skip to main content
10.1145/3587716.3587772acmotherconferencesArticle/Chapter ViewAbstractPublication PagesicmlcConference Proceedingsconference-collections
research-article

TC-GAN: A Transformer-based Conditional Generative Adversarial Network for Low-Dose SPECT Image Reconstruction

Published: 07 September 2023 Publication History

Abstract

Low-dose single photon emission computed tomography (SPECT) technology can reduce radiation damage to the subject but lead to reconstructed image quality degradation, such as severe noise and artifacts. Many existing works utilize convolutional neural networks (CNNs) to improve the quality of reconstructed images at low doses and demonstrate impressive performance. However, the intrinsic locality nature of convolutional operation limits the ability of CNNs to encode explicitly long-range dependencies. Combining the long-range dependencies effectively during reconstruction might improve the quality of reconstructed images. Therefore, we designed a novel transformer-based network, dubbed TC-GAN, which combines the conditional generative adversarial network (CGAN) and CSwin Transformer for high-quality low-dose SPECT reconstruction. In TC-GAN, the CSwin transformer was integrated into the generator of CGAN as a strong feature extractor to establish the intra-relationships in the whole image and provide plenty of global contextual information for image reconstruction. The CGAN was used as the backbone of our network to synthesize reconstructed images with accurate structure and pixel-level details via adversarial training fashion. In the experiment, we used the Monte Carlo simulation software SIMIND to produce low-dose SPECT image datasets. The results show that the proposed method has better reconstruction performance than several recent methods. In terms of visual analysis, the proposed method maintains more accurate image structure and details while noise suppression. In terms of quantitative analysis, the proposed method obtains the best results in global image quality metrics. Our results suggest that the proposed network has good potential to reduce radiation dose without compromising the reconstructed image quality.

References

[1]
Brenner David J, Sachs Rainer K. 2006. Estimating radiation-induced cancer risks at very low doses: Rationale for using a linear no-threshold approach. Radiation and Environmental Biophysics. 44. 253-6
[2]
Bevelacqua J. J. 2010. Practical and Effective ALARA. Health Physics. 98. 39-47
[3]
Jiang Ying, Li Si, Xu Yuesheng. 2019. A Higher-Order Polynomial Method for SPECT Reconstruction. IEEE Transactions on Medical Imaging. 38. 1271-83
[4]
Chen Yun, Huang Jiasheng, Li Si, Lu Yao, Xu Yuesheng. 2020. A content-adaptive unstructured grid based integral equation method with the TV regularization for SPECT reconstruction. Inverse Problems and Imaging. 14. 27-52
[5]
Tang Xinhuang, Schmidtlein Charles Ross, Li Si, Xu Yuesheng. 2021. An integral equation model for PET imaging. International Journal of Numerical Analysis and Modeling. 18. 834-64
[6]
Yang Yan, Sun Jian, Li Huibin, Xu Zongben. 2016. Deep ADMM-Net for compressive sensing MRI. 30th Annual Conference on Neural Information Processing Systems (NIPS). 29.
[7]
Jin Kyong Hwan, McCann Michael T, Froustey Emmanuel, Unser Michael. 2017. Deep convolutional neural network for inverse problems in imaging. IEEE Transactions on Image Processing. 26. 4509-22
[8]
Chen Hu, Zhang Yi, Chen Yunjin, Zhang Junfeng, Zhang Weihua, Sun Huaiqiang, 2018. LEARN: Learned experts’ assessment-based reconstruction network for sparse-data CT. IEEE Transactions on Medical Imaging. 37. 1333-47
[9]
Li Ziheng, Zhang Wenkun, Wang Linyuan, Cai Ailong, Liang Ningning, Yan Bin, 2019. A sinogram inpainting method based on generative adversarial network for limited-angle computed tomography. 15th International Meeting on Fully Three-Dimensional Image Reconstruction in Radiology and Nuclear Medicine. 11072.
[10]
Chrysostomou Charalambos, Koutsantonis Loizos, Lemesios Christos, Papanicolas Costas N. 2020. SPECT Angle Interpolation based on Deep Learning Methodologies. IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC.
[11]
Li Si, Ye Wenquan, Li Fenghuan. 2022. LU-Net: Combining LSTM and U-Net for sinogram synthesis in sparse-view SPECT reconstruction. Mathematical Biosciences and Engineering. 19. 4320-40
[12]
Ronneberger Olaf, Fischer Philipp, Brox Thomas. 2015. U-net: Convolutional networks for biomedical image segmentation. 18th International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI). 9351. 234-41
[13]
Chen Hu, Zhang Yi, Kalra Mannudeep K., Lin Feng, Chen Yang, Liao Peixi, 2017. Low-Dose CT with a residual encoder-decoder convolutional neural network. IEEE Transactions on Medical Imaging. 36. 2524-35
[14]
Wolterink Jelmer M, Leiner Tim, Viergever Max A, Išgum Ivana. 2017. Generative adversarial networks for noise reduction in low-dose CT. IEEE transactions on medical imaging. 36. 2536-45
[15]
Ye Jong Chul, Han Yoseob, Cha Eunju. 2018. Deep convolutional framelets: A general deep learning framework for inverse problems. SIAM Journal on Imaging Sciences. 11. 991-1048
[16]
Gao Yongfeng, Liang Zhengrong, Moore William, Zhang Hao, Pomeroy Marc J, Ferretti John A, 2019. A feasibility study of extracting tissue textures from a previous full-dose CT database as prior knowledge for Bayesian reconstruction of current low-dose CT images. IEEE transactions on medical imaging. 38. 1981-92
[17]
Kang Eunhee, Koo Hyun Jung, Yang Dong Hyun, Seo Joon Bum, Ye Jong Chul. 2019. Cycle‐consistent adversarial denoising network for multiphase coronary CT angiography. Medical physics. 46. 550-62
[18]
Ouyang Jiahong, Chen Kevin T, Gong Enhao, Pauly John, Zaharchuk Greg. 2019. Ultra‐low‐dose PET reconstruction using generative adversarial network with feature matching and task‐specific perceptual loss. Medical physics. 46. 3555-64
[19]
Sano Akira, Nishio Teiji, Masuda Takamitsu, Karasawa Kumiko. 2021. Denoising PET images for proton therapy using a residual U-net. Biomedical Physics and Engineering Express. 7.
[20]
Goodfellow Ian J, Pouget-Abadie Jean, Mirza Mehdi, Xu Bing, Warde-Farley David, Ozair Sherjil, 2014. Generative adversarial nets. 28th Conference on Neural Information Processing Systems (NIPS). 3. 2672-80
[21]
Shamshad Fahad, Khan Salman, Waqas Zamir Syed, Haris Khan Muhammad, Hayat Munawar, Shahbaz Khan Fahad, 2022. Transformers in Medical Imaging: A Survey. arXiv:2201.09873
[22]
Dosovitskiy Alexey, Beyer Lucas, Kolesnikov Alexander, Weissenborn Dirk, Zhai Xiaohua, Unterthiner Thomas, 2020. An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale. arXiv:2010.11929
[23]
Liu Ze, Lin Yutong, Cao Yue, Hu Han, Wei Yixuan, Zhang Zheng, 2021. Swin Transformer: Hierarchical Vision Transformer using Shifted Windows. 18th IEEE/CVF International Conference on Computer Vision (ICCV). 9992-10002
[24]
Dong Xiaoyi, Bao Jianmin, Chen Dongdong, Zhang Weiming, Yu Nenghai, Yuan Lu, 2022. CSWin Transformer: A General Vision Transformer Backbone with Cross-Shaped Windows. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). 12114-24
[25]
Chen Jieneng, Lu Yongyi, Yu Qihang, Luo Xiangde, Adeli Ehsan, Wang Yan, 2021. TransUNet: Transformers Make Strong Encoders for Medical Image Segmentation. arXiv:2102.04306
[26]
Dai Yin, Gao Yifan, Liu Fayu. 2021. Transmed: Transformers advance multi-modal medical image classification. Diagnostics. 11. 1384
[27]
Luthra Achleshwar, Sulakhe Harsh, Mittal Tanish, Iyer Abhishek, Yadav Santosh. 2021. Eformer: Edge Enhancement based Transformer for Medical Image Denoising. arXiv:2109.08044
[28]
Wang Dayang, Wu Zhan, Yu Hengyong. 2021. Ted-net: Convolution-free t2t vision transformer-based encoder-decoder dilation network for low-dose ct denoising. International Workshop on Machine Learning in Medical Imaging. 416-25
[29]
Zhang Zhicheng, Yu Lequan, Liang Xiaokun, Zhao Wei, Xing Lei. 2021. TransCT: Dual-Path Transformer for Low Dose Computed Tomography. International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI). 12906. 55-64
[30]
Krol Andrzej, Li Si, Shen Lixin, Xu Yuesheng. 2012. Preconditioned alternating projection algorithms for maximum a posteriori ECT reconstruction. Inverse Problems. 28.
[31]
Isola Phillip, Zhu Jun-Yan, Zhou Tinghui, Efros Alexei A. 2017. Image-to-image translation with conditional adversarial networks. 30th IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). 5967-76
[32]
Ljungberg Michael, Strand Sven-Erik, King Michael A. 2012. Monte Carlo calculations in nuclear medicine: applications in diagnostic imaging (Boca Raton: CRC Press)

Cited By

View all

Index Terms

  1. TC-GAN: A Transformer-based Conditional Generative Adversarial Network for Low-Dose SPECT Image Reconstruction
          Index terms have been assigned to the content through auto-classification.

          Recommendations

          Comments

          Information & Contributors

          Information

          Published In

          cover image ACM Other conferences
          ICMLC '23: Proceedings of the 2023 15th International Conference on Machine Learning and Computing
          February 2023
          619 pages
          ISBN:9781450398411
          DOI:10.1145/3587716
          Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than the author(s) must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected].

          Publisher

          Association for Computing Machinery

          New York, NY, United States

          Publication History

          Published: 07 September 2023

          Permissions

          Request permissions for this article.

          Check for updates

          Author Tags

          1. CGAN
          2. CSwin transformer
          3. Low dose SPECT reconstruction

          Qualifiers

          • Research-article
          • Research
          • Refereed limited

          Funding Sources

          Conference

          ICMLC 2023

          Contributors

          Other Metrics

          Bibliometrics & Citations

          Bibliometrics

          Article Metrics

          • 0
            Total Citations
          • 50
            Total Downloads
          • Downloads (Last 12 months)33
          • Downloads (Last 6 weeks)5
          Reflects downloads up to 20 Jan 2025

          Other Metrics

          Citations

          Cited By

          View all

          View Options

          Login options

          View options

          PDF

          View or Download as a PDF file.

          PDF

          eReader

          View online with eReader.

          eReader

          HTML Format

          View this article in HTML Format.

          HTML Format

          Media

          Figures

          Other

          Tables

          Share

          Share

          Share this Publication link

          Share on social media