The Annals of Statistics

Efficient nonparametric Bayesian inference for $X$-ray transforms

François Monard, Richard Nickl, and Gabriel P. Paternain

Full-text: Access denied (no subscription detected)

We're sorry, but we are unable to provide you with the full text of this article because we are not able to identify you as a subscriber. If you have a personal subscription to this journal, then please login. If you are already logged in, then you may need to update your profile to register your subscription. Read more about accessing full-text

Abstract

We consider the statistical inverse problem of recovering a function $f:M\to \mathbb{R}$, where $M$ is a smooth compact Riemannian manifold with boundary, from measurements of general $X$-ray transforms $I_{a}(f)$ of $f$, corrupted by additive Gaussian noise. For $M$ equal to the unit disk with “flat” geometry and $a=0$ this reduces to the standard Radon transform, but our general setting allows for anisotropic media $M$ and can further model local “attenuation” effects—both highly relevant in practical imaging problems such as SPECT tomography. We study a nonparametric Bayesian inference method based on standard Gaussian process priors for $f$. The posterior reconstruction of $f$ corresponds to a Tikhonov regulariser with a reproducing kernel Hilbert space norm penalty that does not require the calculation of the singular value decomposition of the forward operator $I_{a}$. We prove Bernstein–von Mises theorems for a large family of one-dimensional linear functionals of $f$, and they entail that posterior-based inferences such as credible sets are valid and optimal from a frequentist point of view. In particular we derive the asymptotic distribution of smooth linear functionals of the Tikhonov regulariser, which attains the semiparametric information lower bound. The proofs rely on an invertibility result for the “Fisher information” operator $I_{a}^{*}I_{a}$ between suitable function spaces, a result of independent interest that relies on techniques from microlocal analysis. We illustrate the performance of the proposed method via simulations in various settings.

Article information

Source
Ann. Statist., Volume 47, Number 2 (2019), 1113-1147.

Dates
Received: August 2017
Revised: February 2018
First available in Project Euclid: 11 January 2019

Permanent link to this document
https://projecteuclid.org/euclid.aos/1547197250

Digital Object Identifier
doi:10.1214/18-AOS1708

Mathematical Reviews number (MathSciNet)
MR3909962

Zentralblatt MATH identifier
07033163

Subjects
Primary: 62G20: Asymptotic properties
Secondary: 58J40: Pseudodifferential and Fourier integral operators on manifolds [See also 35Sxx] 65R10: Integral transforms 62F15: Bayesian inference

Keywords
Inverse problem Bernstein–von Mises theorem MAP estimate Tikhonov regulariser Gaussian prior Radon transform semiparametric efficiency

Citation

Monard, François; Nickl, Richard; Paternain, Gabriel P. Efficient nonparametric Bayesian inference for $X$-ray transforms. Ann. Statist. 47 (2019), no. 2, 1113--1147. doi:10.1214/18-AOS1708. https://projecteuclid.org/euclid.aos/1547197250


Export citation

References

  • [1] Abramowitz, M. and Stegun, I. A., eds. (1966). Handbook of Mathematical Functions, with Formulas, Graphs, and Mathematical Tables. Dover Publications, New York.
  • [2] Boutet de Monvel, L. (1966). Comportement d’un opérateur pseudo-différentiel sur une variété à bord. II. Pseudo-noyaux de Poisson. J. Anal. Math. 17 255–304.
  • [3] Boutet de Monvel, L. (1971). Boundary problems for pseudo-differential operators. Acta Math. 126 11–51.
  • [4] Brown, D. A., Saibaba, A. and Vallélian, S. (2017). Low rank independence samplers in bayesian inverse problems. ArXiv preprint. Available at arXiv:1609.07180v2.
  • [5] Budinger, T. F., Gullberg, G. T. and Huesman, R. H. (1979). Emission computed tomography. In Image Reconstruction from Projections (G. T. Herman, ed.). Topics in Applied Physics 32 147–246. Springer, Berlin.
  • [6] Castillo, I. (2014). On Bayesian supremum norm contraction rates. Ann. Statist. 42 2058–2091.
  • [7] Castillo, I. (2017). Pólya tree posterior distributions on densities. Ann. Inst. Henri Poincaré Probab. Stat. 53 2074–2102.
  • [8] Castillo, I. and Nickl, R. (2013). Nonparametric Bernstein-von Mises theorems in Gaussian white noise. Ann. Statist. 41 1999–2028.
  • [9] Castillo, I. and Nickl, R. (2014). On the Bernstein-von Mises phenomenon for nonparametric Bayes procedures. Ann. Statist. 42 1941–1969.
  • [10] Castillo, I. and Rousseau, J. (2015). A Bernstein–von Mises theorem for smooth functionals in semiparametric models. Ann. Statist. 43 2353–2383.
  • [11] Cavalier, L. (2008). Nonparametric statistical inverse problems. Inverse Probl. 24 Article ID 034004.
  • [12] Cavalier, L., Golubev, G. K., Picard, D. and Tsybakov, A. B. (2002). Oracle inequalities for inverse problems. Ann. Statist. 30 843–874.
  • [13] Chazarain, J. and Piriou, A. (1982). Introduction to the Theory of Linear Partial Differential Equations. Studies in Mathematics and Its Applications 14. North-Holland, Amsterdam. Translated from the French.
  • [14] Creager, K. C. (1992). Anisotropy of the inner core from differential travel times of the phases PKP and PKIPK. Nature 356 309–314.
  • [15] Dairbekov, N. S., Paternain, G. P., Stefanov, P. and Uhlmann, G. (2007). The boundary rigidity problem in the presence of a magnetic field. Adv. Math. 216 535–609.
  • [16] Dashti, M., Law, K. J. H., Stuart, A. M. and Voss, J. (2013). MAP estimators and their consistency in Bayesian nonparametric inverse problems. Inverse Probl. 29 Article ID 095017.
  • [17] Dashti, M. and Stuart, A. (2016). The Bayesian approach to inverse problems. In Handbook of Uncertainty Quantification (R. Ghanem, D. Higdon and H. Owhadi, eds.) Springer, Berlin.
  • [18] Dudley, R. M. (2002). Real Analysis and Probability. Cambridge Studies in Advanced Mathematics 74. Cambridge Univ. Press, Cambridge. Revised reprint of the 1989 original.
  • [19] Duvall, T. L. Jr, Jeffferies, S. M., Harvey, J. W. and Pomerantz, M. A. (1993). Time–distance helioseismology. Nature 362 430–432.
  • [20] Eskin, G. I. (1981). Boundary Value Problems for Elliptic Pseudodifferential Equations. Translations of Mathematical Monographs 52. Amer. Math. Soc., Providence, RI. Translated from the Russian by S. Smith.
  • [21] Frigyik, B., Stefanov, P. and Uhlmann, G. (2008). The X-ray transform for a generic family of curves and weights. J. Geom. Anal. 18 89–108.
  • [22] Ghosal, S. and van der Vaart, A. (2017). Fundamentals of Nonparametric Bayesian Inference. Cambridge Series in Statistical and Probabilistic Mathematics 44. Cambridge Univ. Press, Cambridge.
  • [23] Giné, E. and Nickl, R. (2016). Mathematical Foundations of Infinite-Dimensional Statistical Models. Cambridge Univ. Press, New York.
  • [24] Goldenshluger, A. and Pereverzev, S. V. (2000). Adaptive estimation of linear functionals in Hilbert scales from indirect white noise observations. Probab. Theory Related Fields 118 169–186.
  • [25] Grubb, G. (2014). Local and nonlocal boundary conditions for $\mu$-transmission and fractional elliptic pseudodifferential operators. Anal. PDE 7 1649–1682.
  • [26] Grubb, G. (2015). Fractional Laplacians on domains, a development of Hörmander’s theory of $\mu$-transmission pseudodifferential operators. Adv. Math. 268 478–528.
  • [27] Guillemin, V. and Sternberg, S. (1977). Geometric Asymptotics. Mathematical Surveys 14. Amer. Math. Soc., Providence, RI.
  • [28] Helgason, S. (1999). The Radon Transform, 2nd ed. Progress in Mathematics 5. Birkhäuser, Boston, MA.
  • [29] Hörmander, L. (1965). Ch II, Boundary Problems for “Classical” Pseudo-Differential Operators. Lectures Notes at IAS, Princeton. Available at http://www.math.ku.dk/~grubb/LH65.pdf.
  • [30] Hörmander, L. (1985). The Analysis of Linear Partial Differential Operators. III. Pseudodifferential operators. Grundlehren der Mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences] 274. Springer, Berlin.
  • [31] Johnstone, I. M. and Silverman, B. W. (1990). Speed of estimation in positron emission tomography and related inverse problems. Ann. Statist. 18 251–280.
  • [32] Kerkyacharian, G., Kyriazis, G., Le Pennec, E., Petrushev, P. and Picard, D. (2010). Inversion of noisy Radon transform by SVD based needlets. Appl. Comput. Harmon. Anal. 28 24–45.
  • [33] Knapik, B. T., van der Vaart, A. W. and van Zanten, J. H. (2011). Bayesian inverse problems with Gaussian priors. Ann. Statist. 39 2626–2657.
  • [34] Kuchment, P. (2014). The Radon Transform and Medical Imaging. CBMS-NSF Regional Conference Series in Applied Mathematics 85. SIAM, Philadelphia, PA.
  • [35] Louis, A. K. (1984). Orthogonal function series expansions and the null space of the Radon transform. SIAM J. Math. Anal. 15 621–633.
  • [36] McLean, W. (2000). Strongly Elliptic Systems and Boundary Integral Equations. Cambridge Univ. Press, Cambridge.
  • [37] Monard, F. (2014). Numerical implementation of geodesic X-ray transforms and their inversion. SIAM J. Imaging Sci. 7 1335–1357.
  • [38] Muhometov, R. G. (1977). The reconstruction problem of a two-dimensional Riemannian metric, and integral geometry. Dokl. Akad. Nauk SSSR 232 32–35.
  • [39] Munk, W. and Wunsch, C. (1979). Ocean acoustic tomography: A scheme for large scale monitoring. Deep-Sea Res. 26A 123–161.
  • [40] Natterer, F. (1986). The Mathematics of Computerized Tomography. Wiley, Chichester.
  • [41] Nickl, R. (2017). Bernstein–von Mises theorems for statistical inverse problems I: Schrödinger equation. ArXiv preprint. Available at arxiv:1707.01764.
  • [42] Nickl, R. and Söhl, J. (2017). Nonparametric Bayesian posterior contraction rates for discretely observed scalar diffusions. Ann. Statist. 45 1664–1693.
  • [43] Novikov, R. G. (2002). An inversion formula for the attenuated X-ray transformation. Ark. Mat. 40 145–167.
  • [44] Paternain, G. P., Salo, M., Uhlmann, G. and Zhou, H. (2018). The geodesic $X$-ray transform with matrix weights. Am. J. Math. To appear. Available at arXiv:1605.07894.
  • [45] Pestov, L. and Uhlmann, G. (2004). On characterization of the range and inversion formulas for the geodesic X-ray transform. Int. Math. Res. Not. 80 4331–4347.
  • [46] Pestov, L. and Uhlmann, G. (2005). Two dimensional compact simple Riemannian manifolds are boundary distance rigid. Ann. of Math. (2) 161 1093–1110.
  • [47] Radon, J. Über die Bestimmung Von Funktionen Durch Ihre Integralwerte Längs Gewisser Mannigfaltigkeiten. Berichte über die Verhandlungen der Königlich-Sächsischen Akademie der Wissenschaften zu Leipzig. Mathematisch-Physische Klasse 69 262–277.
  • [48] Ray, K. (2013). Bayesian inverse problems with non-conjugate priors. Electron. J. Stat. 7 2516–2549.
  • [49] Salo, M. and Uhlmann, G. (2011). The attenuated ray transform on simple surfaces. J. Differential Geom. 88 161–187.
  • [50] Sharafutdinov, V., Skokan, M. and Uhlmann, G. (2005). Regularity of ghosts in tensor tomography. J. Geom. Anal. 15 499–542.
  • [51] Sharafutdinov, V. A. (1994). Integral Geometry of Tensor Fields. VSP, Utrecht.
  • [52] Sharafutdinov, V. A. (2005). Geometric symbol calculus for pseudodifferential operators. I. [Translation of Mat. Tr. 7(2) (2004) 159–206; MR2124544]. Siberian Adv. Math. 15 81–125.
  • [53] Stefanov, P. and Uhlmann, G. (2004). Stability estimates for the X-ray transform of tensor fields and boundary rigidity. Duke Math. J. 123 445–467.
  • [54] Stuart, A. M. (2010). Inverse problems: A Bayesian perspective. Acta Numer. 19 451–559.
  • [55] Taylor, M. E. (1996). Partial Differential Equations. I. Basic Theory. Applied Mathematical Sciences 115. Springer, New York.
  • [56] Thorbergsson, G. (1978). Closed geodesics on non-compact Riemannian manifolds. Math. Z. 159 249–258.
  • [57] Trèves, F. (1980). Introduction to Pseudodifferential and Fourier Integral Operators. Vol. 1. Pseudodifferential Operators. Plenum Press, New York.
  • [58] Uhlmann, G. and Vasy, A. (2016). The inverse problem for the local geodesic ray transform. Invent. Math. 205 83–120.
  • [59] van der Vaart, A. W. (1998). Asymptotic Statistics. Cambridge Series in Statistical and Probabilistic Mathematics 3. Cambridge Univ. Press, Cambridge.
  • [60] van der Vaart, A. W. and van Zanten, J. H. (2008). Rates of contraction of posterior distributions based on Gaussian process priors. Ann. Statist. 36 1435–1463.
  • [61] Višik, M. I. and Èskin, G. I. (1965). Convolution equations in a bounded region. Uspekhi Mat. Nauk 20 89–152.
  • [62] Seppecher’s FEM_beta package. Available at http://math.univ-lyon1.fr/homes-www/seppecher/codes.php.