Meeting Banner
Abstract #3980

Correction of Motion-Affected MRI Images via Motion-Adaptive Diffusion Model

Quanhao Sun1, Yaqiong Chai1, Michael Khoo2, and Hosung Kim1
1Neuroimaging with Deep Learning Lab (NIDLL), Stevens Neuroimaging and Informatics Institute, University of Southern California, Los Angeles, CA, United States, 2Cardiorespiratory Sleep Laboratory (CRSL), Alfred E. Mann Department of Biomedical Engineering, University of Southern California, Los Angeles, CA, United States

Synopsis

Keywords: AI Diffusion Models, Brain

Motivation: Motion artifacts in MRI scans hinder diagnostic accuracy, especially in populations like infants and Parkinson’s patients. Existing correction methods struggle to restore image quality in such cases fully.

Goal(s): This study aims to develop a Motion-Adaptive Diffusion Model (MADM) to correct motion artifacts in MR, improving image quality.

Approach: MADM is based on a diffusion model. Gaussian noise was added in the forward process, and a U-Net progressively denoises the images in reverse process. The model was trained on the MR-ART dataset.

Results: MADM significantly outperformed traditional methods, reducing NMSE by 0.0226 and improving PSNR, SSIM, and CCC by 5.5558, 0.1160, and 0.0141.

Impact: This project significantly improves MRI diagnostic accuracy by effectively correcting motion artifacts. It provides a more efficient and reliable solution for both clinical and research applications by reducing the need for repeat scans and enhancing image quality.

How to access this content:

For one year after publication, abstracts and videos are only open to registrants of this annual meeting. Registrants should use their existing login information. Non-registrant access can be purchased via the ISMRM E-Library.

After one year, current ISMRM & ISMRT members get free access to both the abstracts and videos. Non-members and non-registrants must purchase access via the ISMRM E-Library.

After two years, the meeting proceedings (abstracts) are opened to the public and require no login information. Videos remain behind password for access by members, registrants and E-Library customers.

Click here for more information on becoming a member.

Keywords