduce the reliability and precision of the diagnosis and a motion-free reacquisition can become time- and cost-intensive. Numerous motion correction strategies have been proposed to reduce or prevent motion artifacts. These methods have in common that they need to be applied during the actual measurement procedure with a-priori knowledge about the expected motion type and appearance. For retrospective motion correction and without the existence of any a-priori knowledge, this problem is still challenging.
Methods:
We propose the use of deep learning frameworks to perform retrospective motion correction in a reference-free setting by learning from pairs of motion-free and motion-affected images. For this image-to-image translation problem, we propose and compare a variational auto encoder and generative adversarial network. Feasibility and influences of motion type and optimal architecture are investigated by blinded subjective image quality assessment and by quantitative image similarity metrics.
Results:
We observed that generative adversarial network-based motion correction is feasible producing near-realistic motion-free images as confirmed by blinded subjective image quality assessment. Generative adversarial network-based motion correction accordingly resulted in images with high evaluation metrics (normalized root mean squared error <0.08, structural similarity index >0.8, normalized mutual information >0.9).
Conclusion:
Deep learning-based retrospective restoration of motion artifacts is feasible resulting in near-realistic motion-free images. However, the image translation task can alter or hide anatomical features and, therefore, the clinical applicability of this technique has to be evaluated in future studies.
Insights
Deep learning effectively corrects MRI motion artifacts retrospectively without prior knowledge. Generative adversarial networks produce near-realistic, motion-free images, enhancing diagnostic accuracy.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Machine Learning
Background:
- Motion artifacts in Magnetic Resonance Imaging (MRI) significantly degrade image quality and diagnostic accuracy.
- Existing motion correction methods often require prospective application with prior motion knowledge, posing challenges for retrospective correction.
- Retrospective motion correction without prior knowledge remains a significant challenge in MRI.
Purpose of the Study:
- To investigate the feasibility of deep learning frameworks for retrospective, reference-free motion correction in MRI.
- To compare the performance of variational autoencoders and generative adversarial networks for image-to-image translation in motion artifact correction.
- To evaluate the influence of motion type and optimal network architecture on correction efficacy.
Main Methods:
- Proposed deep learning models, specifically variational autoencoders and generative adversarial networks, for retrospective motion correction.
- Trained models on pairs of motion-free and motion-affected MRI images for image-to-image translation.
- Conducted blinded subjective image quality assessments and quantitative image similarity analyses to evaluate performance.
Main Results:
- Generative adversarial network (GAN)-based motion correction demonstrated feasibility, producing near-realistic, motion-free images.
- Subjective assessments confirmed the quality of GAN-corrected images.
- Quantitative metrics showed high performance: normalized root mean squared error <0.08, structural similarity index >0.8, and normalized mutual information >0.9.
Conclusions:
- Deep learning enables feasible retrospective restoration of motion artifacts in MRI, yielding near-realistic, motion-free images.
- While promising, the image translation process may alter or obscure anatomical details.
- Further studies are required to evaluate the clinical applicability of this deep learning-based motion correction technique.
Related Concept Videos
01:15Distance Corrections
01:20Power Factor Correction
01:22Equation of Motion: General Plane motion
Moreover, the body's center of mass experiences a rotational effect as a result of these couple moments. This rotation can be articulated as the...
01:14NMR Spectrometers: Resolution and Error Correction
01:14Avoidance Learning and Learned Helplessness
Avoidance learning occurs when an organism learns that a specific behavior can prevent an unpleasant outcome. For example, a student who receives a bad grade may start studying harder to avoid future poor grades. This behavior persists even when the negative outcome is no longer present. Avoidance learning is powerful because it maintains behavior in the absence of the...
01:24Absolute Motion Analysis- General Plane Motion
As the drone's propellers rotate, an upward force is generated that counteracts the force of gravity, enabling the drone to lift off from the ground. This initial movement of the drone is along a straight path, representing a form of translational motion. In this phase, every point on the...