Dual-domain convolutional neural networks for improving structural information in 3 T MRI

Magn Reson Imaging. 2019 Dec;64:90-100. doi: 10.1016/j.mri.2019.05.023. Epub 2019 Jun 5.

Abstract

We propose a novel dual-domain convolutional neural network framework to improve structural information of routine 3 T images. We introduce a parameter-efficient butterfly network that involves two complementary domains: a spatial domain and a frequency domain. The butterfly network allows the interaction of these two domains in learning the complex mapping from 3 T to 7 T images. We verified the efficacy of the dual-domain strategy and butterfly network using 3 T and 7 T image pairs. Experimental results demonstrate that the proposed framework generates synthetic 7 T-like images and achieves performance superior to state-of-the-art methods.

Keywords: Convolutional neural network; Deep learning; Image super-resolution; Image synthesis; Magnetic resonance imaging.

Publication types

  • Research Support, N.I.H., Extramural
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Algorithms
  • Brain / diagnostic imaging*
  • Epilepsy / diagnostic imaging*
  • Humans
  • Image Interpretation, Computer-Assisted / methods*
  • Magnetic Resonance Imaging / methods*
  • Neural Networks, Computer*