pith. sign in

arxiv: 1906.10814 · v1 · pith:OZBDXXGYnew · submitted 2019-06-26 · 📡 eess.SP

Inversion of Multi-frequency Data with the Cross-Correlated Contrast Source Inversion Method

Shilong Sun , Bert Jan Kooij , Alexander G. Yarovoy This is my paper

Pith reviewed 2026-05-25 15:41 UTC · model grok-4.3

classification 📡 eess.SP
keywords cross-correlated contrast source inversionmulti-frequency inversioninverse scatteringelectromagnetic wavesTM polarizationTE polarizationMR-CSI
0
0 comments X

The pith

Multi-frequency CC-CSI improves reconstruction of complex scatterers over MR-CSI for both TM and TE cases.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper extends the recently proposed cross-correlated contrast source inversion method to multi-frequency electromagnetic data. In the CC-CSI approach a cross-correlated error term is added to the cost functional by exploiting the mismatch between data error and state error. A multi-frequency version is formulated and tested on both numerical examples and laboratory measurements. For transverse-magnetic and transverse-electric polarizations the results show that CC-CSI continues to produce more accurate images of complicated scatterers than the multiplicative-regularized CSI method once multiple frequencies are available.

Core claim

When multi-frequency data are supplied, the multi-frequency CC-CSI algorithm, which augments the cost functional with a cross-correlated error constructed from the mismatch between data and state errors, yields superior reconstructions of complex scatterers compared with MR-CSI; this advantage is confirmed by both numerical simulations and physical experiments for TM and TE polarizations.

What carries the argument

The cross-correlated error term added to the cost functional, formed by exploiting the mismatch between the data error and the state error during iterative minimization.

If this is right

  • CC-CSI produces more accurate images than MR-CSI of complicated scatterers when multi-frequency data are used.
  • The performance gain holds for both transverse-magnetic and transverse-electric polarizations.
  • The advantage appears in both simulated and measured data sets.
  • The method remains a non-linear iterative inversion technique that can be applied directly to existing multi-frequency measurement setups.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same cross-correlation construction could be inserted into other contrast-source or Born-approximation solvers to test whether the mismatch exploitation is broadly useful.
  • Real-time or limited-aperture imaging applications that already collect multi-frequency data might adopt the approach without new hardware.
  • Three-dimensional extensions would follow the same cost-functional modification once the forward solver is updated.

Load-bearing premise

The cross-correlated error term continues to improve performance when the method is extended from single-frequency to multi-frequency data.

What would settle it

A numerical or experimental test on a complicated scatterer in which multi-frequency CC-CSI produces equal or inferior images to MR-CSI would falsify the claimed advantage.

Figures

Figures reproduced from arXiv: 1906.10814 by Alexander G. Yarovoy, Bert Jan Kooij, Shilong Sun.

Figure 1
Figure 1. Figure 1: Inversion error curves of MF-CC-CSI and MF-MR-CSI in terms of iteration number in case 1 Figure 1: Inversion error curves of MF-CC-CSI and MF-MR-CSI in terms of iteration number [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Inversion error curves of MF-CC-CSI and MF-MR-CSI in terms of iteration number in case 2 Figure 2: Inversion error curves of MF-CC-CSI and MF-MR-CSI in terms of iteration number [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Illustrative example: the 0.1-0.5 GHz TM example in Case 1. Behavior of the cost function, log10  CMF-CC-CSI β1, β2  , in the parametric range −1.5 ≤ β1 ≤ 1.5 and −1.5 ≤ β2 ≤ 1.5. (a) SNR= 30 dB; (b) SNR= 10 dB. 223 224 225 where, x act is the actual solution, while x opt MF-CC-CSI and x opt MF-MR-CSI 226 are the solutions of MF-CC-CSI 227 and MR-MR-CSI, respectively. Now let us consider the 0.1-0.5 GHz … view at source ↗
Figure 4
Figure 4. Figure 4: Illustrative example: the 0.1-0.5 GHz TE example in Case 1. Behavior of the reconstruction error and the cost function (log10 {CMF-CC-CSI}) of MF-CC-CSI. (a) SNR= 30 dB; (b) SNR= 10 dB. 246 247 244 the measurement configuration also affect the cost function curve and the convergence rate of the 245 iterative inversion algorithms. 248 One reasonable strategy in practice is to first estimate the noise level.… view at source ↗
Figure 5
Figure 5. Figure 5: Relative permittivity (left) and conductivity (right) of the inverted contrast by processing the mult Figure 5: Relative permittivity (left) and conductivity (right) of the inverted contrast by pro [PITH_FULL_IMAGE:figures/full_fig_p011_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Relative permittivity (left) and conductivity (right) of the inverted contrast by processing the mult Figure 6: Relative permittivity (left) and conductivity (right) of the inverted contrast by pro [PITH_FULL_IMAGE:figures/full_fig_p012_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Inversion error curves of MF-CC-CSI and MF-MR-CSI in terms of iteration number in processing Figure 7: Inversion error curves of MF-CC-CSI and MF-MR-CSI in terms of iteration number in [PITH_FULL_IMAGE:figures/full_fig_p013_7.png] view at source ↗
read the original abstract

Cross-correlated contrast source inversion (CC-CSI) is a non-linear iterative inversion method that is proposed recently for solving the inverse scattering problems. In CC-CSI, a cross-correlated error is constructed and introduced to the cost functional, which improves the inversion ability when compared to the classical design of the cost functional by exploiting the mismatch between the data error and state error. In this paper, the multi-frequency inversion for electromagnetic waves is considered and a multi-frequency version of CC-CSI is proposed. Numerical and experimental inversion results of both transverse magnetic (TM) and transverse electric (TE) polarization demonstrate that, when multi-frequency data are available, CC-CSI still outperforms the multiplicative-regularized CSI method (MR-CSI) in the inversion of more complicated scatterers.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript extends the recently proposed cross-correlated contrast source inversion (CC-CSI) method to the multi-frequency electromagnetic inverse scattering setting for both TM and TE polarizations. It presents numerical simulations and physical experiments showing that the multi-frequency CC-CSI continues to outperform the multiplicative-regularized CSI (MR-CSI) when reconstructing more complicated scatterers from multi-frequency data.

Significance. If the empirical outperformance holds under closer scrutiny, the work supplies a practical, incremental improvement to an existing nonlinear inversion technique by showing that the cross-correlation term between data and state errors remains beneficial when data from multiple frequencies are available. The combination of simulation and experiment for both polarizations adds modest practical value for the inverse scattering community.

major comments (2)
  1. [Method] Method section (around the definition of the multi-frequency cost functional): the precise construction of the cross-correlated error term when multiple frequencies are combined is not stated explicitly, preventing verification that the single-frequency derivation carries over without additional assumptions or parameters.
  2. [Results] Results section (numerical and experimental comparisons): only qualitative visual comparisons of reconstructed images are shown; no quantitative error norms (e.g., relative L2 reconstruction error or data misfit values) are reported, which is load-bearing for the central claim that CC-CSI “still outperforms” MR-CSI on complicated scatterers.
minor comments (1)
  1. [Abstract] The abstract and introduction could state the number of frequencies and the specific scatterer geometries used in the experiments to allow readers to assess the scope of the claimed improvement.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed and constructive review. The comments highlight two areas where the manuscript can be strengthened for clarity and rigor. We address each point below and will incorporate revisions as indicated.

read point-by-point responses

  1. Referee: [Method] Method section (around the definition of the multi-frequency cost functional): the precise construction of the cross-correlated error term when multiple frequencies are combined is not stated explicitly, preventing verification that the single-frequency derivation carries over without additional assumptions or parameters.

    Authors: We agree that an explicit statement is needed. The multi-frequency cost functional is formed by summing the individual single-frequency data and state misfit terms (including the cross-correlation between data and state errors) over the available frequencies, with the same weighting parameters as in the single-frequency case. No new assumptions or parameters are introduced; the cross-correlation operator is applied frequency-wise before summation. In the revision we will insert the explicit multi-frequency expressions immediately following the single-frequency definitions to allow direct verification. revision: yes

  2. Referee: [Results] Results section (numerical and experimental comparisons): only qualitative visual comparisons of reconstructed images are shown; no quantitative error norms (e.g., relative L2 reconstruction error or data misfit values) are reported, which is load-bearing for the central claim that CC-CSI “still outperforms” MR-CSI on complicated scatterers.

    Authors: The referee correctly notes that quantitative metrics would make the performance comparison more objective. We will add a table (or tables) reporting the relative L2 reconstruction error of the contrast (and conductivity where relevant) for both CC-CSI and MR-CSI across all presented numerical and experimental cases. Data-misfit values at convergence will also be included for completeness. These additions will directly support the claim of continued outperformance. revision: yes

Circularity Check

0 steps flagged

No significant circularity

full rationale

The paper's core contribution is an empirical demonstration that a multi-frequency extension of CC-CSI outperforms MR-CSI on TM/TE data for complex scatterers, supported by separate numerical simulations and experimental measurements. The cost-functional modification is described as a direct generalization of the single-frequency version; its claimed benefit is assessed against an external baseline rather than derived from the same fitted quantities. No equations reduce to self-definition, no predictions are statistically forced by parameter fits, and no load-bearing uniqueness claims or ansatzes are imported via self-citation chains. The derivation chain is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review supplies no explicit free parameters, axioms, or invented entities; the central claim rests on the stated benefit of the cross-correlated error term whose detailed construction is not provided here.

pith-pipeline@v0.9.0 · 5663 in / 1057 out tokens · 41850 ms · 2026-05-25T15:41:15.261353+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

33 extracted references · 33 canonical work pages

  1. [1]

    Colton and R

    D. Colton and R. Kress, Inverse acoustic and electromagnetic scattering theory , vol. 93. New York, USA: Springer, 3 ed., 2013

    work page 2013

  2. [2]

    A computational technique based on a real-coded genetic algorithm for microwave imaging purposes,

    S. Caorsi, A. Massa, and M. Pastorino, “A computational technique based on a real-coded genetic algorithm for microwave imaging purposes,” IEEE Transactions on Geoscience and Remote Sensing, vol. 38, pp. 1697–1708, Jul 2000

    work page 2000

  3. [3]

    Evolutionary op- timization as applied to inverse scattering problems,

    P. Rocca, M. Benedetti, M. Donelli, D. Franceschini, and A. Massa, “Evolutionary op- timization as applied to inverse scattering problems,” Inverse Problems, vol. 25, no. 12, p. 123003 (41pp), 2009

    work page 2009

  4. [4]

    Differential evolution as applied to electromagnetics,

    P. Rocca, G. Oliveri, and A. Massa, “Differential evolution as applied to electromagnetics,” IEEE Antennas and Propagation Magazine , vol. 53, pp. 38–49, Feb 2011

    work page 2011

  5. [5]

    Multifrequency particle swarm optimiza- tion for enhanced multiresolution GPR microwave imaging,

    M. Salucci, L. Poli, N. Anselmi, and A. Massa, “Multifrequency particle swarm optimiza- tion for enhanced multiresolution GPR microwave imaging,” IEEE Transactions on Geo- science and Remote Sensing , vol. 55, no. 3, pp. 1305–1317, 2017

    work page 2017

  6. [6]

    Cakoni, D

    F. Cakoni, D. Colton, and P. Monk, The linear sampling method in inverse electromagnetic scattering. Newark, Delaware, USA: SIAM, 2011

    work page 2011

  7. [7]

    The linear sampling method as a way to quantitative inverse scattering,

    L. Crocco, I. Catapano, L. Di Donato, and T. Isernia, “The linear sampling method as a way to quantitative inverse scattering,” IEEE Transactions on Antennas and Propagation, vol. 60, no. 4, pp. 1844–1853, 2012

    work page 2012

  8. [8]

    An iterative solution of the two-dimensional electromagnetic inverse scattering problem,

    Y. Wang and W. C. Chew, “An iterative solution of the two-dimensional electromagnetic inverse scattering problem,” International Journal of Imaging Systems and Technology , vol. 1, no. 1, pp. 100–108, 1989

    work page 1989

  9. [9]

    A modified gradient method for two-dimensional problems in tomography,

    R. Kleinman and P. van den Berg, “A modified gradient method for two-dimensional problems in tomography,” Journal of Computational and Applied Mathematics , vol. 42, no. 1, pp. 17–35, 1992

    work page 1992

  10. [10]

    An extended range-modified gradient technique for profile inversion,

    R. E. Kleinman and P. van den Berg, “An extended range-modified gradient technique for profile inversion,” Radio Science, vol. 28, no. 05, pp. 877–884, 1993

    work page 1993

  11. [11]

    A contrast source inversion method,

    P. M. van den Berg and R. E. Kleinman, “A contrast source inversion method,” Inverse problems, vol. 13, no. 6, pp. 1607–1620, 1997

    work page 1997

  12. [12]

    Numerical linear algebra for nonlinear microwave imaging,

    F. Di Benedetto, C. Estatico, J. G. Nagy, and M. Pastorino, “Numerical linear algebra for nonlinear microwave imaging,” Electronic Transactions on Numerical Analysis, vol. 33, pp. 105–125, 2009

    work page 2009

  13. [13]

    Extended contrast source inversion,

    P. M. van den Berg, A. Van Broekhoven, and A. Abubakar, “Extended contrast source inversion,” Inverse problems, vol. 15, no. 5, pp. 1325–1344, 1999. 15

    work page 1999

  14. [14]

    Iteratively regularized Gauss-Newton method for nonlinear inverse problems with random noise,

    F. Bauer, T. Hohage, and A. Munk, “Iteratively regularized Gauss-Newton method for nonlinear inverse problems with random noise,” SIAM Journal on Numerical Analysis , vol. 47, no. 3, pp. 1827–1846, 2009

    work page 2009

  15. [15]

    Linearized three-dimensional electromagnetic contrast source inversion and its applications to half-space configurations,

    S. Sun, B. J. Kooij, and A. Yarovoy, “Linearized three-dimensional electromagnetic contrast source inversion and its applications to half-space configurations,” IEEE Transactions on Geoscience and Remote Sensing , vol. 55, pp. 3475–3487, June 2017

    work page 2017

  16. [16]

    A new methodology based on an iterative multiscaling for microwave imaging,

    S. Caorsi, M. Donelli, D. Franceschini, and A. Massa, “A new methodology based on an iterative multiscaling for microwave imaging,” IEEE transactions on microwave theory and techniques, vol. 51, no. 4, pp. 1162–1173, 2003

    work page 2003

  17. [17]

    A contrast source inversion method in the wavelet domain,

    M. Li, O. Semerci, and A. Abubakar, “A contrast source inversion method in the wavelet domain,” Inverse Problems, vol. 29, no. 2, p. 025015, 2013

    work page 2013

  18. [18]

    A compressive sensing data acquisition and imaging method for stepped frequency gprs,

    A. C. Gurbuz, J. H. McClellan, and W. R. Scott, “A compressive sensing data acquisition and imaging method for stepped frequency gprs,” IEEE Transactions on Signal Processing, vol. 57, pp. 2640–2650, July 2009

    work page 2009

  19. [19]

    A linear model for microwave imaging of highly conductive scatterers,

    S. Sun, B. J. Kooij, and A. G. Yarovoy, “A linear model for microwave imaging of highly conductive scatterers,” IEEE Transactions on Microwave Theory and Techniques , vol. 66, no. 3, pp. 1149–1164, 2018

    work page 2018

  20. [20]

    A linear method for shape reconstruction based on the generalized multiple measurement vectors model,

    S. Sun, B. J. Kooij, A. Yarovoy, and T. Jin, “A linear method for shape reconstruction based on the generalized multiple measurement vectors model,” IEEE Transactions on Antennas and Propagation, vol. 66, no. 4, pp. 2016–2025, 2018

    work page 2016

  21. [21]

    A Bayesian-compressive-sampling-based inversion for imaging sparse scatterers,

    G. Oliveri, P. Rocca, and A. Massa, “A Bayesian-compressive-sampling-based inversion for imaging sparse scatterers,” IEEE Transactions on Geoscience and Remote Sensing, vol. 49, no. 10, pp. 3993–4006, 2011

    work page 2011

  22. [22]

    MT–BCS-based microwave imaging approach through minimum-norm current expansion,

    L. Poli, G. Oliveri, F. Viani, and A. Massa, “MT–BCS-based microwave imaging approach through minimum-norm current expansion,” IEEE Transactions on Antennas and Propa- gation, vol. 61, pp. 4722–4732, Sept 2013

    work page 2013

  23. [23]

    A compressive-sensing-based approach for the detection and characterization of buried objects,

    M. Ambrosanio and V. Pascazio, “A compressive-sensing-based approach for the detection and characterization of buried objects,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , vol. 8, pp. 3386–3395, July 2015

    work page 2015

  24. [24]

    Cross-correlated contrast source inversion,

    S. Sun, B. J. Kooij, T. Jin, and A. G. Yarovoy, “Cross-correlated contrast source inversion,” IEEE Transactions on Antennas and Propagation , vol. 65, pp. 2592–2603, May 2017

    work page 2017

  25. [25]

    Inversion of experimental multi-frequency data using the contrast source inversion method,

    R. F. Bloemenkamp, A. Abubakar, and P. M. van den Berg, “Inversion of experimental multi-frequency data using the contrast source inversion method,”Inverse problems, vol. 17, no. 6, pp. 1611–1622, 2001

    work page 2001

  26. [26]

    Free space experimental scattering database continuation: experimental set-up and measurement precision,

    J.-M. Geffrin, P. Sabouroux, and C. Eyraud, “Free space experimental scattering database continuation: experimental set-up and measurement precision,” inverse Problems, vol. 21, no. 6, pp. S117–S130, 2005

    work page 2005

  27. [27]

    R. P. Brent, Algorithms for minimization without derivatives. Englewood Cliffs, New Jersey, 1973

    work page 1973

  28. [28]

    G. E. Forsythe, M. A. Malcolm, and C. B. Moler, Computer Methods for Mathematical Computations. Prentice-Hall, 1976. 16

    work page 1976

  29. [29]

    Reconstruction of a two-dimensional binary obstacle by controlled evolution of a level-set,

    A. Litman, D. Lesselier, and F. Santosa, “Reconstruction of a two-dimensional binary obstacle by controlled evolution of a level-set,” Inverse problems, vol. 14, no. 3, pp. 685– 706, 1998

    work page 1998

  30. [30]

    Contrast source inversion method: state of art,

    P. M. van den Berg and A. Abubakar, “Contrast source inversion method: state of art,” Journal of Electromagnetic Waves and Applications , vol. 15, no. 11, pp. 1503–1505, 2001

    work page 2001

  31. [31]

    Multiplicative regularization for contrast profile inversion,

    P. M. van den Berg, A. Abubakar, and J. T. Fokkema, “Multiplicative regularization for contrast profile inversion,” Radio Science, vol. 38, no. 2, pp. 1–10, 2003

    work page 2003

  32. [32]

    Shin, 3D finite-difference frequency-domain method for plasmonics and nanophotonics

    W. Shin, 3D finite-difference frequency-domain method for plasmonics and nanophotonics . PhD thesis, Stanford University, The Department of Electrical Engineering, USA, 2013

    work page 2013

  33. [33]

    Special section: Testing inversion algorithms against exper- imental data,

    K. Belkebir and M. Saillard, “Special section: Testing inversion algorithms against exper- imental data,” Inverse Problems, vol. 17, no. 6, pp. 1565–1571, 2001. 17

    work page 2001