Related Experiment Video
Updated: May 3, 2026
06:06
Optogenetic Functional MRI
Published on: April 19, 2016
14.8K
GAN-Based Motion Artifact Correction of 3D MR Volumes Using an Image-to-Image Translation Algorithm
Vishnu Vardhan Reddy Kanamata Reddy1,2, Chandan Ganesh Bangalore Yogananda3, Nghi C D Truong3
1University of Texas at Dallas, Center for Imaging and Surgical Innovation, Richardson, TX.
Summary
This study presents a 3D deep learning framework using generative adversarial networks (GANs) to remove motion artifacts from brain MRI scans. The method effectively restores image quality, offering a promising solution for clinical applications.
Area of Science:
- Medical Imaging
- Artificial Intelligence in Medicine
- Neuroscience
Background:
- Motion artifacts significantly degrade brain MRI quality, causing blurring and ghosting.
- These artifacts stem from patient movement during scans, impacting diagnostic accuracy.
- Existing motion correction methods may be insufficient for complex artifacts.
Purpose of the Study:
- To introduce a novel 3D deep learning framework for restoring motion-corrupted brain MRI volumes.
- To enhance image quality by mitigating artifacts caused by respiratory and head movements.
- To validate the framework's efficacy on diverse motion-affected MR datasets.
Main Methods:
- Development of a 3D deep learning framework integrating a densely connected 3D U-net architecture.
- Augmentation of training with generative adversarial network (GAN)-informed techniques.
- Implementation of a novel volumetric reconstruction loss function specifically for 3D GANs.
Main Results:
- The proposed framework successfully restored motion-corrupted brain MR volumes.
- Generated high-quality MR volumes exhibited volumetric signatures comparable to motion-free scans.
- The system demonstrated significant potential in rectifying motion artifacts.
Conclusions:
- The 3D deep learning system offers a powerful tool for motion artifact correction in brain MRI.
- This technology holds promise for improving the quality and reliability of clinical neuroimaging.
- Further applications in advanced clinical settings are anticipated.
