Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Competitive performance of a modularized deep neural network compared to commercial algorithms for low-dose CT image reconstruction


Commercial iterative reconstruction techniques help to reduce the radiation dose of computed tomography (CT), but altered image appearance and artefacts can limit their adoptability and potential use. Deep learning has been investigated for low-dose CT (LDCT). Here, we design a modularized neural network for LDCT and compare it with commercial iterative reconstruction methods from three leading CT vendors. Although popular networks are trained for an end-to-end mapping, our network performs an end-to-process mapping so that intermediate denoised images are obtained with associated noise reduction directions towards a final denoised image. The learned workflow allows radiologists in the loop to optimize the denoising depth in a task-specific fashion. Our network was trained with the Mayo LDCT Dataset and tested on separate chest and abdominal CT exams from Massachusetts General Hospital. The best deep learning reconstructions were systematically compared to the best iterative reconstructions in a double-blinded reader study. This study confirms that our deep learning approach performs either favourably or comparably in terms of noise suppression and structural fidelity, and is much faster than commercial iterative reconstruction algorithms.

This is a preview of subscription content, access via your institution

Relevant articles

Open Access articles citing this article.

Access options

Rent or buy this article

Get just this article for as long as you need it


Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Proposed MAP-NN for LDCT denoising and progressively denoised images on LDCT and NDCT images with mapping depth D.
Fig. 2: Comparison between the best DL reconstruction and the best IR reconstruction for abdomen and chest regions across three major vendors (A, B and C) and three readers (R1, R2 and R3).
Fig. 3: Individual metric comparison between the best DL and best IR reconstructions keyed to body regions (abdomen and chest) and CT vendors (A, B and C).
Fig. 4: Best DL and best IR reconstructions, as well as NDCT FBP for abdomen and chest regions from three vendors.

Data availability

The training and validation sets were obtained from the 2016 NIH-AAPM-Mayo Clinic Low-Dose CT Grand Challenge, and are publicly available ( The testing data from MGH for this study are protected for patient privacy and will be available for data sharing upon request and after going through an external Institutional Review Board procedure at MGH.

Code availability

The source code and the trained models for this study are publicly available on Github (


  1. Wang, G. A perspective on deep imaging. IEEE Access 4, 8914–8924 (2016).

    Article  Google Scholar 

  2. Wang, G., Ye, J. C., Mueller, K. & Fessler, J. A. Image reconstruction is a new frontier of machine learning. IEEE Trans. Med. Imaging 37, 1289–1296 (2018).

    Article  Google Scholar 

  3. Zhu, B., Liu, J. Z., Cauley, S. F., Rosen, B. R. & Rosen, M. S. Image reconstruction by domain-transform manifold learning. Nature 555, 487–492 (2018).

    Article  Google Scholar 

  4. Brenner, D. J. & Hall, E. J. Computed tomography—an increasing source of radiation exposure. New Engl. J. Med. 357, 2277–2284 (2007).

    Article  Google Scholar 

  5. de González, A. B. et al. Projected cancer risks from computed tomographic scans performed in the United States in 2007. Arch. Intern. Med. 169, 2071–2077 (2009).

    Article  Google Scholar 

  6. Smith-Bindman, R. et al. Radiation dose associated with common computed tomography examinations and the associated lifetime attributable risk of cancer. Arch. Intern. Med. 169, 2078–2086 (2009).

    Article  Google Scholar 

  7. National Lung Screening Trial Research Team Reduced lung-cancer mortality with low-dose computed tomographic screening. New Engl. J. Med. 2011, 395–409 (2011).

    Article  Google Scholar 

  8. Wang, J., Lu, H., Li, T. & Liang, Z. Sinogram noise reduction for low-dose CT by statistics-based nonlinear filters. Proc. SPIE 5747, 2059 (2005).

    Google Scholar 

  9. Manduca, A. et al. Projection space denoising with bilateral filtering and CT noise modeling for dose reduction in CT. Med. Phys. 36, 4911–4919 (2009).

    Article  Google Scholar 

  10. Wang, J., Li, T., Lu, H. & Liang, Z. Penalized weighted least-squares approach to sinogram noise reduction and image reconstruction for low-dose X-ray computed tomography. IEEE Trans. Med. Imaging 25, 1272–1283 (2006).

    Article  Google Scholar 

  11. Geyer, L. L. et al. State of the art: iterative CT reconstruction techniques. Radiology 276, 339–357 (2015).

    Article  Google Scholar 

  12. Willemink, M. J. et al. Iterative reconstruction techniques for computed tomography. Part 2: Initial results in dose reduction and image quality. Eur. Radiol. 23, 1632–1642 (2013).

    Article  Google Scholar 

  13. Zheng, X., Ravishankar, S., Long, Y. & Fessler, J. A. PWLS-ULTRA: an efficient clustering and learning-based approach for low-dose 3D CT image reconstruction. IEEE Trans. Med. Imaging 37, 1498–1510 (2018).

    Article  Google Scholar 

  14. Chen, H. et al. Low-dose CT via convolutional neural network. Biomed. Opt. Express 8, 679–694 (2017).

    Article  Google Scholar 

  15. Chen, H. et al. Low-dose CT with a residual encoder–decoder convolutional neural network. IEEE Trans. Med. Imaging 36, 2524–2535 (2017).

    Article  Google Scholar 

  16. Wolterink, J. M., Leiner, T., Viergever, M. A. & Išgum, I. Generative adversarial networks for noise reduction in low-dose CT. IEEE Trans. Med. Imaging 36, 2536–2545 (2017).

    Article  Google Scholar 

  17. Yang, Q. et al. Low-dose CT image denoising using a generative adversarial network with Wasserstein distance and perceptual loss. IEEE Trans. Med. Imaging 37, 1348–1357 (2018).

    Article  Google Scholar 

  18. Kang, E., Chang, W., Yoo, J. & Ye, J. C. Deep convolutional framelet denosing for low-dose CT via wavelet residual network. IEEE Trans. Med. Imaging 37, 1358–1369 (2018).

    Article  Google Scholar 

  19. Shan, H. et al. 3D convolutional encoder–decoder network for low-dose CT via transfer learning from a 2D trained network. IEEE Trans. Med. Imaging 37, 1522–1534 (2018).

    Article  Google Scholar 

  20. Choi, K., Kim, S. W. & Lim, J. S. Real-time image reconstruction for low-dose CT using deep convolutional generative adversarial networks (GANs). In Medical Imaging 2018: Physics of Medical Imaging Vol. 10573, 1057332 (International Society for Optics and Photonics, 2018).

  21. Kim, B., Han, M., Shim, H. & Baek, J. Performance comparison of convolutional neural network based denoising in low dose CT images for various loss functions. In Medical Imaging 2019: Physics of Medical Imaging Vol. 10948, 1094849 (International Society for Optics and Photonics, 2019).

  22. Liu, Y. & Zhang, Y. Low-dose CT restoration via stacked sparse denoising autoencoders. Neurocomputing 284, 80–89 (2018).

    Article  Google Scholar 

  23. You, C. et al. Structurally-sensitive multi-scale deep neural network for low-dose CT denoising. IEEE Access 6, 41839–41855 (2018).

    Article  Google Scholar 

  24. Hu, Z. et al. Artifact correction in low-dose dental CT imaging using Wasserstein generative adversarial networks. Med. Phys. 46, 1686–1696 (2019).

    Article  Google Scholar 

  25. Gholizadeh-Ansari, M., Alirezaie, J. & Babyn, P. Deep learning for low-dose CT denoising using perceptual loss and edge detection layer. Preprint at (2019).

  26. Yi, X. & Babyn, P. Sharpness-aware low-dose CT denoising using conditional generative adversarial network. J. Digit. Imaging 31, 655–669 (2018).

    Article  Google Scholar 

  27. Ronneberger, O., Fischer, P. & Brox, T. U-Net: convolutional networks for biomedical image segmentation. In Proceedings of Medical Image Computing and Computer Assisted Intervention (MICCAI) 234–241 (Springer, 2015).

  28. Pascanu, R., Mikolov, T. & Bengio, Y. On the difficulty of training recurrent neural networks. In Proceedings of the International Conference on Machine Learning 1310–1318 (JMLR, 2013).

  29. Tai, Y., Yang, J. & Liu, X. Image super-resolution via deep recursive residual network. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2790–2798 (IEEE, 2017).

  30. Gulrajani, I., Ahmed, F., Arjovsky, M., Dumoulin, V. & Courville, A. C. Improved training of Wasserstein GANs. In Advances in Neural Information Processing Systems (NIPS), 5769–5779 (Curran Associates, 2017).

  31. Goodfellow, I. et al. Generative adversarial nets. In Advances in Neural Information Processing Systems (NIPS) 2672–2680 (Curran Associates, 2014).

  32. Arjovsky, M., Chintala, S. & Bottou, L. Wasserstein generative adversarial networks. In Proceedings of the International Conference on Machine Learning (ICML) 214–223 (JMLR, 2017).

  33. Kingma, D. & Ba, J. Adam: a method for stochastic optimization. Preprint at (2014).

  34. Abadi, M. et al. Tensorflow: large-scale machine learning on heterogeneous distributed systems. Preprint at (2016).

  35. Conover, W. J. & Conover, W. J. Practical Nonparametric Statistics Ch. 3 (Wiley, 1980).

Download references


The authors thank M. Vannier and B. De Man for helpful discussions, and NVIDIA Corporation for the donation of GPUs used for this research. H.S. and G.W. were partially supported by the National Institutes of Health/National Institute of Biomedical Imaging and Bioengineering (U01EB017140).

Author information

Authors and Affiliations



G.W. initiated the project. G.W., H.S. and M.K.K. designed the experiments. H.S. and G.W. performed machine learning research. A.P. and F.H. collected CT data and conducted the reader studies. M.K.K., C.N. and R.D.K. were the readers. H.S., G.W., U.K. and M.K.K. analysed the data. H.S. and G.W. wrote the paper. M.K.K. and U.K. edited the paper.

Corresponding authors

Correspondence to Mannudeep K. Kalra or Ge Wang.

Ethics declarations

Competing interests

G.W. and H.S. have received unrelated industrial research grants from General Electric and Hologic Inc. M.K.K. has received unrelated industrial research grants from Siemens Healthineers and Riverain Inc.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Supplementary Notes 1–3, Supplementary Figs. 1–4, Supplementary Tables 1 and 2

Reporting Summary

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Shan, H., Padole, A., Homayounieh, F. et al. Competitive performance of a modularized deep neural network compared to commercial algorithms for low-dose CT image reconstruction. Nat Mach Intell 1, 269–276 (2019).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:

This article is cited by


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing