张暄轩,曹旭,张久楼,张琳

• 1:西安邮电大学通信与信息工程学院
• 2:西安电子科技大学生命科学技术学院
• 3:南京医科大学医学影像学院
• 4:山东师范大学信息科学与工程学院

摘要(Abstract)：

对生物发光断层成像(Bioluminescence Tomography, BLT)进行研究，提出一种结合有限元方法与深度学习的BLT图像重建方法。该方法使用一种基于有限元网格拓扑结构的神经网络和监督学习实现图像重建，利用的神经网络由一个网格拓扑层和若干全连接层组成，该神经网络通过监督学习方式被训练成一个表面光子密度测量值到被成像物体内部光子密度分布的表达器。在通过神经网络补全获取光子密度分布后，再利用有限元计算获得发光源分布。通过与端到端的全连接神经网络和基于可行域的迭代重建方法对比，实验结果表明，所提方法在500个样本的测试集上的均方差误差显著低于所对比方法，其表现优于直接端到端的全连接网络以及基于可行域的迭代重建方法。

关键词(KeyWords)： 生物发光断层成像;深度学习;有限元方法;全连接网络

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金项目(62001379);; 陕西省自然科学基础研究计划项目(2021JQ-707)

作者(Author): 张暄轩,曹旭,张久楼,张琳

DOI: 10.13682/j.issn.2095-6533.2022.01.011

参考文献(References)：

• [1] DARNE C,LU Y,SEVICK-MURACA E M.Small animal fluorescence and bioluminescence tomography:A review of approaches,algorithms and technology update[J].Physics in Medicine & Biology,2013,59(1):1-64.
• [2] QIN C,FENG J,ZHU S,et al.Recent advances in bioluminescence tomography:Methodology and system as well as application[J].Laser & Photonics Reviews,2014,8(1):94-114.
• [3] SHI J,UDAYAKUMAR T S,XU K,et al.Bioluminescence tomography guided small-animal radiation therapy and tumor response assessment[J].International Journal of Radiation Oncology Biology Physics,2018,102(4):848-857.
• [4] O'NEILL K,LYONS S K,GALLAGHER W M,et al.Bioluminescent imaging:A critical tool in pre-clinical oncology research[J].The Journal of Pathology:A Journal of the Pathological Society of Great Britain and Ireland,2010,220(3):317-327.
• [5] XU T,CLOSE D,HANDAGAMA W,et al.The expanding toolbox of in vivo bioluminescent imaging[J].Frontiers in Oncology,2016,6:150.
• [6] XU S,UDDIN K M S,ZHU Q.Improving DOT reconstruction with a born iterative method and US-guided sparse regularization[J].Biomedical Optics Express,2019,10(5):2528-2541.
• [7] ZHANG X,CAO X,ZHU S.Reconstruction of fluorescence molecular tomography with a cosinoidal level set method[J].Biomedical Engineering Online,2017,16(1):1-17.
• [8] ARRIDGE S R.Optical tomography in medical imaging[J].Inverse Problems,1999,15(2):R41.
• [9] ZHANG B,YIN W,LIU H,et al.Bioluminescence tomography with structural information estimated via statistical mouse atlas registration[J].Biomedical Optics Express,2018,9(8):3544-3558.
• [10] DEHGHANI H,GUGGENHEIM J A,TAYLOR S L,et al.Quantitative bioluminescence tomography using spectral derivative data[J].Biomedical Optics Express,2018,9(9):4163-4174.
• [11] ALLARD M,COTé D,DAVIDSON L,et al.Combined magnetic resonance and bioluminescence imaging of live mice[J].Journal of Biomedical Optics,2007,12(3):034018.1-034018.11.
• [12] ZHANG J,CHEN D,LIANG J,et al.Incorporating MRI structural information into bioluminescence tomography:System,heterogeneous reconstruction and in vivo quantification[J].Biomedical Optics Express,2014,5(6):1861-1876.
• [13] KLOSE A D,BEATTIE B J,DEHGHANI H,et al.In vivo bioluminescence tomography with a blocking-off finite-difference method and MRI/CT coregistration[J].Medical Physics,2010,37(1):329-338.
• [14] NASER M A,PATTERSON M S.Bioluminescence tomography using eigenvectors expansion and iterative solution for the optimized permissible source region[J].Biomedical Optics Express,2011,2(11):3179-3193.
• [15] JAYET B,MORGAN S P,DEHGHANI H.Incorporation of an ultrasound and model guided permissible region improves quantitative source recovery in bioluminescence tomography[J].Biomedical Optics Express,2018,9(3):1360-1374.
• [16] FENG J,JIA K,YAN G,et al.An optimal permissible source region strategy for multispectral bioluminescence tomography[J].Optics Express,2008,16(20):15640-15654.
• [17] LECUN Y,BENGIO Y,HINTON G.Deep learning[J].Nature,2015,521(7553):436-444.
• [18] LITJENS G,KOOI T,BEJNORDI B E,et al.A survey on deep learning in medical image analysis[J].Medical Image Analysis,2017,42:60-88.
• [19] WANG G,YE J C,MUELLER K,et al.Image reconstruction is a new frontier of machine learning[J].IEEE Transactions on Medical Imaging,2018,37(6):1289-1296.
• [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] ZHANG Z,LIANG X,DONG X,et al.A sparse-view CT reconstruction method based on combination of DenseNet and deconvolution[J].IEEE Transactions on Medical Imaging,2018,37(6):1407-1417.
• [22] WüRFL T,HOFFMANN M,CHRISTLEIN V,et al.Deep learning computed tomography:Learning projection-domain weights from image domain in limited angle problems[J].IEEE Transactions on Medical Imaging,2018,37(6):1454-1463.
• [23] QUAN T M,NGUYEN-DUC T,JEONG W K.Compressed sensing MRI reconstruction using a generative adversarial network with a cyclic loss[J].IEEE Transactions on Medical Imaging,2018,37(6):1488-1497.
• [24] ZHU B,LIU J Z,CAULEY S F,et al.Image reconstruction by domain-transform manifold learning[J].Nature,2018,555(7697):487-492.
• [25] YANG B,YING L,TANG J.Artificial neural network enhanced Bayesian PET image reconstruction[J].IEEE Transactions on Medical Imaging,2018,37(6):1297-1309.
• [26] YEDDER H B,BENTAIEB A,SHOKOUFI M,et al.Deep learning based image reconstruction for diffuse optical tomography[C]//Proceedings of the International Workshop on Machine Learning for Medical Image Reconstruction.[S.l]:Springer,2018:112-119.
• [27] GUO L,LIU F,CAI C,et al.3D deep encoder-decoder network for fluorescence molecular tomography[J].Optics Letters,2019,44(8):1892-1895.
• [28] GAO Y,WANG K,AN Y,et al.Nonmodel-based bioluminescence tomography using a machine-learning reconstruction strategy[J].Optica,2018,5(11):1451-1454.
• [29] BOAS D A,CULVER J P,STOTT J J,et al.Three dimensional Monte Carlo code for photon migration through complex heterogeneous media including the adult human head[J].Optics Express,2002,10(3):159-170.
• [30] SCHWEIGER M,ARRIDGE S R,HIRAOKA M,et al.The finite element method for the propagation of light in scattering media:Boundary and source conditions[J].Medical Physics,1995,22(11):1779-1792.
• [31] KINGMA D P,BA J.Adam:A method for stochastic optimization[J/OL].[2021-05-03].https://arxiv.org/pdf/1412.6980.pdf.
• [32] SRIVASTAVA N,HINTON G,KRIZHEVSKY A,et al.Dropout:A simple way to prevent neural networks from overfitting[J].The Journal of Machine Learning Research,2014,15(1):1929-1958.