Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
102 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
6 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Low-Dose CT Reconstruction Using Dataset-free Learning (2402.11186v3)

Published 17 Feb 2024 in eess.IV and physics.med-ph

Abstract: Low-Dose computer tomography (LDCT) is an ideal alternative to reduce radiation risk in clinical applications. Although supervised-deep-learning-based reconstruction methods have demonstrated superior performance compared to conventional model-driven reconstruction algorithms, they require collecting massive pairs of low-dose and norm-dose CT images for neural network training, which limits their practical application in LDCT imaging. In this paper, we propose an unsupervised and training data-free learning reconstruction method for LDCT imaging that avoids the requirement for training data. The proposed method is a post-processing technique that aims to enhance the initial low-quality reconstruction results, and it reconstructs the high-quality images by neural work training that minimizes the $\ell_1$-norm distance between the CT measurements and their corresponding simulated sinogram data, as well as the total variation (TV) value of the reconstructed image. Moreover, the proposed method does not require to set the weights for both the data fidelity term and the plenty term. Experimental results on the AAPM challenge data and LoDoPab-CT data demonstrate that the proposed method is able to effectively suppress the noise and preserve the tiny structures. Also, these results demonstrate the rapid convergence and low computational cost of the proposed method. The source code is available at \url{https://github.com/linfengyu77/IRLDCT}.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (33)
  1. Computed tomography reconstruction using deep image prior and learned reconstruction methods. Inverse Problems 36, 094004. URL: https://dx.doi.org/10.1088/1361-6420/aba415, doi:10.1088/1361-6420/aba415. publisher: IOP Publishing.
  2. Noise2Self: Blind Denoising by Self-Supervision. URL: http://arxiv.org/abs/1901.11365, doi:10.48550/arXiv.1901.11365. arXiv:1901.11365 [cs, stat].
  3. Cine Cone Beam CT Reconstruction Using Low-Rank Matrix Factorization: Algorithm and a Proof-of-Principle Study. IEEE Trans. Med. Imaging 33, 1581–1591. URL: http://ieeexplore.ieee.org/document/6803058/, doi:10.1109/TMI.2014.2319055.
  4. Low-Dose CT With a Residual Encoder-Decoder Convolutional Neural Network. IEEE Transactions on Medical Imaging 36, 2524–2535. doi:10.1109/TMI.2017.2715284. conference Name: IEEE Transactions on Medical Imaging.
  5. Learned Full-Sampling Reconstruction From Incomplete Data. IEEE Transactions on Computational Imaging 6, 945–957. doi:10.1109/TCI.2020.2996751. conference Name: IEEE Transactions on Computational Imaging.
  6. A dataset-free deep learning method for low-dose CT image reconstruction. Inverse Problems 38, 104003. URL: https://dx.doi.org/10.1088/1361-6420/ac8ac6, doi:10.1088/1361-6420/ac8ac6. publisher: IOP Publishing.
  7. PET Image Reconstruction Using Deep Image Prior. IEEE Trans. Med. Imaging 38, 1655–1665. URL: https://ieeexplore.ieee.org/document/8581448/, doi:10.1109/TMI.2018.2888491.
  8. Hybrid-Collaborative Noise2Noise Denoiser for Low-Dose CT Images. IEEE Transactions on Radiation and Plasma Medical Sciences 5, 235–244. doi:10.1109/TRPMS.2020.3002178. conference Name: IEEE Transactions on Radiation and Plasma Medical Sciences.
  9. Iterative Reconstruction for Low-Dose CT Using Deep Gradient Priors of Generative Model. IEEE Transactions on Radiation and Plasma Medical Sciences 6, 741–754. doi:10.1109/TRPMS.2022.3148373. conference Name: IEEE Transactions on Radiation and Plasma Medical Sciences.
  10. Noise2Inverse: Self-supervised deep convolutional denoising for tomography. IEEE Trans. Comput. Imaging 6, 1320–1335. URL: http://arxiv.org/abs/2001.11801, doi:10.1109/TCI.2020.3019647. arXiv:2001.11801 [cs, eess, stat].
  11. Deep Convolutional Neural Network for Inverse Problems in Imaging. IEEE Transactions on Image Processing 26, 4509–4522. doi:10.1109/TIP.2017.2713099. conference Name: IEEE Transactions on Image Processing.
  12. DUG-RECON: A Framework for Direct Image Reconstruction Using Convolutional Generative Networks. IEEE Transactions on Radiation and Plasma Medical Sciences 5, 44–53. doi:10.1109/TRPMS.2020.3033172. conference Name: IEEE Transactions on Radiation and Plasma Medical Sciences.
  13. A deep convolutional neural network using directional wavelets for low-dose X-ray CT reconstruction. Med Phys 44, e360–e375. doi:10.1002/mp.12344.
  14. Sparse-View Spectral CT Reconstruction Using Spectral Patch-Based Low-Rank Penalty. IEEE Trans. Med. Imaging 34, 748–760. URL: http://ieeexplore.ieee.org/document/6985637/, doi:10.1109/TMI.2014.2380993.
  15. Noise2Noise: Learning Image Restoration without Clean Data. URL: http://arxiv.org/abs/1803.04189, doi:10.48550/arXiv.1803.04189. arXiv:1803.04189 [cs, stat].
  16. LoDoPaB-CT, a benchmark dataset for low-dose computed tomography reconstruction. Sci Data 8, 109. URL: https://www.nature.com/articles/s41597-021-00893-z, doi:10.1038/s41597-021-00893-z. number: 1 Publisher: Nature Publishing Group.
  17. Quantitative Comparison of Deep Learning-Based Image Reconstruction Methods for Low-Dose and Sparse-Angle CT Applications. Journal of Imaging 7, 44. URL: https://www.mdpi.com/2313-433X/7/3/44, doi:10.3390/jimaging7030044. number: 3 Publisher: Multidisciplinary Digital Publishing Institute.
  18. Sparse-sampling CT Sinogram Completion using Generative Adversarial Networks, in: 2020 13th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI), IEEE, Chengdu, China. pp. 640–644. URL: https://ieeexplore.ieee.org/document/9263571/, doi:10.1109/CISP-BMEI51763.2020.9263571.
  19. Iterative image reconstruction for cerebral perfusion CT using a pre-contrast scan induced edge-preserving prior. Phys. Med. Biol. 57, 7519–7542. URL: https://iopscience.iop.org/article/10.1088/0031-9155/57/22/7519, doi:10.1088/0031-9155/57/22/7519.
  20. PyTorch: An Imperative Style, High-Performance Deep Learning Library, in: Advances in Neural Information Processing Systems 32. Curran Associates, Inc., pp. 8024–8035. URL: http://papers.neurips.cc/paper/9015-pytorch-an-imperative-style-high-performance-deep-learning-library.pdf.
  21. Low-Dose X-ray CT Reconstruction via Dictionary Learning. IEEE Trans. Med. Imaging 31, 1682–1697. URL: http://ieeexplore.ieee.org/document/6188527/, doi:10.1109/TMI.2012.2195669.
  22. Image reconstruction in circular cone-beam computed tomography by constrained, total-variation minimization. Phys Med Biol 53, 4777–4807. doi:10.1088/0031-9155/53/17/021.
  23. Radiation dose associated with common computed tomography examinations and the associated lifetime attributable risk of cancer. Arch Intern Med 169, 2078–2086. doi:10.1001/archinternmed.2009.427.
  24. Deep Image Prior. Int J Comput Vis 128, 1867–1888. URL: http://arxiv.org/abs/1711.10925, doi:10.1007/s11263-020-01303-4. arXiv:1711.10925 [cs, stat].
  25. Image Quality Assessment: From Error Visibility to Structural Similarity. IEEE Trans. on Image Process. 13, 600–612. URL: http://ieeexplore.ieee.org/document/1284395/, doi:10.1109/TIP.2003.819861.
  26. Computationally efficient deep neural network for computed tomography image reconstruction. Med Phys 46, 4763–4776. doi:10.1002/mp.13627.
  27. Dynamic PET Image Reconstruction Using Nonnegative Matrix Factorization Incorporated With Deep Image Prior, in: 2019 IEEE/CVF International Conference on Computer Vision (ICCV), pp. 3126–3135. doi:10.1109/ICCV.2019.00322. iSSN: 2380-7504.
  28. Time-Dependent Deep Image Prior for Dynamic MRI. IEEE Trans. Med. Imaging 40, 3337–3348. URL: https://ieeexplore.ieee.org/document/9442767/, doi:10.1109/TMI.2021.3084288.
  29. Low-dose CT count-domain denoising via convolutional neural network with filter loss, in: Bosmans, H., Chen, G.H., Gilat Schmidt, T. (Eds.), Medical Imaging 2019: Physics of Medical Imaging, SPIE, San Diego, United States. p. 26. URL: https://www.spiedigitallibrary.org/conference-proceedings-of-spie/10948/2513479/Low-dose-CT-count-domain-denoising-via-convolutional-neural-network/10.1117/12.2513479.full, doi:10.1117/12.2513479.
  30. Half2Half: deep neural network based CT image denoising without independent reference data. Phys. Med. Biol. 65, 215020. URL: https://dx.doi.org/10.1088/1361-6560/aba939, doi:10.1088/1361-6560/aba939. publisher: IOP Publishing.
  31. Spectral CT Reconstruction With Image Sparsity and Spectral Mean. IEEE Trans. Comput. Imaging 2, 510–523. URL: http://ieeexplore.ieee.org/document/7567594/, doi:10.1109/TCI.2016.2609414.
  32. A dual-domain deep learning-based reconstruction method for fully 3D sparse data helical CT. Phys Med Biol 65, 245030. doi:10.1088/1361-6560/ab8fc1.
  33. Image reconstruction by domain-transform manifold learning. Nature 555, 487–492. URL: https://www.nature.com/articles/nature25988, doi:10.1038/nature25988. number: 7697 Publisher: Nature Publishing Group.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (3)
  1. Feng Wang (408 papers)
  2. Renfang Wang (5 papers)
  3. Hong Qiu (5 papers)
Citations (1)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now
X Twitter Logo Streamline Icon: https://streamlinehq.com

Tweets