Abstract
Single image super-resolution reconstruction (SISR) can effectively and economically improve the spatial resolution of magnetic resonance (MR) images, and it helps more accurate early clinical diagnosis and subsequent analysis. To increase the imaging speed and reduce the patient’s pain and motion artifacts, many studies have moved from only considering the quality of the reconstructed image to proposing some lightweight models. However, the model’s lightweight will limit its performance, and high-resolution MR images are often reconstructed with a single target (LOSS). In this work, we propose a lightweight generative adversarial network to alleviate this problem. The network mainly contains generators and discriminators. The generator uses a global cascade module to extract image features, and multi-scale up sampling of high-frequency and low-frequency features of different depths. As the cascaded modules lead to similar features, a consistent spatial attention module is used to weigh them and share the up-sampling module to reduce network parameters. The discriminator judges the authenticity of the input MR image, and it constructs two losses with the pre-trained VGG network to assist the generator training and provide diversified standards for the generation of MR images. In addition, we use knowledge transfer to train the network to explore the toplimit of network performance. Qualitative and quantitative experiments on the FASTMRI dataset show that the MR images generated by the designed multiple targets (loss) have better visual effects in detail. The proposed network has advantages in running time and parameter memory and achieved the highest precision results compared with state-of-the-art methods.
Similar content being viewed by others
References
Agustsson E, Timofte R (2017) NTIRE 2017 challenge on single image super-resolution: dataset and study. In: 2017 IEEE conference on computer vision and pattern recognition workshops (CVPRW), pp 1122–1131
Ahn N, Kang B, Sohn K-a (2018) Fast, accurate, and, lightweight super-resolution with cascading residual network. ArXiv abs/180308664 n. pag
Bosse S, Maniry D, Müller K, Wiegand T, Samek W (2018) Deep neural networks for no-reference and full-reference image quality assessment. IEEE Trans Image Process 27:206–219
Carmi E, Liu S, Alon N, Fiat A, Fiat D (2006) Resolution enhancement in MRI. Magn Reson Imaging 24(2):133–154
Chen Y, Xie Y, Zhou Z, Shi F, Christodoulou AG, Li D (2018) Brain MRI super resolution using 3D deep densely connected neural networks. In: 2018 IEEE 15th international symposium on biomedical imaging (ISBI 2018), pp 739–742
Chen Y, Shi F, Christodoulou A, Zhou Z, Xie Y, Li D (2018) Efficient and accurate MRI super-resolution using a generative adversarial network and 3D multi-level densely connected network. ArXiv abs/1803.01417: n. pag
Chu X, Zhang B, Ma H, Xu R, Li J, Li Q (2021) Fast, accurate and lightweight super-resolution with neural architecture search. 2020 25th International Conference on Pattern Recognition (ICPR):59–64
Dai T, Cai J, Zhang Y, Xia S, Zhang L (2019) Second-order attention network for single image super-resolution. In: 2019 IEEE/CVF conference on computer vision and pattern recognition (CVPR), pp 11057–11066
Dong C, Loy CC, He K, Tang X (2016) Image super-resolution using deep convolutional networks. IEEE Trans Pattern Anal Mach Intell 38:295–307
Dong C, Loy CC, Tang X (2016) Accelerating the super-resolution convolutional neural network. ECCV https://doi.org/10.48550/arXiv.1608.00367
Farsiu S, Robinson M, Elad M, Milanfar P (2004) Fast and robust multiframe super resolution. IEEE Trans Image Process 13:1327–1344
Fitzgibbon, AW, Pollefeys M, Van Gool L, Zisserman A (2006) European conference on computer vision (ECCV). ECCV 2006
Giannakidis A, Oktay O, Keegan J, Spadotto V, Firmin DN (2017) Super-resolution reconstruction of late gadolinium enhancement cardiovascular magnetic resonance images using a residual convolutional neural network. In: 25th scientific meeting of the International Society for Magnetic Resonance in medicine (ISMRM 2017)
Goodfellow I, Pouget-Abadie J, Mirza M, Xu B, Warde-Farley D, Ozair S, Courville A, Bengio Y (2014) Generative adversarial networks. Adv Neural Inf Proces Syst 3:139–144. https://doi.org/10.1145/3422622
Greenspan H, Oz G, Kiryati N, Peled S (2002) MRI inter-slice reconstruction using super-resolution. Magn Reson Imaging 20:437–446
He K, Zhang X, Ren S, Sun J (2016) Deep residual learning for image recognition. In: 2016 IEEE conference on computer vision and pattern recognition (CVPR), pp 770–778
Hui Z, Gao X, Yang Y, Wang X (2019) Lightweight image super-resolution with information multi-distillation network. In: Proceedings of the 27th ACM international conference on multimedia, n. pag
Ioffe S, Szegedy C (2015) Batch normalization: accelerating deep network training by reducing internal covariate shift. ICML
Jolicoeur-Martineau A (2019) The relativistic discriminator: a key element missing from standard GAN. ArXiv abs/180700734 n. pag
Kim J, Lee JK, Lee KM (2016) Accurate image super-resolution using very deep convolutional networks. In: 2016 IEEE conference on computer vision and pattern recognition (CVPR), pp 1646–1654
Kim S, Hong J-H, Kang I, Kwak N (2019) Semantic sentence matching with densely-connected recurrent and co-attentive information. ArXiv abs/180511360 n. pag
Lai W-S, Huang J-B, Ahuja N, Yang M-H (2017) Deep Laplacian pyramid networks for fast and accurate super-resolution. In: 2017 IEEE conference on computer vision and pattern recognition (CVPR), pp 5835–5843
Ledig C, Theis L, Huszár F, Caballero J, Aitken AP, Tejani A, Totz J, Wang Z, Shi W (2017) Photo-realistic single image super-resolution using a generative adversarial network. In: 2017 IEEE conference on computer vision and pattern recognition (CVPR), pp 105–114
Li X, Orchard M (2001) New edge-directed interpolation. IEEE Trans Image Process 10(10):1521–1527
Li Y, Iwamoto Y, Lin L, Xu R, Tong R, Chen Y-W (2021) VolumeNet: a lightweight parallel network for super-resolution of MR and CT volumetric data. IEEE Trans Image Process 30:4840–4854
Lim B, Son S, Kim H, Nah S, Lee KM (2017) Enhanced deep residual networks for single image super-resolution. 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), pp 1132–1140. https://doi.org/10.48550/arXiv.1707.02921
Liu J, Tang J, Wu G (2020) Residual feature distillation network for lightweight image super-resolution. ECCV Workshops
Lyu Q, Shan H, Wang G (2020) MRI super-resolution with ensemble learning and complementary priors. IEEE Trans Comput Imaging 6:615–624
Manjón J, Coupé P, Buades A, Collins D, Robles M (2010) MRI Superresolution using self-similarity and image priors. Int J Biomed Imaging, n. pag
McDonagh SG, Hou B, Alansary A, Oktay O, Kamnitsas K, Rutherford MA, Hajnal JV, Kainz B (2017) Context-sensitive super-resolution for fast fetal magnetic resonance imaging. CMMI/RAMBO/SWITCH@MICCAI
Mittal A, Soundararajan R, Bovik A (2013) Making a “completely blind” image quality analyzer. IEEE Signal Process Lett 20:209–212
Pan SJ, Yang Q (2010) A survey on transfer learning. IEEE Trans Knowl Data Eng 22:1345–1359
Park SC, Park M, Kang M (2003) Super-resolution image reconstruction: a technical overview. IEEE Signal Process Mag 20:21–36
Ramzi Z, Ciuciu P, Starck J (2020) Benchmarking MRI reconstruction neural networks on large public datasets. Appl Sci 10:1816
Shi W, Caballero J, Huszár F, Totz J, Aitken AP, Bishop R, Rueckert D, Wang Z (2016) Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network. In: 2016 IEEE conference on computer vision and pattern recognition (CVPR), pp 1874–1883
Shi J, Li Z, Ying S, Wang C, Liu Q, Zhang Q, Yan P (2019) MR image super-resolution via wide residual networks with fixed skip connection. IEEE J Biomed Health Inform 23:1129–1140
Shilling RZ, Robbie TQ, Bailloeul T, Mewes K, Mersereau R, Brummer M (2009) A super-resolution framework for 3-D high-resolution and high-contrast imaging using 2-D multislice MRI. IEEE Trans Med Imaging 28:633–644
Song D, Wang Y, Chen H, Xu C, Xu C, Tao D (2021) AdderSR: towards energy efficient image super-resolution. 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR):15643–15652
Tai Y, Yang J, Liu X (2017) Image super-resolution via deep recursive residual network. 2017 IEEE conference on computer vision and pattern recognition (CVPR):2790–2798
Tai Y, Yang J, Liu X, Xu C (2017) MemNet: a persistent memory network for image restoration. 2017 IEEE international conference on computer vision (ICCV):4549–4557
Tanno R, Worrall DE, Ghosh A, Kaden E, Sotiropoulos SN, Criminisi A, Alexander DC (2017) Bayesian image quality transfer with CNNs: exploring uncertainty in dMRI super-resolution. MICCAI
Tian C, Zhuge R, Wu Z, Xu Y, Zuo W, Chen C, Lin C (2020) Lightweight image super-resolution with enhanced CNN. Knowl Based Syst 205:106235
Timofte R, De Smet V, Van Gool L (2014) A+: adjusted anchored neighborhood regression for fast super-resolution. ACCV
Tsai R, Huang T. (1984) Multiframe image restoration and registration. Adv Comput Vis Image Process 317–339
Wang Z, Bovik A, Sheikh H, Simoncelli EP (2004) Image quality assessment: from error visibility to structural similarity. IEEE Trans Image Process 13:600–612
Wang Z, Liu D, Yang J, Han W, Huang T (2015) Deep networks for image super-resolution with sparse prior. In: 2015 IEEE international conference on computer vision (ICCV), pp 370–378
Wang X, Yu K, Wu S, Gu J, Liu Y, Dong C, Loy CC, Qiao Y, Tang X (2018) ESRGAN: enhanced super-resolution generative adversarial networks. ArXiv abs/180900219: n. pag
Woo S, Park J, Lee J-Y, Kweon I-S (2018) CBAM: convolutional block attention module. ECCV
Xia Y, Ravikumar N, Greenwood J, Neubauer S, Petersen S, Frangi AF (2021) Super-resolution of cardiac MR cine imaging using conditional GANs and unsupervised transfer learning. Med Image Anal 71:102037
Xue X, Wang Y, Li J, Jiao Z, Ren Z, Gao X (2020) Progressive sub-band residual-learning network for MR image super resolution. IEEE J Biomed Health Inform 24:377–386
Yang J, Wright J, Huang T, Ma Y (2010) Image super-resolution via sparse representation. IEEE Trans Image Process 19:2861–2873
Yuan Y, Liu S, Zhang J, Zhang Y, Dong C, Lin L (2018) Unsupervised image super-resolution using cycle-in-cycle generative adversarial networks. In: 2018 IEEE/CVF conference on computer vision and pattern recognition workshops (CVPRW) (2018): 814–81409
Yue L, Shen H, Li J, Yuan Q, Zhang H, Zhang L (2016) Image super-resolution: The techniques, applications, and future. Signal Process 128:389–408
Zbontar J, Knoll F, Sriram A, Muckley M, Bruno M, Defazio A, Parente M, Geras K, Katsnelson J, Chandarana H, Zhang Z, Drozdzal M, Romero A, Rabbat MG, Vincent P, Pinkerton J, Wang D, Yakubova N, Owens E, Zitnick CL, Recht M, Sodickson D, Lui Y (2018) fastMRI: an open dataset and benchmarks for accelerated MRI. ArXiv abs/181108839 (2018): n. pag
Zeng K, Zheng H, Cai C, Yang Y, Zhang K, Chen Z (2018) Simultaneous single- and multi-contrast super-resolution for brain MRI images based on a convolutional neural network. Comput Biol Med 99:133–141
Zhang C, Ma Y. (2012) Ensemble machine learning: methods and applications. https://link.springer.com/book/10.1007%2F978-1-4419-9326-7
Zhang Y, Tian Y, Kong Y, Zhong B, Fu YR (2018) Residual dense network for image super-resolution. In: 2018 IEEE/CVF conference on computer vision and pattern recognition, pp 2472–2481
Zhang Y, Li K, Li K, Wang L, Zhong B, Fu YR (2018) Image super-resolution using very deep Residual Channel attention networks. ECCV
Zhang W, Liu Y, Dong C, Qiao Y (2019) RankSRGAN: generative adversarial networks with ranker for image super-resolution. 2019 IEEE/CVF International Conference on Computer Vision (ICCV), pp 3096–3105
Zhao C, Carass A, Dewey BE, Prince JL (2018) Self super-resolution for magnetic resonance images using deep networks. In: 2018 IEEE 15th international symposium on biomedical imaging (ISBI 2018), pp 365–368
Zhao X, Zhang Y, Zhang T, Zou X (2019) Channel splitting network for single MR image super-resolution. IEEE Trans Image Process 28:5649–5662
Zhao X, Zhang H, Zhou Y, Bian W, Zhang T, Zou X (2019) Gibbs-ringing artifact suppression with knowledge transfer from natural images to MR images. Multimed Tools Appl:1–23
Zhao X, Hu X, Liao Y, He T, Zhang T, Zou X, Tian J (2020) Accurate MR image super-resolution via lightweight lateral inhibition network. Comput Vis Image Underst 201:103075
Zhu X, Cheng D, Zhang Z, Lin S, Dai J (2019) An empirical study of spatial attention mechanisms in deep networks. In: 2019 IEEE/CVF international conference on computer vision (ICCV), pp 6687–6696
Acknowledgements
We wish to express our gratitude to the anonymous reviewers for their insightful comments.
Funding
The work is supported by Beijing Outstanding Talents Training Fund Youth Top Individual Project,
Premium Funding Project for Academic Human Resources Development in Beijing Union University under.
grant BPHR2020EZ01.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
The authors declare that they have no competing interests.
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Li, H., Xuan, Z., Zhou, J. et al. Fast and accurate super-resolution of MR images based on lightweight generative adversarial network. Multimed Tools Appl 82, 2465–2487 (2023). https://doi.org/10.1007/s11042-022-13326-9
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11042-022-13326-9