A Survey on Clinical Oriented CT Image Denoising

doi:10.11871/jfdc.10-1649.2021.06.003

Abstract

Abstract:

[Objective] Low-dose CT scanning technology has been widely used in clinical diagnoses such as early screening of lung nodules and lung cancer. However, the imaging noise caused by the reduction of radiation dose has restricted the further improvement of diagnostic accuracy.[Methods] This article comprehensively studies the development of low-dose CT image noise reduction technology. Starting with the traditional methods such as iterative optimization, it analyzes the current methods of low-dose CT image noise reduction based on machine learning. [Results] The existing methods based on machine learning, on the one hand, use artificially assumed noise distribution and construct artificial data sets to train and test the model, ignoring the diverse characteristics and intensity of noise in clinical applications; on the other hand, due to the "black box" characteristics of deep neural networks, the existing low-dose CT image noise reduction models based on deep neural networks are insufficient in interpretability. [Conclusions] CT image noise reduction should be oriented to the requirements of clinical scenarios, fully consider the mechanism of noise formation and the true distribution of image noise, and combine high-level tasks such as lesion detection and the reading behavior of clinicians to explore the brand-new paradigm of low-dose CT image noise reduction for clinical demands.

Key words: CT image, deep learning, image denoising, image quality evaluation

PU Xiaorong,HUANG Jiaxin,LIU Junchi,SUN Jiayu,LUO Jixiang,ZHAO Yue,CHEN Kecheng,REN Yazhou. A Survey on Clinical Oriented CT Image Denoising[J]. Frontiers of Data and Computing, 2021, 3(6): 35-49.

Figures/Tables 5

Fig.1

Table 1

Table 2

Fig.2

Fig.3

References 61

[1]	Pan T. Computed Tomography: from Photon Statistics to Modern Cone-Beam CT[J]. Journal of Nuclear Medicine, 2009,50(7):1194.
[2]	张秀文, 张永寿, 刘乃智. PET-CT工作原理及应用[J]. 中国医学装备, 2012, ( 11):22-25.
[3]	石明国. 实用CT影像技术学[M]. 西安: 陕西科学技术出版社, 1995: 233-235.
[4]	朱兆丰. CT图像中噪声产生机理与维修[J]. 中国医疗设备, 2005,20(11):50-51.
[5]	Naidich D P, Marshall C H, Gribbin C, et al. Low-dose CT of the lungs: Preliminary observations[J]. Radiology, 1990,175(3):729-731.
[6]	吴宁, 赵世俊. 积极规范地开展低剂量螺旋CT肺癌筛查[J]. 中华放射学杂志, 2015(5):2.
[7]	Guo L, Chen Q, Shen Y, et al. Evaluation of a Low-Dose Computed Tomography Lung Cancer Screening Program in Henan, China[J]. JAMA Network Open, 2020,3(11):e2019039.
[8]	Shi J, Wang L, Wang S, Chen Y, et al. 2020. Applications of deep learning in medical imaging: a survey. Journal of Image and Graphics, 25(10):1953-1981.
[9]	Wang J, Lu H, Li T, et al. Sinogram noise reduction for low-dose CT by statistics-based nonlinear filters[C]. Image Processing pt.3, Progress in Biomedical Optics and Imaging, 2005,6(24):2058-2066.
[10]	Mileto A, Guimaraes L, Mccollough C, et al. State of the Art in Abdominal CT: The Limits of Iterative Recons-truction Algorithms[J]. Radiology, 2019,293(3):491-503.
[11]	Li M, Hsu W, Xie X, et al. SACNN: Self-Attention Con-volutional Neural Network for Low-Dose CT Denoising with Self-supervised Perceptual Loss Network[J]. IEEE Transactions on Medical Imaging, 2020,39(7):2289-2301.
[12]	Hara A, Paden R, AC Silva, et al. Iterative reconstruction technique for reducing body radiation dose at CT: feasi-bility study[J]. American Journal of Roentgenology, 2009,193(3):764-771.
[13]	Sidky E, Pan X. Image reconstruction in circular cone-beam computed tomography by constrained, total-variation minimization[J]. Physics in Medicine & Bio-logy, 2008,53(17):4777-4807.
[14]	Yang C, Gao D, Cong N, et al. Bayesian statistical recon-struction for low-dose X-ray computed tomography using an adaptive-weighting nonlocal prior[J]. Computerized Medical Imaging & Graphics, 2009,33(7):495-500.
[15]	Xu Q, Yu H, Mou X, et al. Low-dose X-ray CT reconstr-uction via dictionary learning[J]. IEEE transactions on medical imaging, 2012,31(9):1682-1697.
[16]	Cai J, Jia X, Gao H, et al. Cine cone beam CT reconstruction using low-rank matrix factorization: algorithm and a proof-of-principle study[J]. IEEE transactions on medical imaging, 2014,33(8):1581-1591.
[17]	Li Z, Yu L, Trzasko J D, et al. Adaptive nonlocal means filtering based on local noise level for CT denoising[J]. Medical physics, 2014,41(1):11908.
[18]	Feruglio P, Vinegoni C, Gros J, et al. Block matching 3D random noise filtering for absorption optical projection tomography[J]. Physics in Medicine & Biology, 2010,55(18):5401-5415.
[19]	Chen H, Zhang Y, Zhang W, et al. Low-dose CT den-oising with convolutional neural network[C]. 2017 IEEE 14th International Symposium on Biomedical Imaging (ISBI 2017), 2017: 143-146.
[20]	Chen H, Zhang Y, Kalra M K, et al. Low-dose CT with a residual encoder-decoder convolutional neural network[J]. IEEE transactions on medical imaging, 2017,36(12):2524-2535.
[21]	Mao X, Shen C, Yang Y. Image restoration using very deep convolutional encoder-decoder networks with sym-metric skip connections[C]. In Proceedings of the 30 ^th International Conference on Neural Information Proce-ssing Systems (NIPS’16) , 2016: 2810-2818.
[22]	He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C]. Proceedings of the IEEE confe-rence on computer vision and pattern recognition, 2016: 770-778.
[23]	Hu C, Zhang Y, Zhang W, et al. Low-dose CT via convo-lutional neural network[J]. Biomedical Optics Express, 2017,8(2):679-694.
[24]	Yang Q, Yan P, Kalra M K, et al. CT image denoising with perceptive deep neural networks[J]. arXiv preprint arXiv: 1702. 07019, 2017.
[25]	Simonyan K, Zisserman A. Very Deep Convolutional Networks for Large-Scale Image Recognition[C]. The 3rd International Conference on Learning Representations (ICLR2015). https://arxiv.org/abs/1409.1556.
[26]	张雄, 杨琳琳, 上官宏等. 基于生成对抗网络和噪声水平估计的低剂量CT图像降噪方法[J]. 电子与信息学报, 2021,43(8):2404-2413.
[27]	Gholizadeh-Ansari M, Alirezaie J, Babyn P. Low-dose CT Denoising with Dilated Residual Network[C]. 40 ^th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) , 2018: 5117-5120.
[28]	Kadimesetty V S, Gutta S, Ganapathy S, et al. Convo-lutional Neural Network-Based Robust Denoising of Low-Dose Computed Tomography Perfusion Maps[J]. IEEE Transactions on Radiation and Plasma Medical Sciences, 2018,3(2):137-152.
[29]	Liu P, Li Y, Basha M, et al. Neural Network Evolution Using Expedited Genetic Algorithm for Medical Image Denoising[C]. Medical Image Computing and Computer Assisted Intervention - MICCAI, 2018: 12-20.
[30]	Han Y, Ye J. Framing U-Net via Deep Convolutional Fra-melets: Application to Sparse-View CT[J]. IEEE Tran-sactions on Medical Imaging, 2018,37(6):1418-1429.
[31]	Khoroushadi M, Sadegh M. Enhancement in low-dose computed tomography through image denoising techniques: Wavelets and deep learning[D]. Ph.D. thesis, ProQuest Dissertations Publishing. 2018.
[32]	Green M, Marom E, Konen E, et al. Learning Real Noise for Ultra-Low Dose Lung CT Denoising[C]. Patch-Based Techniques in Medical Imaging, 2018: 3-11.
[33]	Li M, Hsu W, X Xie, et al. SACNN: Self-Attention Con-volutional Neural Network for Low-Dose CT Den-oising with Self-supervised Perceptual Loss Network[J]. IEEE Transactions on Medical Imaging, 2020,39(7):2289-2301.
[34]	Hendriksen A A, Pelt D M, Batenburg K J. Noise2Inverse: Self-supervised deep convolutional denoising for linear inverse problems in imaging[J]. IEEE Transactions on Computational Imaging, 2020,6:1320-1335.
[35]	Wu D, Ren H, Li Q. Self-supervised Dynamic CT Per-fusion Image Denoising with Deep Neural Networks[J]. IEEE Transactions on Radiation and Plasma Medical Sciences, 2021,5(3):350-361.
[36]	Du W, Chen H, Yang H. Learning Invariant Representa-tion for Unsupervised Image Restoration[C]. 2020 IEEE/CVF Conference on Computer Vision and Pattern Recog-nition (CVPR), 2020: 14471-14480.
[37]	Chen K, Long K, Ren Y, Sun J, Pu X. Lesion-Inspired Denoising Network: Connecting Medical Image Denoi-sing and Lesion Detection[C]. Proceedings of the 29th ACM International Conference on Multimedia, 2021: 3283-3292.
[38]	Wolterink J M, Leiner T, Viergever M A, et al. Generative adversarial networks for noise reduction in low-dose CT[J]. IEEE transactions on medical imaging, 2017,36(12):2536-2545.
[39]	Isola P, Zhu J Y, Zhou T, et al. Image-to-Image Tra-nslation with Conditional Adversarial Networks[C]. 2017 IEEE Conference on Computer Vision and Pattern Recog-nition (CVPR), 2017: 5967-5976.
[40]	Goodfellow I J, Pouget-Abadie J, Mirza M, et al. Gene-rative Adversarial Networks[J]. Advances in Neural Infor-mation Processing Systems, 2014,3:2672-2680.
[41]	Burlina P M, Joshi N, Pekala M, et al. Automated grading of age-related macular degeneration from color fundus images using deep convolutional neural networks[J]. JAMA ophthalmology, 2017,135(11):1170-1176.
[42]	Maas A L, Hannun A Y, Ng A Y. Rectifier Nonlinearities improve neural network acoustic models[C]. In Proc. ICML, 2013,30(1):3.
[43]	Glorot X, Bengio Y. Understanding the difficulty of training deep feedforward neural networks[C]. Procee-dings of the thirteenth international conference on artifi-cial intelligence and statistics. JMLR Workshop and Con-ference Proceedings, 2010: 249-256.
[44]	Yu A, Liu X, Wei X, et al. Generative Adversarial Netw-orks with Dense Connection for Optical Coherence Tomo-graphy Images Denoising[C]. 2018 11th Intern-ational Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI), 2018: 1-5.
[45]	Abbasi A, Monadjemi A, Fang L, et al. Three-dim-ensional optical coherence tomography image denoising through multi-input fully-convolutional networks[J]. Computers in Biology and Medicine, 2019,108:1-8.
[46]	Ma Y, Chen X, Zhu W, et al. Speckle noise reduction in optical coherence tomography images based on edge-sensitive cGAN[J]. Biomedical optics express, 2018,9(11):5129-5146.
[47]	Hong Z, Fan X, Jiang T, et al. End-to-End Unpaired Image Denoising with Conditional Adversarial Networks[J]. Proceedings of the AAAI Conference on Artificial Intelli-gence, 2020,34(4):4140-4149.
[48]	Park H S, Baek J, You S K, et al. Unpaired image denoi-sing using a generative adversarial network in X-ray CT[J]. IEEE Access, 2019,7:110414-110425.
[49]	You C, Yang Q, Shan H, et al. Structurally-sensitive Mul-ti-scale Deep Neural Network for Low-Dose CT Denoi-sing[J]. IEEE Access, 2018,6:41839-41855.
[50]	Yang Q, Yan P, Zhang Y, et al. Low-Dose CT Image De-noising Using a Generative Adversarial Network with Wasserstein Distance and Perceptual Loss[J]. IEEE Tran-sactions on Medical Imaging, 2018,37(6):1348-1357.
[51]	Chen K, Huang J, Sun J, et al. Task-Driven Deep Lear-ning for LDCT Image Denoising[C]. In The Fourth Inter-national Symposium on Image Computing and Digital Medicine. Association for Computing Machinery, New York, NY, USA, 2020: 35-39.
[52]	段影影. 医学图像质量评价方法研究[D]. 广东:南方医科大学, 2010.
[53]	Shan H, Padole A, Homayounieh F, et al. Competitive performance of a modularized deep neural network com-pared to commercial algorithms for low-dose CT image reconstruction[J]. Nature Machine Intelligence, 2019,1(6):269-276.
[54]	段影影, 马建华, 陈武凡, 等. 改进的结构相似医学图像质量评价方法[J]. 计算机工程与应用, 2010,46(2):145-149.
[55]	张剑华, 张自然, 汪晓妍, 等. 基于结构显著性的医学图像质量评价[J]. 浙江工业大学学报, 2015,43(6):636-641.
[56]	Virendra K, Gu Y, Satrajit B, et al. Radiomics: the pro-cess and the challenge[J]. Magnetic Resonance Imaging, 2012,30(9):1234-1248.
[57]	陆锐, 何健, 周科峰, 等. 自动管电流调制扫描及ASIR重建下胸部CT噪声指数的优化[J]. 医学影像学杂志, 2016,26(06):1012-1016.
[58]	Junior F, Raniery R, Koenigkam-Santos M, Cipriano F, et al. Radiomics-based features for pattern recognition of lung cancer histopathology and metastases[J]. Computer Methods and Programs in Biomedicine, 2018,159:23-30.
[59]	Zhang Q, Xiao Y, Suo J, et al. Sonoelastomics for Breast Tumor Classification: A Radiomics Approach with Clustering-Based Feature Selection on Sonoelastography[J]. Ultrasound in medicine & biology, 2017,43(5):1058-1069.
[60]	Coroller T, Grossmann P, Hou Y, et al. CT-Based radiomic signature predicts distant metastasis in lung adenocar-cinoma[J]. Radiotherapy and Oncology, 2015,114(3):345-350.
[61]	Visvikis D, Rest C, Jaouen V, et al. Artificial intelligence, machine (deep) learning and radio(geno)mics: definitions and nuclear medicine imaging applications[J]. European Journal of Nuclear Medicine, 2019,46(13):2630-2637.

参考文献	噪声类型	应用场景	关键词	评价指标
Gholizadeh等 (2018) ^[27]	外加白噪声	LDCT/X射线图像降噪	膨胀卷积	PSNR\SSIM
Kadimesetty等 (2019) ^[28]	外加白噪声	LDCT图像降噪	强化学习/批归一化	PSNR\SSIM
Liu等 (2018) ^[29]	外加白噪声	CT灌注图像降噪	遗传算法	PSNR\SSIM\RMSE
Han等(2018)^[30]	真实噪声	CT图像重建	U-Net/跳跃连接	PSNR\SSIM
Khoroushadi等(2019) ^[31]	真实噪声	LDCT图像降噪	小波变换	PSNR\SSIM
Green等(2018) ^[32]	真实噪声	肺部超低剂量CT图像降噪	卷积神经网络	PSNR\双盲评分
Li等(2020) ^[33]	真实噪声	LDCT图像降噪	自注意力/自监督	PSNR\SSIM\RMSE
Hendriksen等(2020) ^[34]	真实噪声	LDCT图像降噪	自监督/无配对样本	PSNR\SSIM\RMSE
Wu等(2020)^[35]	真实噪声	CT灌注图像降噪	自监督/无配对样本	PSNR\SSIM
Du等(2020)^[36]	真实噪声	图像超分/LDCT图像降噪	无监督/域适应	PSNR\SSIM\RMSE
Chen等(2021)^[37]	真实噪声	LDCT图像降噪	协同训练	PSNR\双盲评分
Chen等(2020)^[51]	真实噪声	LDCT图像降噪	构造成对伪样本	PSNR\SSIM

参考文献	噪声类型	应用场景	关键词	评价指标
Yu等 (2018) ^[44]	外加白噪声	光学相干层析成像降噪	深度网络/跳跃连接	PSNR\SSIM\RMSE
Abbasi等(2019) ^[45]	外加白噪声	3D-MRI/LDCT图像降噪	强化学习/全卷积网络	PSNR\SSIM
Ma等(2018) ^[46]	真实噪声	光学相干层析成像降噪	GAN/边缘优先	PSNR\SSIM\RMSE
Hong等(2020) ^[47]	真实噪声	CT图像重建	CGAN/端到端网络	PSNR\SSIM
Park等(2019)^[48]	真实噪声	LDCT图像降噪	GAN/无监督	PSNR\SSIM\RMSE
You等(2018) ^[49]	真实噪声	LDCT图像降噪	GAN/无监督	PSNR\双盲评分
Yang等(2018) ^[50]	真实噪声	LDCT图像降噪	WGAN/感知损失	PSNR\双盲评分
Zhang等(2021)^[26]	真实噪声	LDCT图像降噪	U-Net	PSNR\SSIM\RMSE