Meeting Banner
Abstract #3825

Self-Supervised Denoising Model for 7T MR Angiography with Z-Directional Artery Focused DDPM

SoJin Yun1,2, Sung-Hye You3, Byungjun Kim3, Dong-Hyun Kim2, and Hyunseok Seo1
1Bionics Research Center, Korea Institute of Science and Technology, Seoul, Korea, Republic of, 2Department of Electrical and Electronic Engineering, Yonsei University, Seoul, Korea, Republic of, 3Department of Radiology, Anam Hospital, Korea University College of Medicine, Seoul, Korea, Republic of

Synopsis

Keywords: AI/ML Image Reconstruction, AI/ML Image Reconstruction

Motivation: 7T MRA imaging can offer improved SNR but still poses challenges in diagnosing small vessel abnormalities.

Goal(s): To develop a self-supervised deep learning approach that effectively reduces noise while preserving vascular details in 7T MRA images.

Approach: Utilizing independent noise characteristics between adjacent slices, the method creates high-SNR datasets through slice averaging and trains a conditional DDPM with surrounding slice information.

Results: The proposed method outperforms conventional approaches like BM3D and DnCNN, by effectively reducing noise while preserving fine vascular details and enhancing vessel visibility, even revealing previously undetectable structures.

Impact: This study advances 7T MRA imaging by providing a ground truth-free denoising method, potentially improving the diagnosis accuracy of cerebrovascular diseases and enhancing the clinical utility of 7T MRI in vascular imaging applications.

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