Noise‐assisted hybrid attention networks for low‐dose PET and CT denoising

Hengzhi Xue; Yudong Yao; Yueyang Teng

doi:10.1002/mp.17430

journal article Oct 21, 2024

Noise‐assisted hybrid attention networks for low‐dose PET and CT denoising

Hengzhi Xue Yudong Yao

Yueyang Teng

Medical Physics Vol. 52 No. 1 pp. 444-453 · Wiley

View at Publisher Save 10.1002/mp.17430

Abstract

AbstractBackgroundPositron emission tomography (PET) and computed tomography (CT) play a vital role in tumor‐related medical diagnosis, assessment, and treatment planning. However, full‐dose PET and CT pose the risk of excessive radiation exposure to patients, whereas low‐dose images compromise image quality, impacting subsequent tumor recognition and disease diagnosis.PurposeTo solve such problems, we propose a Noise‐Assisted Hybrid Attention Network (NAHANet) to reconstruct full‐dose PET and CT images from low‐dose PET (LDPET) and CT (LDCT) images to reduce patient radiation risks while ensuring the performance of subsequent tumor recognition.MethodsNAHANet contains two branches: the noise feature prediction branch (NFPB) and the cascaded reconstruction branch. Among them, NFPB providing noise features for the cascade reconstruction branch. The cascaded reconstruction branch comprises a shallow feature extraction module and a reconstruction module which contains a series of cascaded noise feature fusion blocks (NFFBs). Among these, the NFFB fuses the features extracted from low‐dose images with the noise features obtained by NFPB to improve the feature extraction capability. To validate the effectiveness of the NAHANet method, we performed experiments using two public available datasets: the Ultra‐low Dose PET Imaging Challenge dataset and Low Dose CT Grand Challenge dataset.ResultsAs a result, the proposed NAHANet achieved higher performance on common indicators. For example, on the CT dataset, the PSNR and SSIM indicators were improved by 4.1 dB and 0.06 respectively, and the rMSE indicator was reduced by 5.46 compared with the LDCT; on the PET dataset, the PSNR and SSIM was improved by 3.37 dB and 0.02, and the rMSE was reduced by 9.04 compared with the LDPET.ConclusionsThis paper proposes a transformer‐based denoising algorithm, which utilizes hybrid attention to extract high‐level features of low dose images and fuses noise features to optimize the denoising performance of the network, achieving good performance improvements on low‐dose CT and PET datasets.

Topics

No keywords indexed for this article. Browse by subject →

References

42

[1]

10.1109/tnnls.2023.3266551

[2]

10.1016/j.media.2020.101770

[3]

JiangC PanY ShenD.TriDoRNet: Reconstruction of standard‐dose Pet from low‐dose Pet in triple (Projection Image and Frequency) domains.2023 IEEE 20th International Symposium on Biomedical Imaging (ISBI).IEEE;2023:1‐5. 10.1109/isbi53787.2023.10230514

[4]

10.1088/1361-6560/ab4891

[5]

10.1016/j.bspc.2023.105197

[6]

10.1016/j.neuroimage.2018.03.045

[7]

Low-Dose CT With a Residual Encoder-Decoder Convolutional Neural Network

Hu Chen, Mannudeep K. Kalra, Feng Lin et al.

IEEE Transactions on Medical Imaging 10.1109/tmi.2017.2715284

[8]

Low-Dose CT Image Denoising Using a Generative Adversarial Network With Wasserstein Distance and Perceptual Loss

Qingsong Yang, Pingkun Yan, Yanbo Zhang et al.

IEEE Transactions on Medical Imaging 10.1109/tmi.2018.2827462

[9]

10.1109/tmi.2018.2832217

[10]

10.1109/tmi.2019.2963248

[11]

LiangT JinY LiY WangT.EDCNN: Edge enhancement‐based densely connected network with compound loss for low‐dose CT denoising. In:15th IEEE International conference on signal processing (ICSP).IEEE;2020;193‐198. 10.1109/icsp48669.2020.9320928

[12]

CTformer: convolution-free Token2Token dilated vision transformer for low-dose CT denoising

Dayang Wang, Fenglei Fan, Zhan Wu et al.

Physics in Medicine and Biology 10.1088/1361-6560/acc000

[13]

10.1016/j.artmed.2023.102609

[14]

10.1109/tmi.2017.2708987

[15]

10.1016/j.compbiomed.2022.106061

[16]

SharmaV KhuranaA YenamandraS AwateSP.Semi‐supervised deep expectation‐maximization for low‐dose PET‐CT. In:2022 IEEE 19th International Symposium on Biomedical Imaging (ISBI).IEEE;2022:1‐5. 10.1109/isbi52829.2022.9761601

[17]

10.1016/j.compbiomed.2023.107391

[18]

10.1016/j.media.2023.102787

[19]

Transformers in Vision: A Survey

Salman Khan, Muzammal Naseer, Munawar Hayat et al.

ACM Computing Surveys 10.1145/3505244

[20]

DilateFormer: Multi-Scale Dilated Transformer for Visual Recognition

Jiayu Jiao, Yu-Ming Tang, Kun-Yu Lin et al.

IEEE Transactions on Multimedia 10.1109/tmm.2023.3243616

[21]

LiuK DuX LiuS ZhengY WuX JinC.DDT: Dual‐branch Deformable Transformer for Image Denoising. In IEEE International Conference on Multimedia and Expo (ICME). IEEE;2023:2765‐2770. 10.1109/icme55011.2023.00470

[22]

ZhuL WangX KeZ ZhangW LauRW.BiFormer: vision transformer with bi‐level routing attention. In:Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.IEEE;2023:10323‐10333. 10.1109/cvpr52729.2023.00995

[23]

LiY FanY XiangX et al.Efficient and explicit modelling of image hierarchies for image restoration. In:Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.IEEE;2023:18278‐18289. 10.1109/cvpr52729.2023.01753

[24]

ZamirSW AroraA KhanS HayatM KhanFS YangMH.Restormer: efficient transformer for high‐resolution image restoration. In:Proceedings of the IEEE/CVF conference on computer vision and pattern recognition.IEEE;2022:5728‐5739. 10.1109/cvpr52688.2022.00564

[25]

ChenJ LuY YuQ et al.TransUNet: Transformers Make Strong Encoders for Medical Image Segmentation. arXiv preprint arXiv:2102.04306;2021.

[26]

H2Former: An Efficient Hierarchical Hybrid Transformer for Medical Image Segmentation

Along He, KAI WANG, Tao Li et al.

IEEE Transactions on Medical Imaging 10.1109/tmi.2023.3264513

[27]

10.1109/tmi.2023.3250474

[28]

10.1109/tmi.2022.3211764

[29]

ResViT: Residual Vision Transformers for Multimodal Medical Image Synthesis

Onat Dalmaz, Mahmut Yurt, Tolga çukur

IEEE Transactions on Medical Imaging 10.1109/tmi.2022.3167808

[30]

10.1016/j.compmedimag.2022.102150

[31]

10.1109/tmi.2023.3321064

[32]

WangD WuZ YuH.TED‐net: convolution‐free T2T vision transformer‐based encoder‐decoder dilation network for low‐dose CT denoising. In:Machine Learning in Medical Imaging: 12th International Workshop MLMI 2021 Held in Conjunction with MICCAI 2021 Proceedings 12.Springer;2021:416‐425. 10.1007/978-3-030-87589-3_43

[33]

10.1016/j.compbiomed.2023.107162

[34]

10.1002/mp.16550

[35]

10.1109/tmi.2023.3336237

[36]

LiuZ LinY CaoY et al.Swin transformer: hierarchical vision transformer using shifted windows. In:Proceedings of the IEEE/CVF international conference on computer vision.IEEE;2021:10012‐10022. 10.1109/iccv48922.2021.00986

[37]

WuH XiaoB CodellaN et al.CVT: introducing convolutions to vision transformers. In:Proceedings of the IEEE/CVF international conference on computer vision.IEEE;2021:22‐31. 10.1109/iccv48922.2021.00009

[38]

Xiao T "Early convolutions help transformers see better" Adv Neural Inf Process Syst (2021)

[39]

YuanK GuoS LiuZ ZhouA YuF WuW.Incorporating convolution designs into visual transformers. In:Proceedings of the IEEE/CVF International Conference on Computer Vision.IEEE;2021:579‐588. 10.1109/iccv48922.2021.00062

[40]

ChenX WangX ZhouJ QiaoY DongC.Activating more pixels in image super‐resolution transformer. In:Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.IEEE;2023:22367‐22377. 10.1109/cvpr52729.2023.02142

[41]

10.1002/mp.14594

[42]

10.1016/j.media.2021.102187

Metrics

3

Citations

42

References

Details

Published: Oct 21, 2024
Vol/Issue: 52(1)
Pages: 444-453
License: View

Authors

H

Hengzhi Xue

College of Medicine and Biological Information Engineering Northeastern University Shenyang China

Y

Yudong Yao

College of Medicine and Biological Information Engineering Northeastern University Shenyang China; Department of Electrical and Computer Engineering Steven Institute of Technology Hoboken New Jersey USA

Y

Yueyang Teng

College of Medicine and Biological Information Engineering Northeastern University Shenyang China; Key Laboratory of Intelligent Computing in Medical Image, Ministry of Education Shenyang China

Funding

Natural Science Foundation of Liaoning Province Award: 2022‐MS‐114

Cite This Article

Hengzhi Xue, Yudong Yao, Yueyang Teng (2024). Noise‐assisted hybrid attention networks for low‐dose PET and CT denoising. Medical Physics, 52(1), 444-453. https://doi.org/10.1002/mp.17430

Noise‐assisted hybrid attention networks for low‐dose PET and CT denoising

You May Also Like