2.793

2018影响因子

(CJCR)

  • 中文核心
  • EI
  • 中国科技核心
  • Scopus
  • CSCD
  • 英国科学文摘

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

融合MRI信息的PET图像去噪: 基于图小波的方法

易利群 盛玉霞 柴利

易利群, 盛玉霞, 柴利. 融合MRI信息的PET图像去噪: 基于图小波的方法. 自动化学报, 2021, 47(x): 1−10 doi: 10.16383/j.aas.c201036
引用本文: 易利群, 盛玉霞, 柴利. 融合MRI信息的PET图像去噪: 基于图小波的方法. 自动化学报, 2021, 47(x): 1−10 doi: 10.16383/j.aas.c201036
Yi Li-Qun, Sheng Yu-Xia, Chai Li. Dynamic PET images denoising with MRI information: A graph wavelet based method. Acta Automatica Sinica, 2021, 47(x): 1−10 doi: 10.16383/j.aas.c201036
Citation: Yi Li-Qun, Sheng Yu-Xia, Chai Li. Dynamic PET images denoising with MRI information: A graph wavelet based method. Acta Automatica Sinica, 2021, 47(x): 1−10 doi: 10.16383/j.aas.c201036

融合MRI信息的PET图像去噪: 基于图小波的方法

doi: 10.16383/j.aas.c201036
基金项目: 国家自然科学基金(61625305, 61501337)资助
详细信息
    作者简介:

    易利群:武汉科技大学信息科学与工程学院硕士研究生. 主要研究方向为医学图像处理, 信号处理. E-mail: ylqgenuine@sina.cn

    盛玉霞:武汉科技大学信息科学与工程学院副教授. 2014年获武汉科技大学控制科学与工程专业博士学位. 主要研究方向为图像处理, 图信号处理. 本文通信作者. E-mail: shengyuxia@wust.edu.cn

    柴利:武汉科技大学信息科学与工程学院教授. 2002年获香港科技大学电子工程系博士学位. 主要研究方向为分布式优化, 滤波器组框架, 图信号处理, 网络化控制系统. E-mail: chaili@wust.edu.cn

Dynamic PET Images Denoising with MRI Information: A Graph Wavelet Based Method

Funds: Supported by National Natural Science Foundation of P. R. China (61625305, 61501337)
More Information
    Author Bio:

    YI Li-Qun Master student at the School of Information Science and Engineering, Wuhan University of Science and Technology. Her research interests include medical image processing and signal processing

    SHENG Yu-Xia Associate professor at the School of Information Science and Engineering, Wuhan University of Science and Technology. She received the Ph. D. degree in control science and engineering from Wuhan University of Science and Technology in 2014. Her research interests include image processing and graph signal processing. Corresponding author of this paper

    CHAI Li Professor at the School of Information Science and Engineering, Wuhan University of Science and Technology. He received the Ph. D. degree in electrical engineering from Hong Kong University of Science and Technology in 2002. His research interests include distributed optimization, filter bank frames, graph signal processing, and networked control systems

  • 摘要: 正电子发射断层成像(Positron emission tomography, PET)是一种强大的核医学功能成像模式, 广泛地应用于临床诊断, 但PET图像的空间分辨率低且含有噪声, 有必要对PET图像进行去噪来提升PET图像的质量. 随着PET/MR等一体化成像设备的出现, 磁共振成像(Magnetic resonance imaging, MRI)的先验信息可用于PET图像去噪, 提高PET图像质量. 针对动态PET图像, 提出了一种融合MRI先验信息的PET图像图小波去噪新方法. 首先构建PET合成图像; 再将PET合成图像与MRI信息通过硬阈值方法进行融合; 接着在融合图像上构造图拉普拉斯矩阵; 最后通过图小波变换对动态PET图像去噪. 仿真实验结果表明, 与单独的图滤波、图小波去噪方法, 以及其他结合MRI的PET图像去噪方法相比, 本文方法有更高的信噪比, 更好地保留了病灶信息; 本文方法的去噪性能与VGG深度神经网络等基于学习的方法相当, 但不需要大量数据的训练, 计算复杂度低.
  • 图  1  从左到右依次为: (a) MR图像, (b) 合成图像

    Fig.  1  From left to right: (a) MR image, (b) composite image

    图  2  无病灶PET图像去噪结果

    Fig.  2  Denoising results of normal PET images

    图  3  从左到右依次为: (a)单病灶MR图像, (b)单病灶动态PET合成图像, (c)单病灶动态PET融合图像

    Fig.  3  From left to right: (a) abnormal MR image, (b) abnormal composite image, (c) abnormal fusion PET image

    图  4  单病灶PET图像去噪结果

    Fig.  4  Denoising results of abnormal PET images

    图  5  单病灶PET图像去噪残差图

    Fig.  5  Denoising residual map of abnormal PET images

    图  6  不同方法病灶点CRC与背景区域STD曲线图

    Fig.  6  CRC-STD curves of the different denoising methods for abnornal PET images

    表  1  本文方法参数设置

    Table  1  Parameter setting in this paper

    参数符号nk$\theta $$\eta $JThreshold
    参数设置511200.000840.75
    下载: 导出CSV

    表  2  无病灶情况下结合MRI的PET图像去噪方法比较

    Table  2  Comparison of PET image denoising methods incorporated with MRI on the normal dataset

    帧数文献[5]文献[13]本文方法
    6SNR10.45297.934311.7851
    RMSE1.65722.21461.4215
    12SNR10.94579.518412.2803
    RMSE3.08383.63452.6445
    18SNR11.168011.615312.6945
    RMSE19.510618.531216.3661
    24SNR11.067511.523212.6693
    RMSE20.492319.445017.0413
    下载: 导出CSV

    表  3  无病灶情况下由PET合成图像构图的方法与本文方法比较

    Table  3  Comparison of the methods of constructing graph by composite image with the proposed method on the normal dataset

    帧数图滤波图小波本文方法
    6SNR7.767011.406411.7851
    RMSE2.25771.48491.4215
    12SNR9.725811.796712.2803
    RMSE3.54882.79602.6445
    18SNR11.531712.539512.6945
    RMSE18.710416.660816.3661
    24SNR11.483112.463712.6693
    RMSE19.535017.449517.0413
    下载: 导出CSV

    表  4  单病灶情况下结合MRI的PET图像去噪方法比较

    Table  4  Comparison of PET image denoising methods incorporated with MRI on the abnormal dataset

    帧数文献[5]文献[13]本文方法
    6SNR10.44288.605011.6739
    RMSE1.66142.05291.4419
    12SNR11.009010.076912.2405
    RMSE3.06923.41692.6635
    18SNR11.122911.695512.7017
    RMSE19.700518.443716.4262
    24SNR10.981011.587412.5982
    RMSE20.829519.424917.2910
    下载: 导出CSV

    表  5  单病灶情况下由PET合成图像构图的方法与本文方法比较

    Table  5  Comparison of the methods of constructing graph by composite image with the proposed method on the abnormal dataset

    帧数图滤波图小波本文方法
    6SNR8.745611.363211.6739
    RMSE2.01991.49981.4419
    12SNR10.217211.703612.2405
    RMSE3.36212.83332.6635
    18SNR11.585012.519612.7017
    RMSE18.679716.774216.4262
    24SNR11.484912.409912.5982
    RMSE19.655417.669817.2910
    下载: 导出CSV

    表  6  单病灶PET图像在不同光子数时各种去噪方法比较

    Table  6  Comparison of different denoising methods for abnormal PET images with different photon numbers

    光子数帧数文献[5]文献[13]图滤波图小波本文方法
    $7 \times {10^8}$12SNR11.440311.478911.247411.302811.9611
    RMSE2.92052.90762.98612.96712.7506
    24SNR11.287311.475111.238711.584911.9411
    RMSE20.107819.677720.220619.430518.6497
    $7 \times {10^9}$12SNR11.439811.469011.663011.360911.9734
    RMSE2.92072.91092.84662.94732.7467
    24SNR11.288411.280711.497111.831512.0638
    RMSE20.105320.122919.627818.886618.3881
    下载: 导出CSV
  • [1] Dutta J, Leahy R M, Li Q. Non-local means denoising of dynamic PET images. PloS ONE, 2013, 8(12): e81390 doi: 10.1371/journal.pone.0081390
    [2] Mansoor A, Bagci U, Mollura D J. Optimally stabilized PET image denoising using trilateral filtering. In: Proceedings of International Conference on Medical Image Computing and Computer-Assisted Intervention, Boston, America: IEEE 2014. 130–137.
    [3] Buades A, Coll B, Morel J M. A non-local algorithm for image denoising. In: Proceedings of 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, San Diego, CA, USA: IEEE, 2005. 60–65.
    [4] Jommaa H, Mabrouk R, Khlifa N, Morain-Nicolier, F. Denoising of dynamic PET images using a multi-scale transform and non-local means filter. Biomedical Signal Processing and Control, 2018, 41: 69−80 doi: 10.1016/j.bspc.2017.11.002
    [5] Yan J, Lim J C S, Townsend D W. MRI-guided brain PET image filtering and partial volume correction. Physics in Medicine and Biology, 2015, 60(3): 961−976 doi: 10.1088/0031-9155/60/3/961
    [6] Song T A, Yang F, Chowdhury S R, Kim K, Johnson K A, Fakhri, G A et al. PET image deblurring and super-resolution with an MR-based joint entropy prior. IEEE Transactions on Computational Imaging, 2019, 5(4): 530−539 doi: 10.1109/TCI.2019.2913287
    [7] Song T A, Chowdhury S R, Yang F, Dutta J. PET image super-resolution using generative adversarial networks. Neural Networks, 2020, 125: 83−91 doi: 10.1016/j.neunet.2020.01.029
    [8] Li T, Jiang C H, Gao J, Yang Y F, Liang D, Liu X, et al. Low-count PET image restoration using sparse representation. Nuclear Instruments and Methods in Physics Research Section A Accelerators Spectrometers Detectors and Associated Equipment, 2018, 888: 222−227 doi: 10.1016/j.nima.2018.01.083
    [9] Wang Y, Ma G, An L, Shi F, Zhang P, Lalush D S, et al. Semisupervised tripled dictionary learning for standard-dose PET image prediction using low-dose PET and multimodal MRI. IEEE Transactions on Biomedical Engineering, 2017, 64(3): 569−579 doi: 10.1109/TBME.2016.2564440
    [10] An L, Zhang P, Adeli E, Wang Y, Ma G K, Shi F, et al. Multi-level canonical correlation analysis for standard-dose PET image estimation. IEEE Transactions Image Process, 2016, 25(7): 3303−3315 doi: 10.1109/TIP.2016.2567072
    [11] Zhang Y, Zhang X. PET-MRI joint reconstruction with common edge weighted total variation regularization. Inverse Problems, 2018, 34(6): 065006 doi: 10.1088/1361-6420/aabce9
    [12] Bergounioux M, Papoutsellis E, Stute S, Tauber C. Infimal convolution spatiotemporal PET reconstruction using total variation based priors. 2018, hal-01694064.
    [13] Cheng J C K, Matthews J, Boellaard R, Sossi V. A MR guided denoising for PET using IHYPR-LR. In: Proceedings of IEEE Nuclear Science Symposium and Medical Imaging Conference (NSS/MIC), Atlanta, Georgia: IEEE, 2017. 1–3.
    [14] Tahaei M S, Reader A J, Collins D L. Two novel PET image restoration methods guided by PET-MR kernels: Application to brain imaging. Medical Physics, 2019, 46(5): 2085−2102 doi: 10.1002/mp.13418
    [15] Gong K, Berg E, Cherry S R, Qi J. Machine learning in PET: From photon detection to quantitative image reconstruction. Proceedings of the IEEE, 2020, 108(1): 51−68 doi: 10.1109/JPROC.2019.2936809
    [16] 田娟秀, 刘国才, 谷珊珊, 鞠忠建, 刘劲光, 顾冬冬. 医学图像分析深度学习方法研究与挑战. 自动化学报, 2018, 44(3): 401−424

    Tian Juan-Xiu, Liu Guo-Cai, Gu Shan-Shan, Ju Zhong-Jian, Liu Jin-Guang, Gu Dong-Dong. Deep learning in medical image analysis and its challenges. Acta Automatica Sinica, 2018, 44(3): 401−424
    [17] 施俊, 汪琳琳, 王珊珊, 陈艳霞, 王乾, 魏冬铭, 等. 深度学习在医学影像中的应用综述. 中国图象图形学报, 2020, 25(10): 1953−1981 doi: 10.11834/jig.200255

    Shi Jun, Wang Lin-Lin, Wang Shan-Shan, Chen Yan-Ming, Wang Qian, Wei Dong-Ming et al. Applications of deep learning in medical imaging: A survey. Journal of Image and Graphics, 2020, 25(10): 1953−1981 doi: 10.11834/jig.200255
    [18] 范家伟, 张如如, 陆萌, 何佳雯, 康霄阳, 柴文俊, 等. 深度学习方法在糖尿病视网膜病变诊断中的应用. 自动化学报, 2021, 47(3): 1−20

    Fan Jia-Wei, Zhang Ru-Ru, Lu Meng, He Jia-Wen, Kang Xiao-Yang, Chai Wen-Jun, et al. Applications of deep learning techniques for diabetic retinal diagnosis. Acta Automatica Sinica, 2021, 47(3): 1−20
    [19] Reader A J, Corda G, Mehranian A, Costa-Luis C D, Ellis S, Schnabel J A. Deep learning for PET image reconstruction. IEEE Transactions on Radiation and Plasma Medical Sciences, 2020, 5(1): 1−25
    [20] Gong K, Guan J H, Liu C, Liu C C, Qi J Y. PET image denoising using a deep neural network through fine tuning. IEEE Transactions on Radiation and Plasma Medical Sciences, 2019, 3(2): 153−161 doi: 10.1109/TRPMS.2018.2877644
    [21] Gong Y, Shan H, Teng Y, Zheng H, Wang G, Wang S. Deeply-supervised multi-dose prior learning for low-dose PET imaging. In: Proceedings of IEEE 17th International Symposium on Biomedical Imaging Workshops (ISBI Workshops), Iowa, USA: IEEE, 2020. 1–4.
    [22] Gong Y, Shan H, Teng Y, Zheng H, Wang G, Wang S. Low-dose PET image restoration with 2D and 3D network prior learning. In: Proceedings of IEEE 17th International Symposium on Biomedical Imaging Workshops (ISBI Workshops), Iowa, USA: IEEE, 2020. 1–4.
    [23] Gong Y, Shan H G, Teng Y Y, Tu N, Li M, Liang G D, et al. Parameter-Transferred Wasserstein Generative Adversarial Network (PT-WGAN) for low-dose PET image denoising. IEEE Transactions on Radiation and Plasma Medical Sciences, 2021, 5(2): 213−223 doi: 10.1109/TRPMS.2020.3025071
    [24] Wang Y, Zhou L P, Yu B T, Wang L, Zu C, Lalush D S, et al. 3D auto-context-based locality adaptive multi-modality GANs for PET synthesis. IEEE Transactions on Medical Imaging, 2019, 38(6): 1328−1339 doi: 10.1109/TMI.2018.2884053
    [25] Shuman D I, Narang S K, Frossard P, Ortega A, Vandergheynst P. The emerging field of signal processing on graphs: Extending high-dimensional data analysis to networks and other irregular domains. IEEE Signal Processing Magazine, 2013, 30(3): 83−98 doi: 10.1109/MSP.2012.2235192
    [26] Sandryhaila A F, Moura J M. Big data analysis with signal processing on graphs: Representation and processing of massive data sets with irregular structure. IEEE Signal Processing Magazine, 2014, 31(5): 80−90 doi: 10.1109/MSP.2014.2329213
    [27] 杨杰, 赵磊, 郭文彬. 基于图谱域移位的带限图信号重构算法. 自动化学报, 2021, x(x): 1−11

    Yang Jie, Zhao Lei, Guo Wen-Bin. Graph band-limited signals reconstruction method based graph spectral domain shifting. Acta Automatica Sinica, 2021, x(x): 1−11
    [28] Cheung G, Magli E, Tanaka Y, Ng M K. Graph spectral image processing. Proceedings of the IEEE, 2018, 106(5): 907−930 doi: 10.1109/JPROC.2018.2799702
    [29] Guo S Y, Sheng Y X, Chai L, et al. Graph filtering approach to PET image denoising. In: Proceedings of 1st International Conference on Industrial Artificial Intelligence, Shenyang, China: IEEE, 2019. 1–6.
    [30] Hammond D K, VandergheynstT P, Gribonval R. Wavelets on graphs via spectral graph theory. Applied and Computational Harmonic Analysis, 2011, 30(2): 129−150 doi: 10.1016/j.acha.2010.04.005
    [31] Hammond D K, Vandergheynst P, Gribonval R. The spectral graph wavelet transform: Fundamental theory and fast computation. Vertex-Frequency Analysis of Graph Signals, Springer, Cham, 2019. 141–175.
    [32] Deutsch S, Ortega A, Medioni G. Manifold denoising based on spectral graph wavelets. In: Proceedings of 2016 IEEE International Conference on Acoustics, Speech and Signal Processing, Shanghai, China: IEEE, 2016. 4673–4677.
    [33] Wang G, Qi J. PET image reconstruction using kernel method. IEEE Transactions on Medical Imaging, 2014, 34(1): 61−71
    [34] Shepp L A, Vardi Y. Maximum likelihood reconstruction for positron emission tomography. IEEE Transactions on Medical Imaging, 1982, 1(2): 113−122 doi: 10.1109/TMI.1982.4307558
    [35] Aubert-Broche B, Griffin M, Pike G B, Evans A C, Collins D L. Twenty new digital brain phantoms for creation of validation image databases. IEEE Transactions on Medical Imaging, 2006, 25(11): 1410−1416 doi: 10.1109/TMI.2006.883453
    [36] Feng D, Wong K P, Wu C M, Siu W C. A technique for extracting physiological parameters and the equired input function simultaneously from PET image measurements: Theory and simulation study. IEEE Transactions on Information Technology in Biomedicine, 1997, 1(4): 243−254 doi: 10.1109/4233.681168
  • 加载中
计量
  • 文章访问数:  126
  • HTML全文浏览量:  47
  • 被引次数: 0
出版历程
  • 收稿日期:  2020-12-15
  • 录用日期:  2021-04-29
  • 网络出版日期:  2021-06-13

目录

    /

    返回文章
    返回