Journal article

xQSM: quantitative susceptibility mapping with octave convolutional and noise-regularized neural networks

Yang Gao, Xuanyu Zhu, Bradford A Moffat, Rebecca Glarin, Alan H Wilman, G Bruce Pike, Stuart Crozier, Feng Liu, Hongfu Sun

NMR in Biomedicine | WILEY | Published : 2020

Abstract

Quantitative susceptibility mapping (QSM) provides a valuable MRI contrast mechanism that has demonstrated broad clinical applications. However, the image reconstruction of QSM is challenging due to its ill-posed dipole inversion process. In this study, a new deep learning method for QSM reconstruction, namely xQSM, was designed by introducing noise regularization and modified octave convolutional layers into a U-net backbone and trained with synthetic and in vivo datasets, respectively. The xQSM method was compared with two recent deep learning (QSMnet+ and DeepQSM) and two conventional dipole inversion (MEDI and iLSQR) methods, using both digital simulations and in vivo experiments. Recons..

View full abstract

University of Melbourne Researchers

Grants

Awarded by Natural Sciences and Engineering Research Council of Canada


Awarded by Canadian Institutes of Health Research


Awarded by Australian Research Council


Awarded by University of Queensland


Funding Acknowledgements

Natural Sciences and Engineering Research Council of Canada, Grant/Award Number: DG-03880; Canadian Institutes of Health Research, Grant/Award Number: FDN-143290; Australian Research Council, Grant/Award Number: DE210101297; University of Queensland, Grant/Award Number: UQECR2057605