Skip to main navigation Skip to search Skip to main content

Learned Primal Dual Splitting for Self-Supervised Noise-Adaptive MRI Reconstruction

  • Nikola Janjušević
  • , Amirhossein Khalilian-Gourtani
  • , Yao Wang
  • , Li Feng

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

3 Scopus citations

Abstract

Magnetic resonance imaging (MRI) reconstruction has largely been dominated by deep neural networks (DNN); however, many state-of-the-art architectures use black-box structures, which hinder interpretability and improvement. Here, we propose an interpretable DNN architecture for self-supervised MRI reconstruction and denoising by directly parameterizing and learning the classical primal-dual splitting, dubbed LPDSNet. This splitting algorithm allows us to decouple the observation model from the signal prior. Experimentally, we show other interpretable architectures without this decoupling property exhibit failure in the self-supervised learning regime. We report state-of-the-art self-supervised joint MRI reconstruction and denoising performance and novel noise-level generalization capabilities, where in contrast black-box networks fail to generalize.

Original languageEnglish
Title of host publicationISBI 2025 - 2025 IEEE 22nd International Symposium on Biomedical Imaging, Proceedings
PublisherIEEE Computer Society
ISBN (Electronic)9798331520526
DOIs
StatePublished - 2025
Externally publishedYes
Event22nd IEEE International Symposium on Biomedical Imaging, ISBI 2025 - Houston, United States
Duration: 14 Apr 202517 Apr 2025

Publication series

NameProceedings - International Symposium on Biomedical Imaging
ISSN (Print)1945-7928
ISSN (Electronic)1945-8452

Conference

Conference22nd IEEE International Symposium on Biomedical Imaging, ISBI 2025
Country/TerritoryUnited States
CityHouston
Period14/04/2517/04/25

Keywords

  • Deep-learning
  • interpretable neural network
  • joint MRI reconstruction and denoising
  • unrolled network

Fingerprint

Dive into the research topics of 'Learned Primal Dual Splitting for Self-Supervised Noise-Adaptive MRI Reconstruction'. Together they form a unique fingerprint.

Cite this