Related Experiment Video
Updated: Jun 16, 2025
Standardized Data Acquisition for Neuromelanin-Sensitive Magnetic Resonance Imaging of the Substantia Nigra
Published on: September 8, 2021
Explicit Abnormality Extraction for Unsupervised Motion Artifact Reduction in Magnetic Resonance Imaging
Abstract:
Motion artifacts compromise the quality of magnetic resonance imaging (MRI) and pose challenges to achieving diagnostic outcomes and image-guided therapies. In recent years, supervised deep learning approaches have emerged as successful solutions for motion artifact reduction (MAR). One disadvantage of these methods is their dependency on acquiring paired sets of motion artifact-corrupted (MA-corrupted) and motion artifact-free (MA-free) MR images for training purposes. Obtaining such image pairs is difficult and therefore limits the application of supervised training. In this paper, we propose a novel UNsupervised Abnormality Extraction Network (UNAEN) to alleviate this problem. Our network is capable of working with unpaired MA-corrupted and MA-free images. It converts the MA-corrupted images to MA-reduced images by extracting abnormalities from the MA-corrupted images using a proposed artifact extractor, which intercepts the residual artifact maps from the MA-corrupted MR images explicitly, and a reconstructor to restore the original input from the MA-reduced images. The performance of UNAEN was assessed by experimenting with various publicly available MRI datasets and comparing them with state-of-the-art methods. The quantitative evaluation demonstrates the superiority of UNAEN over alternative MAR methods and visually exhibits fewer residual artifacts. Our results substantiate the potential of UNAEN as a promising solution applicable in real-world clinical environments, with the capability to enhance diagnostic accuracy and facilitate image-guided therapies.
Insights
This study introduces a novel unsupervised deep learning network (UNAEN) for motion artifact reduction in MRI. UNAEN effectively reduces artifacts using unpaired images, improving diagnostic accuracy and image-guided therapies.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Deep Learning
Background:
- Motion artifacts significantly degrade Magnetic Resonance Imaging (MRI) quality, hindering diagnosis and image-guided therapies.
- Supervised deep learning methods for motion artifact reduction (MAR) require paired corrupted and artifact-free images, which are difficult to obtain.
- This limitation restricts the practical application of supervised MAR techniques in clinical settings.
Purpose of the Study:
- To propose a novel unsupervised deep learning network, UNAEN, for motion artifact reduction in MRI.
- To enable MAR using unpaired corrupted and artifact-free MR images, overcoming the limitations of supervised methods.
- To enhance the quality of MRI scans for improved diagnostic accuracy and image-guided therapies.
Main Methods:
- Developed a UNsupervised Abnormality Extraction Network (UNAEN) that operates on unpaired MRI datasets.
- Implemented an artifact extractor to identify and isolate artifact maps from corrupted MR images.
- Utilized a reconstructor to restore image quality from the artifact-reduced images.
Main Results:
- UNAEN demonstrated superior performance compared to state-of-the-art MAR methods on various public MRI datasets.
- Quantitative evaluations confirmed the effectiveness of UNAEN in reducing motion artifacts.
- Visual assessments showed significantly fewer residual artifacts in images processed by UNAEN.
Conclusions:
- UNAEN offers a promising unsupervised solution for motion artifact reduction in MRI.
- The network's ability to work with unpaired data makes it suitable for real-world clinical applications.
- UNAEN has the potential to enhance diagnostic accuracy and facilitate advanced image-guided therapies.
