Provably robust estimation of modulo 1 samples of a smooth function with applications to phase unwrapping

Consider an unknown smooth function $f: [0,1]^d \rightarrow \mathbb{R}$, and say we are given $n$ noisy mod 1 samples of $f$, i.e., $y_i = (f(x_i) + \eta_i)\mod 1$, for $x_i \in [0,1]^d$, where $\eta_i$ denotes the noise. Given the samples $(x_i,y_i)_{i=1}^{n}$, our goal is to recover smooth, robust estimates of the clean samples $f(x_i) \bmod 1$. We formulate a natural approach for solving this problem, which works with angular embeddings of the noisy mod 1 samples over the unit circle, inspired by the angular synchronization framework. This amounts to solving a smoothness regularized least-squares problem -- a quadratically constrained quadratic program (QCQP) -- where the variables are constrained to lie on the unit circle. Our approach is based on solving its relaxation, which is a trust-region sub-problem and hence solvable efficiently. We provide theoretical guarantees demonstrating its robustness to noise for adversarial, and random Gaussian and Bernoulli noise models. To the best of our knowledge, these are the first such theoretical results for this problem. We demonstrate the robustness and efficiency of our approach via extensive numerical simulations on synthetic data, along with a simple least-squares solution for the unwrapping stage, that recovers the original samples of $f$ (up to a global shift). It is shown to perform well at high levels of noise, when taking as input the denoised modulo $1$ samples. Finally, we also consider two other approaches for denoising the modulo 1 samples that leverage tools from Riemannian optimization on manifolds, including a Burer-Monteiro approach for a semidefinite programming relaxation of our formulation. For the two-dimensional version of the problem, which has applications in radar interferometry, we are able to solve instances of real-world data with a million sample points in under 10 seconds, on a personal laptop.

[1]  Jorge J. Moré,et al.  Computing a Trust Region Step , 1983 .

[2]  Nicolas Boumal,et al.  The non-convex Burer-Monteiro approach works on smooth semidefinite programs , 2016, NIPS.

[3]  Bamdev Mishra,et al.  Manopt, a matlab toolbox for optimization on manifolds , 2013, J. Mach. Learn. Res..

[4]  Mariano Rivera,et al.  Half-quadratic cost functions for phase unwrapping. , 2004, Optics letters.

[5]  Mihai Cucuringu,et al.  Synchronization over Z2 and community detection in signed multiplex networks with constraints , 2015, J. Complex Networks.

[6]  Akiko Takeda,et al.  Solving the Trust-Region Subproblem By a Generalized Eigenvalue Problem , 2017, SIAM J. Optim..

[7]  Roman Vershynin,et al.  Introduction to the non-asymptotic analysis of random matrices , 2010, Compressed Sensing.

[8]  Laurent Jacques,et al.  Robust phase unwrapping by convex optimization , 2014, 2014 IEEE International Conference on Image Processing (ICIP).

[9]  Mariano Rivera,et al.  Quadratic regularization functionals for phase unwrapping , 1995 .

[10]  Mario Costantini,et al.  A three-dimensional phase unwrapping algorithm for processing of multitemporal SAR interferometric measurements , 2002, IEEE International Geoscience and Remote Sensing Symposium.

[11]  David P. Towers,et al.  Automatic interferogram analysis techniques applied to quasi-heterodyne holography and ESPI , 1991 .

[12]  Michael Elad,et al.  Image Denoising Via Sparse and Redundant Representations Over Learned Dictionaries , 2006, IEEE Transactions on Image Processing.

[13]  Shuzhong Zhang,et al.  Complex Quadratic Optimization and Semidefinite Programming , 2006, SIAM J. Optim..

[14]  Yonina C. Eldar,et al.  Phase Retrieval via Matrix Completion , 2013, SIAM J. Imaging Sci..

[15]  Todd M. Venema,et al.  Optical phase unwrapping in the presence of branch points. , 2008, Optics express.

[16]  Jaakko Astola,et al.  Absolute phase estimation: adaptive local denoising and global unwrapping. , 2008, Applied optics.

[17]  J. Ohta,et al.  An Implantable CMOS Image Sensor With Self-Reset Pixels for Functional Brain Imaging , 2016, IEEE Transactions on Electron Devices.

[18]  L. Rudin,et al.  Nonlinear total variation based noise removal algorithms , 1992 .

[19]  J. M. Huntley,et al.  Temporal phase-unwrapping algorithm for automated interferogram analysis. , 1993, Applied optics.

[20]  D Kerr,et al.  Unwrapping of interferometric phase-fringe maps by the discrete cosine transform. , 1996, Applied optics.

[21]  D. Sorensen Newton's method with a model trust region modification , 1982 .

[22]  Louis A. Romero,et al.  Minimum Lp-norm two-dimensional phase unwrapping , 1996 .

[23]  Stephen P. Boyd,et al.  Linear Matrix Inequalities in Systems and Control Theory , 1994 .

[24]  Pratibha Mishra,et al.  Advanced Engineering Mathematics , 2013 .

[25]  Mark D. Pritt,et al.  Phase unwrapping by means of multigrid techniques for interferometric SAR , 1996, IEEE Trans. Geosci. Remote. Sens..

[26]  Nicholas I. M. Gould,et al.  Solving the Trust-Region Subproblem using the Lanczos Method , 1999, SIAM J. Optim..

[27]  K Itoh,et al.  Analysis of the phase unwrapping algorithm. , 1982, Applied optics.

[28]  Yonina C. Eldar,et al.  Phase Retrieval: An Overview of Recent Developments , 2015, ArXiv.

[29]  Yonina C. Eldar,et al.  Phase Retrieval via Matrix Completion , 2011, SIAM Rev..

[30]  S. Chavez,et al.  Understanding phase maps in MRI: a new cutline phase unwrapping method , 2002, IEEE Transactions on Medical Imaging.

[31]  E. Kreyszig,et al.  Advanced Engineering Mathematics. , 1974 .

[32]  Danny C. Sorensen,et al.  A New Matrix-Free Algorithm for the Large-Scale Trust-Region Subproblem , 2000, SIAM J. Optim..

[33]  Gabriele Steidl,et al.  Second Order Differences of Cyclic Data and Applications in Variational Denoising , 2014, SIAM J. Imaging Sci..

[34]  M. Rudelson,et al.  Hanson-Wright inequality and sub-gaussian concentration , 2013 .

[35]  J. M. Huntley Noise-immune phase unwrapping algorithm. , 1989, Applied optics.

[36]  Yaron Lipman,et al.  Sensor network localization by eigenvector synchronization over the euclidean group , 2012, TOSN.

[37]  Claudio Prati,et al.  SAR Interferometry: A 2-D Phase Unwrapping Technique Based On Phase And Absolute Values Informations , 1990, 10th Annual International Symposium on Geoscience and Remote Sensing.

[38]  Alexander I. Barvinok,et al.  Problems of distance geometry and convex properties of quadratic maps , 1995, Discret. Comput. Geom..

[39]  A. Singer Angular Synchronization by Eigenvectors and Semidefinite Programming. , 2009, Applied and computational harmonic analysis.

[40]  Y Liu,et al.  Path-independent phase unwrapping using phase gradient and total-variation (TV) denoising. , 2012, Optics express.

[41]  Mihai Cucuringu,et al.  Sync-Rank: Robust Ranking, Constrained Ranking and Rank Aggregation via Eigenvector and Semidefinite Programming Synchronization , 2015, ArXiv.

[42]  William W. Hager,et al.  Minimizing a Quadratic Over a Sphere , 2001, SIAM J. Optim..

[43]  Mitsuo Takeda,et al.  Phase unwrapping by a maximum cross‐amplitude spanning tree algorithm: a comparative study , 1996 .

[44]  William H. Press,et al.  Numerical Recipes 3rd Edition: The Art of Scientific Computing , 2007 .

[45]  Mark D. Pritt,et al.  Least-squares two-dimensional phase unwrapping using FFT's , 1994, IEEE Trans. Geosci. Remote. Sens..

[46]  Andrew Hooper,et al.  Phase unwrapping in three dimensions with application to InSAR time series. , 2007, Journal of the Optical Society of America. A, Optics, image science, and vision.

[47]  Natan T Shaked,et al.  Four dimensional phase unwrapping of dynamic objects in digital holography. , 2018, Optics express.

[48]  José M. Bioucas-Dias,et al.  Phase Unwrapping via Graph Cuts , 2007, IEEE Trans. Image Process..

[49]  Robert E. Mahony,et al.  Optimization Algorithms on Matrix Manifolds , 2007 .

[50]  Y. Y. Hung Shearography for non-destructive evaluation of composite structures , 1996 .

[51]  Peter Schröder,et al.  Multiscale Representations for Manifold-Valued Data , 2005, Multiscale Model. Simul..

[52]  Michael Braun,et al.  Two-dimensional phase unwrapping using a minimum spanning tree algorithm , 1992, IEEE Trans. Image Process..

[53]  D J Bone,et al.  Fourier fringe analysis: the two-dimensional phase unwrapping problem. , 1991, Applied optics.

[54]  Hiroaki Takajo,et al.  Noniterative method for obtaining the exact solution for the normal equation in least-squares phase estimation from the phase difference , 1988 .

[55]  R. Goldstein,et al.  Topographic mapping from interferometric synthetic aperture radar observations , 1986 .

[56]  E. Yaz Linear Matrix Inequalities In System And Control Theory , 1998, Proceedings of the IEEE.

[57]  L. Ying PHASE UNWRAPPING , 2005 .

[58]  P. Lauterbur,et al.  Image Formation by Induced Local Interactions: Examples Employing Nuclear Magnetic Resonance , 1973, Nature.

[59]  Takahiro Yamaguchi,et al.  Implantable self-reset CMOS image sensor and its application to hemodynamic response detection in living mouse brain , 2016 .

[60]  Louis A. Romero,et al.  Robust two-dimensional weighted and unweighted phase unwrapping that uses fast transforms and iterative methods , 1994 .

[61]  Ramesh Raskar,et al.  On unlimited sampling , 2017, 2017 International Conference on Sampling Theory and Applications (SampTA).

[62]  P. Absil,et al.  Erratum to: ``Global rates of convergence for nonconvex optimization on manifolds'' , 2016, IMA Journal of Numerical Analysis.

[63]  Gonçalo Valadão,et al.  CAPE: combinatorial absolute phase estimation. , 2009, Journal of the Optical Society of America. A, Optics, image science, and vision.

[64]  D C Ghiglia,et al.  Direct phase estimation from phase differences using fast elliptic partial differential equation solvers. , 1989, Optics letters.

[65]  Mark Hedley,et al.  A new two‐dimensional phase unwrapping algorithm for MRI images , 1992, Magnetic resonance in medicine.

[66]  Stanley Osher,et al.  Image Recovery via Nonlocal Operators , 2010, J. Sci. Comput..

[67]  Mark Jenkinson,et al.  Fast, automated, N‐dimensional phase‐unwrapping algorithm , 2003, Magnetic resonance in medicine.

[68]  Paul M. Meaney,et al.  The Multidimensional Phase Unwrapping Integral and Applications to Microwave Tomographical Image Reconstruction , 2006, IEEE Transactions on Image Processing.

[69]  Domenica Paoletti,et al.  Manipulation of speckle fringes for non-destructive testing of defects in composites , 1994 .

[70]  R. Pratt,et al.  The application of diffraction tomography to cross-hole seismic data , 1988 .

[71]  Gene H. Golub,et al.  On direct methods for solving Poisson's equation , 1970, Milestones in Matrix Computation.

[72]  Pierre-Antoine Absil,et al.  Trust-Region Methods on Riemannian Manifolds , 2007, Found. Comput. Math..

[73]  Renato D. C. Monteiro,et al.  Digital Object Identifier (DOI) 10.1007/s10107-004-0564-1 , 2004 .

[74]  Hemant Tyagi,et al.  On denoising modulo 1 samples of a function , 2018, AISTATS.

[75]  Franz Rendl,et al.  A semidefinite framework for trust region subproblems with applications to large scale minimization , 1997, Math. Program..

[76]  Ronny Bergmann,et al.  A Second-Order TV-Type Approach for Inpainting and Denoising Higher Dimensional Combined Cyclic and Vector Space Data , 2015, Journal of Mathematical Imaging and Vision.

[77]  Michael Unser,et al.  Isotropic inverse-problem approach for two-dimensional phase unwrapping. , 2015, Journal of the Optical Society of America. A, Optics, image science, and vision.

[78]  Batuhan Osmanoglu,et al.  Three-Dimensional Phase Unwrapping for Satellite Radar Interferometry, I: DEM Generation , 2014, IEEE Transactions on Geoscience and Remote Sensing.

[79]  Abbas El Gamal,et al.  Quantitative study of high-dynamic-range image sensor architectures , 2004, IS&T/SPIE Electronic Imaging.

[80]  H. Takajo,et al.  Least-squares phase estimation from the phase difference , 1988 .

[81]  Youngjoong Joo,et al.  Wide dynamic range CMOS image sensor with pixel level ADC , 2003 .

[82]  Amit Singer,et al.  Eigenvector Synchronization, Graph Rigidity and the Molecule Problem , 2011, Information and inference : a journal of the IMA.

[83]  M. Fiedler Algebraic connectivity of graphs , 1973 .

[84]  L. C. Graham,et al.  Synthetic interferometer radar for topographic mapping , 1974 .

[85]  Bobby R. Hunt,et al.  Matrix formulation of the reconstruction of phase values from phase differences , 1979 .