Meeting Banner
Abstract #3815

Spatially Adaptive DWI Denoising using Plug-and-Play Diffusion Models

Mahmoud Mostapha1, Radu Miron2, Nirmal Janardhanan1, Mariappan S. Nadar1, Omar Darwish3, Till Huelnhagen3, Tobias Würfl3, David Grodzki3, and Rainer Schneider3
1Siemens Healthineers, Princeton, NJ, United States, 2Siemens Industry Software România, Brasov, Romania, 3Siemens Healthineers AG, Erlangen, Germany

Synopsis

Keywords: AI Diffusion Models, AI/ML Image Reconstruction

Motivation: Diffusion-weighted imaging (DWI) is essential in clinical settings but often has a low and spatially variable signal-to-noise ratio (SNR), particularly in under-sampled high-b acquisitions of small organs.

Goal(s): To accelerate DWI scans by reducing the number of image repetitions needed and overall acquisition time while maintaining image quality.

Approach: Integrate a plug-and-play (PnP) method with a diffusion sampling framework that uses calculated noise map information to adaptively denoise accelerated DWI data affected by spatially varying noise.

Results: Our method successfully preserves the content of images and prevents over-smoothing, even at high denoising settings. This efficiency enables a significant reduction in DWI scan times.

Impact: We present a denoising method that accelerates DWI scans through a PnP diffusion model that utilizes noise maps for guidance. This approach improves scanning efficiency while preserving image quality, showcasing promise for future DWI clinical 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