Nothing Special   »   [go: up one dir, main page]
Skip to main content
Springer Nature Link
Log in

Ultra-Low-Dose Spectral CT Based on a Multi-level Wavelet Convolutional Neural Network

  • Published:
Journal of Digital Imaging Aims and scope Submit manuscript

Abstract

Spectral computed tomography (CT) based on a photon-counting detector (PCD) is a promising technique with the potential to improve lesion detection, tissue characterization, and material decomposition. PCD-based scanners have several technical issues including operation in the step-and-scan mode and long data acquisition time. One straightforward solution to these issues is to reduce the number of projection views. However, if the projection data are under-sampled or noisy, it would be challenging to produce a correct solution without precise prior information. Recently, deep-learning approaches have demonstrated impressive performance for under-sampled CT reconstruction. In this work, the authors present a multilevel wavelet convolutional neural network (MWCNN) to address the limitations of PCD-based scanners. Data properties of the proposed method in under-sampled spectral CT are analyzed with respect to the proposed deep-running-network-based image reconstruction using two measures: sampling density and data incoherence. This work presents the proposed method and four different methods to restore sparse sampling. We investigate and compare these methods through a simulation and real experiments. In addition, data properties are quantitatively analyzed and compared for the effect of sparse sampling on the image quality. Our results indicate that both sampling density and data incoherence affect the image quality in the studied methods. Among the different methods, the proposed MWCNN shows promising results. Our method shows the highest performance in terms of various evaluation parameters such as the structural similarity, root mean square error, and resolution. Based on the results of imaging and quantitative evaluation, this study confirms that the proposed deep-running network structure shows excellent image reconstruction in sparse-view PCD-based CT. These results demonstrate the feasibility of sparse-view PCD-based CT using the MWCNN. The advantage of sparse view CT is that it can significantly reduce the radiation dose and obtain images with several energy bands by fusing PCDs. These results indicate that the MWCNN possesses great potential for sparse-view PCD-based CT.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Subscribe and save

Springer+ Basic
$34.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or eBook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12
Fig. 13

Similar content being viewed by others

References

  1. Wang G, Yu H, De Man B. An outlook on x-ray CT research and development. Med Phys. 2008;35(3):1051‐1064. https://doi.org/10.1118/1.2836950

    Article  PubMed  Google Scholar 

  2. Alvarez RE, Macovski A. Energy-selective reconstructions in X-ray computerized tomography. Phys Med Biol. 1976;21(5):733‐744. https://doi.org/10.1088/0031-9155/21/5/002

    Article  CAS  PubMed  Google Scholar 

  3. De Man B, Nuyts J, Dupont P, Marchal G, Suetens P. An iterative maximum-likelihood polychromatic algorithm for CT. IEEE Trans Med Imaging. 2001;20(10):999‐1008. https://doi.org/10.1109/42.959297

    Article  PubMed  Google Scholar 

  4. Graser A, Johnson TR, Chandarana H, Macari M. Dual energy CT: preliminary observations and potential clinical applications in the abdomen. Eur Radiol. 2009;19(1):13‐23. https://doi.org/10.1007/s00330-008-1122-7

    Article  PubMed  Google Scholar 

  5. Yu L, Leng S, McCollough CH. Dual-energy CT-based monochromatic imaging. AJR Am J Roentgenol. 2012;199(5 Suppl):S9‐S15. https://doi.org/10.2214/AJR.12.9121

    Article  PubMed  Google Scholar 

  6. Lv P, Lin XZ, Chen K, Gao J. Spectral CT in patients with small HCC: investigation of image quality and diagnostic accuracy. Eur Radiol. 2012;22(10):2117‐2124. https://doi.org/10.1007/s00330-012-2485-3

    Article  PubMed  Google Scholar 

  7. Zhao LQ, He W, Li JY, Chen JH, Wang KY, Tan L. Improving image quality in portal venography with spectral CT imaging. Eur J Radiol. 2012;81(8):1677‐1681. https://doi.org/10.1016/j.ejrad.2011.02.063

    Article  PubMed  Google Scholar 

  8. Heismann BJ, Schmidt BT, Flohr T. Spectral computed tomography. Bellingham, WA: SPIE; 2012.

    Book  Google Scholar 

  9. Yang Q, Cong W, Xi Y, Wang G. Spectral X-Ray CT Image Reconstruction with a Combination of Energy-Integrating and Photon-Counting Detectors. PLoS One. 2016;11(5):e0155374. Published 2016 May 12. https://doi.org/10.1371/journal.pone.0155374

  10. Shikhaliev PM, Fritz SG. Photon counting spectral CT versus conventional CT: comparative evaluation for breast imaging application. Phys Med Biol. 2011;56(7):1905‐1930. https://doi.org/10.1088/0031-9155/56/7/001

    Article  CAS  PubMed  Google Scholar 

  11. Leng S, Yu L, Wang J, Fletcher JG, Mistretta CA, McCollough CH. Noise reduction in spectral CT: reducing dose and breaking the trade-off between image noise and energy bin selection. Med Phys. 2011;38(9):4946‐4957. https://doi.org/10.1118/1.3609097

    Article  PubMed  Google Scholar 

  12. Yu Z, Leng S, Li Z, McCollough CH. Spectral prior image constrained compressed sensing (spectral PICCS) for photon-counting computed tomography. Phys Med Biol. 2016;61(18):6707‐6732. https://doi.org/10.1088/0031-9155/61/18/6707

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Zeng D, Huang J, Zhang H, et al. Spectral CT Image Restoration via an Average Image-Induced Nonlocal Means Filter. IEEE Trans Biomed Eng. 2016;63(5):1044‐1057. https://doi.org/10.1109/TBME.2015.2476371

  14. Zeng D, Gao Y, Huang J, et al. Penalized weighted least-squares approach for multienergy computed tomography image reconstruction via structure tensor total variation regularization. Comput Med Imaging Graph. 2016;53:19‐29. https://doi.org/10.1016/j.compmedimag.2016.07.002

    Article  PubMed  Google Scholar 

  15. Roessl E, Proksa R. K-edge imaging in x-ray computed tomography using multi-bin photon counting detectors. Phys Med Biol. 2007;52(15):4679‐4696. https://doi.org/10.1088/0031-9155/52/15/020

    Article  CAS  PubMed  Google Scholar 

  16. He P, Wei B, Cong W, Wang G. Optimization of K-edge imaging with spectral CT. Med Phys. 2012;39(11):6572‐6579. https://doi.org/10.1118/1.4754587

    Article  PubMed  PubMed Central  Google Scholar 

  17. Schirra CO, Roessl E, Koehler T, et al. Statistical reconstruction of material decomposed data in spectral CT. IEEE Trans Med Imaging. 2013;32(7):1249‐1257. https://doi.org/10.1109/TMI.2013.2250991

    Article  PubMed  Google Scholar 

  18. Elbakri IA, Fessler JA. Statistical image reconstruction for polyenergetic X-ray computed tomography. IEEE Trans Med Imaging. 2002;21(2):89‐99. https://doi.org/10.1109/42.993128

    Article  PubMed  Google Scholar 

  19. Sawatzky A, Xu Q, Schirra CO, Anastasio MA. Proximal ADMM for multi-channel image reconstruction in spectral X-ray CT. IEEE Trans Med Imaging. 2014;33(8):1657‐1668. https://doi.org/10.1109/TMI.2014.2321098

    Article  PubMed  Google Scholar 

  20. Kim K, Ye JC, Worstell W, et al. Sparse-view spectral CT reconstruction using spectral patch-based low-rank penalty. IEEE Trans Med Imaging. 2015;34(3):748‐760. https://doi.org/10.1109/TMI.2014.2380993

    Article  PubMed  Google Scholar 

  21. Zhang Y, Mou X, Wang G, Yu H. Tensor-Based Dictionary Learning for Spectral CT Reconstruction. IEEE Trans Med Imaging. 2017;36(1):142‐154. https://doi.org/10.1109/TMI.2016.2600249

    Article  PubMed  Google Scholar 

  22. Wang M, Zhang Y, Liu R, Guo S, Yu H. An adaptive reconstruction algorithm for spectral CT regularized by a reference image. Phys Med Biol. 2016;61(24):8699‐8719. https://doi.org/10.1088/1361-6560/61/24/8699

    Article  PubMed  PubMed Central  Google Scholar 

  23. Li S, Zeng D, Peng J, et al. An Efficient Iterative Cerebral Perfusion CT Reconstruction via Low-Rank Tensor Decomposition With Spatial-Temporal Total Variation Regularization. IEEE Trans Med Imaging. 2019;38(2):360‐370. https://doi.org/10.1109/TMI.2018.2865198

    Article  PubMed  Google Scholar 

  24. Chen GH, Tang J, Leng S. Prior image constrained compressed sensing (PICCS): a method to accurately reconstruct dynamic CT images from highly undersampled projection data sets. Med Phys. 2008;35(2):660‐663. https://doi.org/10.1118/1.2836423

    Article  PubMed  Google Scholar 

  25. Sidky EY, Pan X. Image reconstruction in circular cone-beam computed tomography by constrained, total-variation minimization. Phys Med Biol. 2008;53(17):4777‐4807. https://doi.org/10.1088/0031-9155/53/17/021

    Article  PubMed  PubMed Central  Google Scholar 

  26. Singh S, Kalra MK, Gilman MD, et al. Adaptive statistical iterative reconstruction technique for radiation dose reduction in chest CT: a pilot study. Radiology. 2011;259(2):565‐573. https://doi.org/10.1148/radiol.11101450

    Article  PubMed  Google Scholar 

  27. Andersen AH, Kak AC. Simultaneous algebraic reconstruction technique (SART): a superior implementation of the art algorithm. Ultrason Imaging. 1984;6(1):81‐94. https://doi.org/10.1177/016173468400600107

    Article  CAS  PubMed  Google Scholar 

  28. Donoho DL. Compressed sensing. IEEE Transactions on information theory. 2006;52(4):1289-1306.

    Article  Google Scholar 

  29. Krizhevsky A, Sutskever I, Hinton GE. Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems. 2012.

  30. Ronneberger O, Fischer P, Brox T. U-net: Convolutional networks for biomedical image segmentation. International Conference on Medical image computing and computer-assisted intervention. Springer, Cham, 2015.

  31. Zhang K, Zuo W, Chen Y, Meng D, Zhang L. Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising. IEEE Trans Image Proces. 2017;26(7):3142‐3155. https://doi.org/10.1109/TIP.2017.2662206

    Article  PubMed  Google Scholar 

  32. Shi, W., Caballero, J., Huszár, F., Totz, J., Aitken, A. P., Bishop, R., ... & Wang, Z. Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network. Proceedings of the IEEE conference on computer vision and pattern recognition. 2016. p. 1874–1883.

  33. Kang E, Min J, Ye JC. A deep convolutional neural network using directional wavelets for low-dose X-ray CT reconstruction. Med Phys. 2017;44(10):e360‐e375. https://doi.org/10.1002/mp.12344

    Article  CAS  PubMed  Google Scholar 

  34. Kang, E., & Ye, J. C. "Wavelet domain residual network (WavResNet) for low-dose X-ray CT reconstruction." arXiv preprint arXiv 1703.01383; 2017.

  35. Kang E, Chang W, Yoo J, Ye JC. Deep Convolutional Framelet Denosing for Low-Dose CT via Wavelet Residual Network. IEEE Trans Med Imaging. 2018;37(6):1358‐1369. https://doi.org/10.1109/TMI.2018.2823756

    Article  PubMed  Google Scholar 

  36. Mao X, Shen C, Yang YB. Image restoration using very deep convolutional encoder-decoder networks with symmetric skip connections. Advances in neural information processing systems. 2016.

  37. Xie J, Xu L, Chen E. Image denoising and inpainting with deep neural networks. Advances in neural information processing systems. 2012.

  38. Kyong Hwan Jin, McCann MT, Froustey E, Unser M. Deep Convolutional Neural Network for Inverse Problems in Imaging. IEEE Trans Image Process. 2017;26(9):4509‐4522. https://doi.org/10.1109/TIP.2017.2713099

  39. Han YS, Yoo J, Ye JC. Deep residual learning for compressed sensing CT reconstruction via persistent homology analysis. 2016. arXiv preprint arXiv 1611.06391.

  40. Daubechies I. The wavelet transform, time-frequency localization and signal analysis. IEEE transactions on information theory. 1990;36(5):961-1005.

    Article  Google Scholar 

  41. Daubechies I. Ten lectures on wavelets. Vol. 61. Siam; 1992.

  42. Taguchi K, Stierstorfer K, Polster C, Lee O, Kappler S. Spatio-energetic cross-talk in photon counting detectors: Numerical detector model (PcTK) and workflow for CT image quality assessment. Med Phys. 2018;45(5):1985‐1998. https://doi.org/10.1002/mp.12863

    Article  PubMed  Google Scholar 

  43. ICRU Report 44. Tissue Substitutes in Radiation Dosimetry and Measurement. Bethesda, MD: International Commission on Radiation Units and Measurements (ICRU); 1989.

  44. Yu F, and Koltun V. Multi-scale context aggregation by dilated convolutions. 2015. arXiv preprint arXiv 1511.07122.

  45. Zeyde R, Elad M, Protter M. On single image scale-up using sparse-representations. International conference on curves and surfaces. Springer, Berlin, Heidelberg; 2010.

  46. Wang P, Chen P, Yuan Y, et al. Understanding convolution for semantic segmentation. 2018 IEEE winter conference on applications of computer vision (WACV). IEEE; 2018.

  47. Ye JC, Han Y, Cha E. Deep convolutional framelets: A general deep learning framework for inverse problems. SIAM Journal on Imaging Sciences. 2018;11(2):991-1048.

    Article  Google Scholar 

  48. Lehmann LA, Alvarez RE, Macovski A, et al. Generalized image combinations in dual KVP digital radiography. Med Phys. 1981;8(5):659‐667. https://doi.org/10.1118/1.595025.

    Article  CAS  PubMed  Google Scholar 

  49. Kelcz F, Joseph PM, Hilal SK. Noise considerations in dual energy CT scanning. Med Phys. 1979;6(5):418‐425. https://doi.org/10.1118/1.594520.

    Article  CAS  PubMed  Google Scholar 

  50. Cho HM, Barber WC, Ding H, Iwanczyk JS, Molloi S. Characteristic performance evaluation of a photon counting Si strip detector for low dose spectral breast CT imaging. Med Phys. 2014;41(9):091903. https://doi.org/10.1118/1.4892174.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  51. Lee MJ, Lee DH, Kim DH, et al. Development of a non-linear dual-energy technique in chest radiography. Radiation Physics and Chemistry. 2020;108811.

  52. Xue H, Zhang L, Cheng Z, Xing Y, Xiao Y. An improved TV minimization algorithm for incomplete data problem in computer tomography. In: IEEE Nuclear Science Symposuim & Medical Imaging Conference. IEEE; 2010.

  53. Gao Y, Bian Z, Huang J, et al. Low-dose X-ray computed tomography image reconstruction with a combined low-mAs and sparse-view protocol. Opt Express. 2014;22(12):15190-15210.

    Article  Google Scholar 

  54. Lee M, Kim H, Kim HJ. Sparse-view CT reconstruction based on multi-level wavelet convolution neural network. Phys Med. 2020;80:352-362. https://doi.org/10.1016/j.ejmp.2020.11.021.

    Article  PubMed  Google Scholar 

  55. Taguchi K, Iwanczyk JS. Vision 20/20: Single photon counting x-ray detectors in medical imaging. Med Phys. 2013;40(10):100901. https://doi.org/10.1118/1.4820371.

    Article  PubMed  PubMed Central  Google Scholar 

  56. Leng S, Bruesewitz M, Tao S, et al. Photon-counting Detector CT: System Design and Clinical Applications of an Emerging Technology. Radiographics. 2019;39(3):729-743. https://doi.org/10.1148/rg.2019180115.

    Article  PubMed  Google Scholar 

  57. Willemink MJ, Persson M, Pourmorteza A, Pelc NJ, Fleischmann D. Photon-counting CT: Technical Principles and Clinical Prospects. Radiology. 2018;289(2):293-312. https://doi.org/10.1148/radiol.2018172656

    Article  PubMed  Google Scholar 

  58. Mahmoudi G, Fouladi MR, Ay MR, Rahmim A, Ghadiri H. Sparse-view statistical image reconstruction with improved total variation regularization for X-ray micro-CT imaging. J Instru. 2019;14(08):P08023-P08023.

    Article  Google Scholar 

  59. Szczykutowicz, TP, Chen, GH. Dual energy CT using slow kVp switching acquisition and prior image constrained compressed sensing. Phys Med Biol 2010; 55.21: 6411.

    Article  Google Scholar 

  60. Abbas, S et al. Effects of sparse sampling schemes on image quality in low‐dose CT. Medical Phys 2013; 40.11: 111915.

    Article  Google Scholar 

  61. Lee, T et al. Moving beam-blocker-based low-dose cone-beam CT. IEEE Trans Nuclear Sci 2016; 63.5: 2540–2549.

    Article  Google Scholar 

  62. Pan, X, Sidky, EM, Vannier, M. Why do commercial CT scanners still employ traditional, filtered back-projection for image reconstruction?. Inverse Problems 2009; 25.12: 123009.

    Article  Google Scholar 

  63. Bian, J et al. Evaluation of sparse-view reconstruction from flat-panel-detector cone-beam CT. Phys Med Biol 2010; 55.22: 6575.

    Article  Google Scholar 

Download references

Funding

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (NTIS Number: 2020R1F1A1075741). This work was supported by the Korea Medical Device Development Fund grant funded by the Korea government (the Ministry of Science and ICT, the Ministry of Trade, Industry and Energy, the Ministry of Health & Welfare, the Ministry of Food and Drug Safety) (Project Number: 1711137897, KMDF_PR_20200901_0014). This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (NRF-2021R1I1A1A01059875).

Author information

Authors and Affiliations

  1. Department of Radiation Convergence Engineering, Yonsei University, 1 Yonseidae-gil, Wonju, 26493, Republic of Korea

    Minjae Lee & Hee-Joung Kim

  2. Department of Radiological Science, Yonsei University, 1 Yonseidae-gil, Wonju, 26493, Republic of Korea

    Hyemi Kim & Hee-Joung Kim

  3. Korea Research Institute of Standards and Science, Daejoen, Republic of Korea

    Hyo-Min Cho

Authors
  1. Minjae Lee
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Hyemi Kim
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Hyo-Min Cho
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Hee-Joung Kim
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Hee-Joung Kim.

Ethics declarations

Conflict of Interest

The authors declare no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendix. Image-based VMI technique

Appendix. Image-based VMI technique

VMIs can be created with reconstructed low- and high-energy images. The effective mass attenuation coefficients of the two basis materials at low and high-energy scans are [48, 49]

$${\left(\frac{\mu }{\rho }\right)}_{i}^{j}j=L,H,i=\mathrm{1,2}$$
(7)

where L and H are low- and high-energy bins, respectively, and 1 and 2 are two basic materials. By solving the two linear equations, the mass densities of the two basis materials are obtained as follows:

$$\begin{array}{c}{\rho }_{1}=\frac{{\mu }^{L}\cdot {\left(\frac{\mu }{\rho }\right)}_{2}^{H}-{\mu }^{H}\cdot {\left(\frac{\mu }{\rho }\right)}_{2}^{L}}{{\left(\frac{\mu }{\rho }\right)}_{1}^{L}\cdot {\left(\frac{\mu }{\rho }\right)}_{2}^{H}-{\left(\frac{\mu }{\rho }\right)}_{1}^{H}\cdot {\left(\frac{\mu }{\rho }\right)}_{2}^{L}}\\ {\rho }_{2}=\frac{-{\mu }^{L}\cdot {\left(\frac{\mu }{\rho }\right)}_{1}^{H}+{\mu }^{H}\cdot {\left(\frac{\mu }{\rho }\right)}_{1}^{L}}{{\left(\frac{\mu }{\rho }\right)}_{1}^{L}\cdot {\left(\frac{\mu }{\rho }\right)}_{2}^{H}-{\left(\frac{\mu }{\rho }\right)}_{1}^{H}\cdot {\left(\frac{\mu }{\rho }\right)}_{2}^{L}}\end{array}$$
(8)

The monochromatic image at energy E is given by

$$\mu \left(E\right)={\left(\frac{\mu }{\rho }\right)}_{1}\left(E\right){\rho }_{1}+{\left(\frac{\mu }{\rho }\right)}_{2}\left(E\right){\rho }_{2}$$
(9)

By rewriting the LACs in Eq. (9) in terms of the CT number and by assuming that one of the basis materials is the soft tissue, one can show that the monochromatic image at energy E can be expressed as the weighted average of the images at low- and high-energy bins. Image-based VMI can be obtained as follows:

$$CT\left(E\right)=w\left(E\right)\cdot {CT}^{L}\cdot \left[1-w\left(E\right)\right]\cdot {CT}^{H}$$
(10)

where the weighting factor is given by

$$w\left(E\right)=\frac{{\mu }_{1}\left(E\right)\cdot {\mu }_{2}^{H}-{\mu }_{2}\left(E\right)\cdot {\mu }_{1}^{H}}{{\mu }_{1}^{L}\cdot {\mu }_{2}^{H}-{\mu }_{1}^{H}\cdot {\mu }_{2}^{L}}\cdot \frac{{\mu }_{2}^{H}}{{\mu }_{2}\left(E\right)}$$
(11)

The obtained monochromatic image is simply a linear combination of the two energy-bin reconstruction images.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Lee, M., Kim, H., Cho, HM. et al. Ultra-Low-Dose Spectral CT Based on a Multi-level Wavelet Convolutional Neural Network. J Digit Imaging 34, 1359–1375 (2021). https://doi.org/10.1007/s10278-021-00467-w

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10278-021-00467-w

Keywords

Search

Navigation