FA-GAN: fused attentive generative adversarial networks for MRI image super-resolution
File(s)
Published version
Author(s)
Type
Journal Article
Abstract
High-resolution magnetic resonance images can provide fine-grained anatomical
information, but acquiring such data requires a long scanning time. In this paper, a
framework called the Fused Attentive Generative Adversarial Networks(FA-GAN) is
proposed to generate the super- resolution MR image from low-resolution magnetic
resonance images, which can reduce the scanning time effectively but with high
resolution MR images. In the framework of the FA-GAN, the local fusion feature
block, consisting of different three-pass networks by using different convolution
kernels, is proposed to extract image features at different scales. And the global
feature fusion module, including the channel attention module, the self-attention
module, and the fusion operation，is designed to enhance the important features of the
MR image. Moreover, the spectral normalization process is introduced to make the
discriminator network stable. 40 sets of 3D magnetic resonance images (each set of
images contains 256 slices) are used to train the network, and 10 sets of images are
used to test the proposed method. The experimental results show that the PSNR and
SSIM values of the super-resolution magnetic resonance image generated by the
proposed FA-GAN method are higher than the state-of-the-art reconstruction
methods.
information, but acquiring such data requires a long scanning time. In this paper, a
framework called the Fused Attentive Generative Adversarial Networks(FA-GAN) is
proposed to generate the super- resolution MR image from low-resolution magnetic
resonance images, which can reduce the scanning time effectively but with high
resolution MR images. In the framework of the FA-GAN, the local fusion feature
block, consisting of different three-pass networks by using different convolution
kernels, is proposed to extract image features at different scales. And the global
feature fusion module, including the channel attention module, the self-attention
module, and the fusion operation，is designed to enhance the important features of the
MR image. Moreover, the spectral normalization process is introduced to make the
discriminator network stable. 40 sets of 3D magnetic resonance images (each set of
images contains 256 slices) are used to train the network, and 10 sets of images are
used to test the proposed method. The experimental results show that the PSNR and
SSIM values of the super-resolution magnetic resonance image generated by the
proposed FA-GAN method are higher than the state-of-the-art reconstruction
methods.
Date Issued
2021-09
Date Acceptance
2021-08-06
Citation
Computerized Medical Imaging and Graphics, 2021, 92, pp.1-11
ISSN
0895-6111
Publisher
Elsevier
Start Page
1
End Page
11
Journal / Book Title
Computerized Medical Imaging and Graphics
Volume
92
Copyright Statement
© 2021 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/)
License URL
Sponsor
British Heart Foundation
European Research Council Horizon 2020
Commission of the European Communities
Innovative Medicines Initiative
Medical Research Council (MRC)
Identifier
https://www.sciencedirect.com/science/article/pii/S089561112100118X?via%3Dihub
Grant Number
PG/16/78/32402
H2020-SC1-FA-DTS-2019-1 952172
101005122
101005122
MR/V023799/1
Subjects
Attention
Generative adversarial networks
MRI
Mechanism
Super-resolution
eess.IV
eess.IV
cs.CV
Nuclear Medicine & Medical Imaging
0903 Biomedical Engineering
1103 Clinical Sciences
Publication Status
Published
Date Publish Online
2021-08-10