Meeting Banner
Abstract #2115

Integrating Classical Image Filters with Physics-Driven Deep Learning for Sharper Image Reconstruction

Junno Yun1,2 and Mehmet Akçakaya1,2
1Electrical and Computer Engineering, University of Minnesota, Minneapolis, MN, United States, 2Center for Magnetic Resonance Research, University of Minnesota, Minneapolis, MN, United States

Synopsis

Keywords: Machine Learning/Artificial Intelligence, AI/ML Image Reconstruction

Motivation: To improve the sharpness of physics-driven deep learning (PD-DL) reconstruction by incorporating Laplacian filter blocks into existing unrolled networks.

Goal(s): This study aims to integrate traditional Laplacian filtering to refine PD-DL reconstruction, specifically targeting blurring artifacts.

Approach: Laplacian sharpening filters with a single tunable weight are incorporated to the output of regularization units in unrolled PD-DL networks. These networks are compared to conventional unrolled networks with matching architectures under same supervised learning conditions.

Results: The proposed approach improves image clarity and detail in reconstructed MR images, indicating that integrating a classic building block with deep learning may enhance overall performance.

Impact: The proposed approach incorporates simple Laplacian sharpening filters into unrolled networks, which is shown to enhance sharpness, with visual improvements. This hybrid methodology represents a promising direction, merging traditional techniques with DL for superior 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