Journal of radiological science and technology (대한방사선기술학회지:방사선기술과학)

Korean Society of Radiological Science (대한방사선과학회)
권호 표지
DOI QR코드

DOI QR Code

The Evaluation of Denoising PET Image Using Self Supervised Noise2Void Learning Training: A Phantom Study

자기 지도 학습훈련 기반의 Noise2Void 네트워크를 이용한 PET 영상의 잡음 제거 평가: 팬텀 실험

  • Yoon, Seokhwan (Department of Nuclear Medicine, Seoul National University Hospital) ;
  • Park, Chanrok (Department of Radiological Science, Jeonju University)
  • 윤석환 (서울대학교병원 핵의학과) ;
  • 박찬록 (전주대학교 방사선학과)
  • Received : 2021.10.07
  • Accepted : 2021.11.24
  • Published : 2021.12.31
Download PDF
⟨ Previous Next ⟩

Abstract

Positron emission tomography (PET) images is affected by acquisition time, short acquisition times results in low gamma counts leading to degradation of image quality by statistical noise. Noise2Void(N2V) is self supervised denoising model that is convolutional neural network (CNN) based deep learning. The purpose of this study is to evaluate denoising performance of N2V for PET image with a short acquisition time. The phantom was scanned as a list mode for 10 min using Biograph mCT40 of PET/CT (Siemens Healthcare, Erlangen, Germany). We compared PET images using NEMA image-quality phantom for standard acquisition time (10 min), short acquisition time (2min) and simulated PET image (S2 min). To evaluate performance of N2V, the peak signal to noise ratio (PSNR), normalized root mean square error (NRMSE), structural similarity index (SSIM) and radio-activity recovery coefficient (RC) were used. The PSNR, NRMSE and SSIM for 2 min and S2 min PET images compared to 10min PET image were 30.983, 33.936, 9.954, 7.609 and 0.916, 0.934 respectively. The RC for spheres with S2 min PET image also met European Association of Nuclear Medicine Research Ltd. (EARL) FDG PET accreditation program. We confirmed generated S2 min PET image from N2V deep learning showed improvement results compared to 2 min PET image and The PET images on visual analysis were also comparable between 10 min and S2 min PET images. In conclusion, noisy PET image by means of short acquisition time using N2V denoising network model can be improved image quality without underestimation of radioactivity.

Keywords

References

  1. LeCun Y, Bottou L, Bengio Y, Haffner P. Gradientbased learning applied to document recognition. Proceeding of the IEEE. 1998;86(11):2278-324. https://doi.org/10.1109/5.726791
  2. Buades A, Coll B, Morel JM. A Non-local algorithm for image denoising. Proceedings Conference on Computer Vision and Pattern Recognition. 2005:60-5.
  3. Jeong EH, Oh JY, Lee JY, Park HH. Deep Learning Application of Gamma Camera Quality Control in Nuclear Medicine. Journal of Radiological Science and Technology. 2021;43(6):461-7. https://doi.org/10.17946/JRST.2020.43.6.461
  4. Chen J, Wang Y, Wu Y, Cai C. An ensemble of convolution neural networks for image classification based on LSTM. Proceedings of International Conference on Green Informatics(ICGI). 2017:217-22.
  5. Tajbakhsh N, Shin JY, Gurudu SR, Hursr RT, Kendall CB, Gotway MB, et al. Convolution neural networks for medical image analysis: Full training or fine tuning? IEEE Trans Med Imaging. 2016;35:1299-312. https://doi.org/10.1109/TMI.2016.2535302
  6. Anwar SM, Majid M, Qayyum A, Awais M, Alnowami M, Khan MK. Medical image analysis using convolutional neural networks: A review. Journal of Medical System. 2018;42:226. https://doi.org/10.1007/s10916-018-1088-1
  7. Zhang K, Zuo W, Chen Y, Meng D. Beyond a Gaussian denoiser: Residual learning of deep cnn for image denoising. IEEE Transactions on Image Processing. 2017;26(7):3142-55. https://doi.org/10.1109/TIP.2017.2662206
  8. Burger HC, Schuler C, Hamelling S. Image denoising: Can plain neural networks compete with BM3D? Computer Vision and Pattern Recgnotion(CVPR). IEEE Conference on IEEE; 2012.
  9. Chen Y, Pock T. Trainable nonlinear reaction diffusion: A flexible framework for fast and effective image restoration. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2017;39(6):1256-72. https://doi.org/10.1109/TPAMI.2016.2596743
  10. Jain V, Seung S. Natural image denoising with convolutional networks. Advances in Neural Information Processing Systems. 2009:769-76.
  11. Heo KH, Lim DH. Noise reduction using patchbased CNN in images. Journal of Korean Data Information Science Society. 2019;30(2):349-63. https://doi.org/10.7465/jkdi.2019.30.2.349
  12. Cho YH, Lee SJ, Lee YI, Park CR. Efficiency of Median Wiener Filter Algorithm for Noise Reduction in PET/MR Images: A Phantom Study. Journal of Radiological Science and Technology. 2021;44(3);225-9. https://doi.org/10.17946/JRST.2021.44.3.225
  13. Krull A, Buchholz TO, Jug F. Noise2void-learning denoising from single noisy images. Proceedings of the IEEE Conference on Computer Vision and Pattern Rcognition(CVPR); 2019.
  14. Leonard J, Kramer MA. Improvement of the backpropagation algorithm for training neural networks. Computer & Chemical engineering. 1990;14(3):337-41. https://doi.org/10.1016/0098-1354(90)87070-6
  15. Anwar SM, Brames N. Real image denoising with feature attention. Processing of the IEEE International Conference on Computer Vision(ICCV); 2019.
  16. Lehtinen J, Munkberg J, Hasselgren J, Laine S, Karras T, Aittala M, Alia T. Noise2Nois: Learning image restoration without clean data. International Conference on Machine Learning(ICML). 2018;80:2965-2974.
  17. Aide N, Lasnon C, Veit-Haibach P, Sera T, Sattler B, Boellaard R. EANM/EARL harmonization strategies in PET quantification: From daily practice to multicenter oncological studies. European Journal of Nuclear Medicine Molecular Imaging. 2017;44:17-31.
  18. Ronneberger O, Fischer P, Brox T. U-net: Convolutional networks for biomedical image segmentation. MIC-CAI, Springer, 2015:234-241.
  19. Zhao H, Gallo O, Frosio I, Kautz J. Loss function for Image restoration with neural networks. IEEE Transactions on Computational Imaging. 2017;3(1):47-57. https://doi.org/10.1109/TCI.2016.2644865
  20. Hofheinz F, Langer J, Beuthien-Baumann B, Oehme L, Steinbach J, Kotzerke J, Vanden Hoff J. Suitability of bilateral filtering for edge-preserving noise reduction in PET. European Journal of Nuclear Medial Molecular Imaging Research. 2011;1:23.
  21. Chen KT, Gong E, Carvalho Macruz FB, Xu J, Boumis A, Khalighi M. Ultra low dose 18F-flor-betaben amyloid PET imaging using deep learning with multi contrast MRI inputs. Radiology. 2019;290(3):649-56. https://doi.org/10.1148/radiol.2018180940
  22. Lu W, Onofrey JA, Lu Y, Shi L, Ma T, Liu Y, Liu C. An investigation of quantitative accuracy for deep learning based denosing in oncological PET. Physics in Medicine Biology. 2019;(64):165019.
  23. Jeong YJ, Park HS, Jeong JE, Yoon HJ, Jeon K, Cho K, Kang DY. Restoration of amyloid PET images obtained with short time data using a generative adversarial networks framework. Scientific Reports. 2021;11:4825. https://doi.org/10.1038/s41598-021-84358-8
  24. Kim YJ, Kim KG. Development of an optimized Deep Learning Model for Medical Imaging. Journal of the Korean Society of Radiology. 2020;81(6):1274-89. https://doi.org/10.3348/jksr.2020.0171