Comparison of Neural Network Architectures for Physics-Driven Deep Learning MRI Reconstruction

Burhaneddin Yaman, Seyed Amir Hossein Hosseini, Steen Moeller, Mehmet Akcakaya

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Scopus citations

Abstract

Machine learning techniques have recently received interest as a means of improving MRI reconstruction. Conventionally, ill-conditioned reconstruction problems are solved using iterative optimization algorithms that alternate between applying data consistency and a proximal operator based on a regularizer. This iterative procedure can also be unrolled for a finite number of iterations to generate a feed-forward model. In physics-driven machine learning approaches, the known forward encoding model is used for enforcing data consistency in an unrolled iterative regularized least squares reconstruction. A neural network, which may or may not share weights across different unrolled iterations, is used as the regularizer prior. In this study, we aim to compare several neural network architectures, namely U-Net, ResNet and DenseNet for such physicsdriven reconstruction. The performance of these architectures are evaluated on the publicly available fastMRI knee database. Comparisons are made for uniform and random undersampling masks. The results indicate that a DenseNet regularization unit performs as well as the other strategies for both uniform and random undersampling patterns, even though it has considerably fewer trainable parameters.

Original languageEnglish (US)
Title of host publication2019 IEEE 10th Annual Information Technology, Electronics and Mobile Communication Conference, IEMCON 2019
EditorsSatyajit Chakrabarti, Himadri Nath Saha
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages155-159
Number of pages5
ISBN (Electronic)9781728125305
DOIs
StatePublished - Oct 2019
Event10th IEEE Annual Information Technology, Electronics and Mobile Communication Conference, IEMCON 2019 - Vancouver, Canada
Duration: Oct 17 2019Oct 19 2019

Publication series

Name2019 IEEE 10th Annual Information Technology, Electronics and Mobile Communication Conference, IEMCON 2019

Conference

Conference10th IEEE Annual Information Technology, Electronics and Mobile Communication Conference, IEMCON 2019
CountryCanada
CityVancouver
Period10/17/1910/19/19

Keywords

  • Data consistency
  • Deep learning
  • MRI reconstruction
  • Parallel imaging
  • Recurrent neural networks
  • Unrolled network

Fingerprint Dive into the research topics of 'Comparison of Neural Network Architectures for Physics-Driven Deep Learning MRI Reconstruction'. Together they form a unique fingerprint.

Cite this