000596988 001__ 596988 000596988 005__ 20250731015132.0 000596988 0247_ $$2doi$$a10.1007/s42967-023-00333-2 000596988 0247_ $$2ISSN$$a2096-6385 000596988 0247_ $$2ISSN$$a2661-8893 000596988 0247_ $$2datacite_doi$$a10.3204/PUBDB-2023-06365 000596988 0247_ $$2altmetric$$aaltmetric:161535995 000596988 0247_ $$2WOS$$aWOS:001169095400002 000596988 0247_ $$2openalex$$aopenalex:W4392109466 000596988 037__ $$aPUBDB-2023-06365 000596988 041__ $$aEnglish 000596988 082__ $$a510 000596988 1001_ $$0P:(DE-H253)PIP1106483$$aKabri, Samira$$b0$$eCorresponding author 000596988 245__ $$aConvergent Data-driven Regularizations for CT Reconstruction 000596988 260__ $$aSingapore$$bSpringer Singapore$$c2024 000596988 3367_ $$2DRIVER$$aarticle 000596988 3367_ $$2DataCite$$aOutput Types/Journal article 000596988 3367_ $$0PUB:(DE-HGF)16$$2PUB:(DE-HGF)$$aJournal Article$$bjournal$$mjournal$$s1719319394_1933869 000596988 3367_ $$2BibTeX$$aARTICLE 000596988 3367_ $$2ORCID$$aJOURNAL_ARTICLE 000596988 3367_ $$00$$2EndNote$$aJournal Article 000596988 500__ $$aFunding should be my DFG project BU 2327/19/-1 Grundlagen vollüberwachter Deep Learning Verfahren für Inverse ProblemePaper accepted, not yet published (written while SK and MB were still with FAU) 000596988 520__ $$aThe reconstruction of images from their corresponding noisy Radontransform is a typical example of an ill-posed linear inverse problemas arising in the application of computerized tomography (CT). Asthe (naive) solution does not depend on the measured data continu-ously, regularization is needed to re-establish a continuous dependence.In this work, we investigate simple, but yet still provably convergentapproaches to learning linear regularization methods from data. Morespecifically, we analyze two approaches: One generic linear regularizationthat learns how to manipulate the singular values of the linear oper-ator in an extension of our previous work, and one tailored approachin the Fourier domain that is specific to CT-reconstruction. We provethat such approaches become convergent regularization methods as wellas the fact that the reconstructions they provide are typically muchsmoother than the training data they were trained on. Finally, wecompare the spectral as well as the Fourier-based approaches for CT-reconstruction numerically, discuss their advantages and disadvantagesand investigate the effect of discretization errors at different resolutions 000596988 536__ $$0G:(DE-HGF)POF4-623$$a623 - Data Management and Analysis (POF4-623)$$cPOF4-623$$fPOF IV$$x0 000596988 542__ $$2Crossref$$i2024-02-23$$uhttps://creativecommons.org/licenses/by/4.0 000596988 542__ $$2Crossref$$i2024-02-23$$uhttps://creativecommons.org/licenses/by/4.0 000596988 588__ $$aDataset connected to CrossRef, Journals: bib-pubdb1.desy.de 000596988 693__ $$0EXP:(DE-MLZ)NOSPEC-20140101$$5EXP:(DE-MLZ)NOSPEC-20140101$$eNo specific instrument$$x0 000596988 7001_ $$aAuras, Alexander$$b1 000596988 7001_ $$00000-0002-8153-8640$$aRiccio, Danilo$$b2 000596988 7001_ $$aBauermeister, Hartmut$$b3 000596988 7001_ $$aBenning, Martin$$b4 000596988 7001_ $$0P:(DE-H253)PIP1010484$$aMoeller, Michael$$b5 000596988 7001_ $$0P:(DE-H253)PIP1103953$$aBurger, Martin$$b6 000596988 77318 $$2Crossref$$3journal-article$$a10.1007/s42967-023-00333-2$$bSpringer Science and Business Media LLC$$d2024-02-23$$n2$$p1342-1368$$tCommunications on Applied Mathematics and Computation$$v6$$x2096-6385$$y2024 000596988 773__ $$0PERI:(DE-600)2969950-2$$a10.1007/s42967-023-00333-2$$gVol. 6, no. 2, p. 1342 - 1368$$n2$$p1342-1368$$tCommunications on applied mathematics and computation$$v6$$x2096-6385$$y2024 000596988 8564_ $$uhttps://bib-pubdb1.desy.de/record/596988/files/HTML-Approval_of_scientific_publication.html 000596988 8564_ $$uhttps://bib-pubdb1.desy.de/record/596988/files/PDF-Approval_of_scientific_publication.pdf 000596988 8564_ $$uhttps://bib-pubdb1.desy.de/record/596988/files/s42967-023-00333-2.pdf$$yOpenAccess 000596988 8564_ $$uhttps://bib-pubdb1.desy.de/record/596988/files/s42967-023-00333-2.pdf?subformat=pdfa$$xpdfa$$yOpenAccess 000596988 8767_ $$92023-10-30$$d2023-10-30$$eHybrid-OA$$jDEAL$$lSpringerNature$$pCAMC-D-22-00247R1$$v0.00 000596988 8767_ $$92025$$d2025-07-16$$ePayment fee$$jDEAL$$lSpringerNature$$v0.35$$zMPDL Gebühr 000596988 8767_ $$d2025-07-30$$eHybrid-OA$$jStorniert$$lSpringerNature$$zDFG OAPK (Projekt) verrechnet durch V3 000596988 8767_ $$d2025-07-30$$eHybrid-OA$$jZahlung erfolgt$$lSpringerNature$$zDFG OAPK (Projekt) verrechnet durch V3 000596988 909CO $$ooai:bib-pubdb1.desy.de:596988$$popenCost$$pVDB$$pdriver$$pOpenAPC_DEAL$$pdnbdelivery$$pOpenAPC$$popen_access$$popenaire 000596988 9101_ $$0I:(DE-588b)2008985-5$$6P:(DE-H253)PIP1106483$$aDeutsches Elektronen-Synchrotron$$b0$$kDESY 000596988 9101_ $$0I:(DE-HGF)0$$6P:(DE-H253)PIP1010484$$aExternal Institute$$b5$$kExtern 000596988 9101_ $$0I:(DE-588b)2008985-5$$6P:(DE-H253)PIP1103953$$aDeutsches Elektronen-Synchrotron$$b6$$kDESY 000596988 9131_ $$0G:(DE-HGF)POF4-623$$1G:(DE-HGF)POF4-620$$2G:(DE-HGF)POF4-600$$3G:(DE-HGF)POF4$$4G:(DE-HGF)POF$$aDE-HGF$$bForschungsbereich Materie$$lMaterie und Technologie$$vData Management and Analysis$$x0 000596988 9141_ $$y2024 000596988 915__ $$0LIC:(DE-HGF)CCBY4$$2HGFVOC$$aCreative Commons Attribution CC BY 4.0 000596988 915__ $$0StatID:(DE-HGF)3002$$2StatID$$aDEAL Springer$$d2023-10-27$$wger 000596988 915__ $$0StatID:(DE-HGF)0510$$2StatID$$aOpenAccess 000596988 915__ $$0StatID:(DE-HGF)0200$$2StatID$$aDBCoverage$$bSCOPUS$$d2024-12-11 000596988 915__ $$0StatID:(DE-HGF)0300$$2StatID$$aDBCoverage$$bMedline$$d2024-12-11 000596988 915__ $$0StatID:(DE-HGF)0199$$2StatID$$aDBCoverage$$bClarivate Analytics Master Journal List$$d2024-12-11 000596988 915__ $$0StatID:(DE-HGF)0112$$2StatID$$aWoS$$bEmerging Sources Citation Index$$d2024-12-11 000596988 915__ $$0StatID:(DE-HGF)0150$$2StatID$$aDBCoverage$$bWeb of Science Core Collection$$d2024-12-11 000596988 915__ $$0StatID:(DE-HGF)0100$$2StatID$$aJCR$$bCOM APPL MATH COMPUT : 2022$$d2024-12-11 000596988 915__ $$0StatID:(DE-HGF)9900$$2StatID$$aIF < 5$$d2024-12-11 000596988 915pc $$0PC:(DE-HGF)0000$$2APC$$aAPC keys set 000596988 915pc $$0PC:(DE-HGF)0001$$2APC$$aLocal Funding 000596988 915pc $$0PC:(DE-HGF)0002$$2APC$$aDFG OA Publikationskosten 000596988 915pc $$0PC:(DE-HGF)0113$$2APC$$aDEAL: Springer Nature 2020 000596988 9201_ $$0I:(DE-H253)FS-CI-20230420$$kFS-CI$$lComputational Imaging$$x0 000596988 980__ $$ajournal 000596988 980__ $$aVDB 000596988 980__ $$aUNRESTRICTED 000596988 980__ $$aI:(DE-H253)FS-CI-20230420 000596988 980__ $$aAPC 000596988 9801_ $$aAPC 000596988 9801_ $$aFullTexts 000596988 999C5 $$1J Adler$$2Crossref$$9-- missing cx lookup --$$a10.1109/TMI.2018.2799231$$p1322 -$$tIEEE Trans. Med. Imaging$$uAdler, J., Öktem, O.: Learned primal-dual reconstruction. IEEE Trans. Med. Imaging 37(6), 1322–1332 (2018). https://doi.org/10.1109/TMI.2018.2799231$$v37$$y2018 000596988 999C5 $$1M Aharon$$2Crossref$$9-- missing cx lookup --$$a10.1109/TSP.2006.881199$$p4311 -$$tIEEE Trans. Signal Process.$$uAharon, M., Elad, M., Bruckstein, A.: K-SVD: an algorithm for designing overcomplete dictionaries for sparse representation. IEEE Trans. Signal Process. 54(11), 4311–4322 (2006)$$v54$$y2006 000596988 999C5 $$1GS Alberti$$2Crossref$$tAdvances in Neural Information Processing Systems$$uAlberti, G.S., De Vito, E., Lassas, M., Ratti, L., Santacesaria, M.: Learning the optimal Tikhonov regularizer for inverse problems. In: Ranzato, M., Beygelzimer, A., Dauphin, Y., Liang, P.S., Vaughan, J.W. (eds.) Advances in Neural Information Processing Systems, vol. 34, pp. 25205–25216. Curran Associates Inc., New York (2021)$$y2021 000596988 999C5 $$1L Ambrosio$$2Crossref$$9-- missing cx lookup --$$a10.1007/978-3-642-32160-3_1$$p1 -$$uAmbrosio, L., Gigli, N.: A user’s guide to optimal transport. In: Modelling and Optimisation of Flows on Networks, pp. 1–155. Springer, Berlin (2013). https://doi.org/10.1007/978-3-642-32160-3_1$$y2013 000596988 999C5 $$2Crossref$$uAmos, B., Xu, L., Kolter, J.Z.: Input convex neural networks. In: ICML, pp. 146–155. PMLR (2017) 000596988 999C5 $$1S Arridge$$2Crossref$$9-- missing cx lookup --$$a10.1017/S0962492919000059$$p1 -$$tActa Numer.$$uArridge, S., Maass, P., Öktem, O., Schönlieb, C.-B.: Solving inverse problems using data-driven models. Acta Numer. 28, 1–174 (2019)$$v28$$y2019 000596988 999C5 $$1A Aspri$$2Crossref$$9-- missing cx lookup --$$a10.1080/01630563.2020.1740734$$p1190 -$$tNumer. Funct. Anal. Optim.$$uAspri, A., Banert, S., Öktem, O., Scherzer, O.: A data-driven iteratively regularized Landweber iteration. Numer. Funct. Anal. Optim. 41(10), 1190–1227 (2020)$$v41$$y2020 000596988 999C5 $$1DO Baguer$$2Crossref$$9-- missing cx lookup --$$a10.1088/1361-6420/aba415$$tInverse Prob.$$uBaguer, D.O., Leuschner, J., Schmidt, M.: Computed tomography reconstruction using deep image prior and learned reconstruction methods. Inverse Prob. 36(9), 094004 (2020). https://doi.org/10.1088/1361-6420/aba415$$v36$$y2020 000596988 999C5 $$2Crossref$$uBai, S., Kolter, J.Z., Koltun, V. Deep equilibrium models. In: Wallach, H., Larochelle, H., Beygelzimer, A., d'Alché-Buc, F., Fox, E., Garnett, R. (eds.) Advances in Neural Information Processing Systems, vol. 32, Curran Associates, Inc., New York (2019) 000596988 999C5 $$1S Barutcu$$2Crossref$$9-- missing cx lookup --$$a10.1038/s41598-021-97226-2$$p17740 -$$tSci. Rep.$$uBarutcu, S., Aslan, S., Katsaggelos, A.K., Gürsoy, D.: Limited-angle computed tomography with deep image and physics priors. Sci. Rep. 11(1), 17740 (2021). https://doi.org/10.1038/s41598-021-97226-2$$v11$$y2021 000596988 999C5 $$2Crossref$$uBauermeister, H., Burger, M., Moeller, M.: Learning spectral regularizations for linear inverse problems. In: NeurIPS 2020 Workshop on Deep Learning and Inverse Problems (2020) 000596988 999C5 $$1M Benning$$2Crossref$$9-- missing cx lookup --$$a10.1017/S0962492918000016$$p1 -$$tActa Numer.$$uBenning, M., Burger, M.: Modern regularization methods for inverse problems. Acta Numer. 27, 1–111 (2018)$$v27$$y2018 000596988 999C5 $$2Crossref$$uBora, A., Jalal, A., Price, E., Dimakis, A.G.: Compressed sensing using generative models. In: ICML, pp. 537–546. PMLR (2017) 000596988 999C5 $$1Y Chen$$2Crossref$$9-- missing cx lookup --$$a10.1109/TIP.2014.2299065$$p1060 -$$tIEEE Trans. Image Process.$$uChen, Y., Ranftl, R., Pock, T.: Insights into analysis operator learning: from patch-based sparse models to higher order MRFs. IEEE Trans. Image Process. 23(3), 1060–1072 (2014)$$v23$$y2014 000596988 999C5 $$1H Chen$$2Crossref$$9-- missing cx lookup --$$a10.1109/TMI.2018.2805692$$p1333 -$$tIEEE Trans. Med. Imaging$$uChen, H., Zhang, Y., Chen, Y., Zhang, J., Zhang, W., Sun, H., Lyu, Y., Liao, P., Zhou, J., Wang, G.: LEARN: learned experts’ assessment-based reconstruction network for sparse-data CT. IEEE Trans. Med. Imaging 37(6), 1333–1347 (2018)$$v37$$y2018 000596988 999C5 $$1H Chen$$2Crossref$$9-- missing cx lookup --$$a10.1109/TMI.2017.2715284$$p2524 -$$tIEEE Trans. Med. Imaging$$uChen, H., Zhang, Y., Kalra, M.K., Lin, F., Chen, Y., Liao, P., Zhou, J., Wang, G.: Low-dose CT with a residual encoder-decoder convolutional neural network. IEEE Trans. Med. Imaging 36(12), 2524–2535 (2017). https://doi.org/10.1109/TMI.2017.2715284$$v36$$y2017 000596988 999C5 $$1J Chung$$2Crossref$$9-- missing cx lookup --$$a10.1137/100812938$$p3132 -$$tSIAM J. Sci. Comput.$$uChung, J., Chung, M., O’Leary, D.P.: Designing optimal spectral filters for inverse problems. SIAM J. Sci. Comput. 33(6), 3132–3152 (2011)$$v33$$y2011 000596988 999C5 $$1HW Engl$$2Crossref$$9-- missing cx lookup --$$a10.1007/978-94-009-1740-8$$uEngl, H.W., Hanke, M., Neubauer, A.: Regularization of Inverse Problems, vol. 375. Kluwer, Dordrecht (1996)$$y1996 000596988 999C5 $$1J He$$2Crossref$$9-- missing cx lookup --$$a10.1109/TMI.2020.2964266$$p2076 -$$tIEEE Trans. Med. Imaging$$uHe, J., Wang, Y., Ma, J.: Radon inversion via deep learning. IEEE Trans. Med. Imaging 39(6), 2076–2087 (2020)$$v39$$y2020 000596988 999C5 $$1KH Jin$$2Crossref$$9-- missing cx lookup --$$a10.1109/TIP.2017.2713099$$p4509 -$$tIEEE Trans. Image Process.$$uJin, K.H., McCann, M.T., Froustey, E., Unser, M.: Deep convolutional neural network for inverse problems in imaging. IEEE Trans. Image Process. 26(9), 4509–4522 (2017). https://doi.org/10.1109/TIP.2017.2713099$$v26$$y2017 000596988 999C5 $$2Crossref$$9-- missing cx lookup --$$a10.1109/CVPR42600.2020.00757$$uKobler, E., Effland, A., Kunisch, K., Pock, T.: Total deep variation for linear inverse problems. In: CVPR, pp. 7546–7555 (2020) 000596988 999C5 $$2Crossref$$uLatorre, F., Ektekhari, A., Cevher, V. Fast and provable ADMM for learning with generative priors. In: Wallach, H., Larochelle, H., Beygelzimer, A., d'Alché-Buc, F., Fox, E., Garnett, R. (eds.) Advances in Neural Information Processing Systems, vol. 32, Curran Associates, Inc., New York (2019) 000596988 999C5 $$1J Leuschner$$2Crossref$$9-- missing cx lookup --$$a10.1038/s41597-021-00893-z$$p109 -$$tSci. Data$$uLeuschner, J., Schmidt, M., Baguer, D.O., Maass, P.: LoDoPaB-CT, a benchmark dataset for low-dose computed tomography reconstruction. Sci. Data. 8, 109 (2021). https://doi.org/10.1038/s41597-021-00893-z$$v8$$y2021 000596988 999C5 $$1J Leuschner$$2Crossref$$9-- missing cx lookup --$$a10.3390/jimaging7030044$$p44 -$$tJ. Imaging$$uLeuschner, J., Schmidt, M., Ganguly, P.S., Andriiashen, V., Coban, S.B., Denker, A., Bauer, D., Hadjifaradji, A., Batenburg, K.J., Maass, P., van Eijnatten, M.: Quantitative comparison of deep learning-based image reconstruction methods for low-dose and sparse-angle CT applications. J. Imaging 7(3), 44 (2021)$$v7$$y2021 000596988 999C5 $$1H Li$$2Crossref$$9-- missing cx lookup --$$a10.1088/1361-6420/ab6d57$$tInverse Prob.$$uLi, H., Schwab, J., Antholzer, S., Haltmeier, M.: NETT: solving inverse problems with deep neural networks. Inverse Prob. 36(6), 065005 (2020)$$v36$$y2020 000596988 999C5 $$1Y Li$$2Crossref$$9-- missing cx lookup --$$a10.1109/TMI.2019.2910760$$p2469 -$$tIEEE Trans. Med. Imaging$$uLi, Y., Li, K., Zhang, C., Montoya, J., Chen, G.-H.: Learning to reconstruct computed tomography images directly from sinogram data under a variety of data acquisition conditions. IEEE Trans. Med. Imaging 38(10), 2469–2481 (2019). https://doi.org/10.1109/TMI.2019.2910760$$v38$$y2019 000596988 999C5 $$2Crossref$$uMairal, J., Ponce, J., Sapiro, G., Zisserman, A., Bach, F.: Supervised dictionary learning. In: NeurIPS (2008) 000596988 999C5 $$2Crossref$$9-- missing cx lookup --$$a10.1109/ICCV.2017.198$$uMeinhardt, T., Moeller, M., Hazirbas, C., Cremers, D.: Learning proximal operators: using denoising networks for regularizing inverse imaging problems. In: ICCV, pp. 1781–1790 (2017) 000596988 999C5 $$2Crossref$$9-- missing cx lookup --$$a10.1109/ICCV.2019.00335$$uMoeller, M., Mollenhoff, T., Cremers, D.: Controlling neural networks via energy dissipation. In: ICCV, pp. 3256–3265 (2019) 000596988 999C5 $$2Crossref$$uRiccio, D., Ehrhardt, M.J., Benning, M.: Regularization of inverse problems: deep equilibrium models versus bilevel learning. arXiv:2206.13193 (2022) 000596988 999C5 $$2Crossref$$9-- missing cx lookup --$$a10.1109/ICCV.2017.627$$uRick Chang, J., Li, C.-L., Poczos, B., Vijaya Kumar, B., Sankaranarayanan, A.C.: One network to solve them all—solving linear inverse problems using deep projection models. In: ICCV, pp. 5888–5897 (2017) 000596988 999C5 $$1Y Romano$$2Crossref$$9-- missing cx lookup --$$a10.1137/16M1102884$$p1804 -$$tSIAM J. Imaging Sci.$$uRomano, Y., Elad, M., Milanfar, P.: The little engine that could: regularization by denoising (RED). SIAM J. Imaging Sci. 10(4), 1804–1844 (2017)$$v10$$y2017 000596988 999C5 $$1O Ronneberger$$2Crossref$$tMedical Image Computing and Computer-Assisted Intervention—MICCAI 2015$$uRonneberger, O., Fischer, P., Brox, T.: U-net: convolutional networks for biomedical image segmentation. In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F. (eds.) Medical Image Computing and Computer-Assisted Intervention—MICCAI 2015, pp. 234–241. Springer, Cham (2015)$$y2015 000596988 999C5 $$1S Roth$$2Crossref$$9-- missing cx lookup --$$a10.1007/s11263-008-0197-6$$p205 -$$tInt. J. Comput. Vis.$$uRoth, S., Black, M.J.: Fields of experts. Int. J. Comput. Vis. 82(2), 205–229 (2009)$$v82$$y2009 000596988 999C5 $$2Crossref$$9-- missing cx lookup --$$a10.1117/12.533472$$uServieres, M.C.J., Normand, N., Subirats, P., Guedon, J.: Some links between continuous and discrete Radon transform. In: Fitzpatrick, J.M., Sonka, M. (eds.) Medical Imaging 2004: Image Processing, vol. 5370, pp. 1961–1971. SPIE, WA, United States. International Society for Optics and Photonics (2004). https://doi.org/10.1117/12.533472 000596988 999C5 $$2Crossref$$9-- missing cx lookup --$$a10.1109/CVPR.2018.00984$$uUlyanov, D., Vedaldi, A., Lempitsky, V.: Deep image prior. In: CVPR, pp. 9446–9454 (2018) 000596988 999C5 $$1G Wang$$2Crossref$$9-- missing cx lookup --$$a10.1038/s42256-020-00273-z$$p737 -$$tNat. Mach. Intell.$$uWang, G., Ye, J.C., De Man, B.: Deep learning for tomographic image reconstruction. Nat. Mach. Intell. 2(12), 737–748 (2020). https://doi.org/10.1038/s42256-020-00273-z$$v2$$y2020 000596988 999C5 $$1J Xiang$$2Crossref$$9-- missing cx lookup --$$a10.1109/TMI.2021.3054167$$p1329 -$$tIEEE Trans. Med. Imaging$$uXiang, J., Dong, Y., Yang, Y.: FISTA-Net: learning a fast iterative shrinkage thresholding network for inverse problems in imaging. IEEE Trans. Med. Imaging 40(5), 1329–1339 (2021). https://doi.org/10.1109/TMI.2021.3054167$$v40$$y2021 000596988 999C5 $$1M Xu$$2Crossref$$9-- missing cx lookup --$$a10.1109/TRPMS.2020.2991887$$p78 -$$tIEEE Trans. Radiat. Plasma Med. Sci.$$uXu, M., Hu, D., Luo, F., Liu, F., Wang, S., Wu, W.: Limited-angle x-ray CT reconstruction using image gradient $$\ell _0$$-norm with dictionary learning. IEEE Trans. Radiat. Plasma Med. Sci. 5(1), 78–87 (2021). https://doi.org/10.1109/TRPMS.2020.2991887$$v5$$y2021 000596988 999C5 $$1M Zhang$$2Crossref$$9-- missing cx lookup --$$a10.1007/s40747-022-00724-7$$p5545 -$$tComplex Intell. Syst.$$uZhang, M., Gu, S., Shi, Y.: The use of deep learning methods in low-dose computed tomography image reconstruction: a systematic review. Complex Intell. Syst. 8, 5545–5561 (2022). https://doi.org/10.1007/s40747-022-00724-7$$v8$$y2022 000596988 999C5 $$1B Zhu$$2Crossref$$9-- missing cx lookup --$$a10.1038/nature25988$$p487 -$$tNature$$uZhu, B., Liu, J.Z., Cauley, S.F., Rosen, B.R., Rosen, M.S.: Image reconstruction by domain-transform manifold learning. Nature 555(7697), 487–492 (2018). https://doi.org/10.1038/nature25988$$v555$$y2018