Pith. sign in

REVIEW 4 major objections 5 minor 59 references

Harnessing modal fields retrieved from speckle for multi-dimensional metrology

T0 review · 4 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper shows that compressing a speckle image into 2N-1 fiber-mode coefficients makes speckle-based metrology train 800 times faster without sacrificing accuracy.

desk verdict The core idea — train on modal coefficients instead of raw speckle — is plausible and the 5x/800x training claims are striking, but the paper overstates the physics grounding and underreports the mode decomposition cost in inference. read the letter →

arxiv 2509.03976 v1 pith:RMZAB23I submitted 2025-09-04 physics.optics

Qingbo Liu , Zhongyang Xu , Guangkui Tao , Xiuyuan Sun , Min Xue , Weihao Yuan , Shilong Pan This is my paper · ORCID
classification physics.optics
keywords specklemetrologyfew-modefibermodedecompositionmodalfieldsmachinelearningfiber-opticsensingtactilemulti-dimensional
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper's central claim is that the main bottleneck in speckle-based optical metrology — large training sets and long training times — comes from treating the speckle pattern as raw pixels, and that compressing each specklegram into the modal field coefficients of a few-mode fiber removes that bottleneck. It proposes a two-stage anti-noise fast mode decomposition: an SPGD calibration stage builds a mode-to-speckle matrix for the noisy environment, followed by fast matrix inversion and BFGS refinement that recovers the amplitudes and relative phases of the fiber's LP modes. Feeding these modal coefficients to a LightGBM regressor, rather than 256x256 pixel specklegrams to a CNN, yields high-accuracy estimation of fiber curvature, bending position, bending angle, and torsion, and enables 2D tactile pattern reconstruction. On the reported experiments the approach needs about five times fewer training samples and 800 times less training time (9 h 45 min to 40 s). If the decomposition is physically faithful, this makes speckle sensing substantially cheaper to calibrate and deploy, and opens the same sensor to simultaneous multi-parameter measurement.

What carries the argument

The central object is the modal-coefficient vector of length 2N-1: the normalized amplitudes and relative phases of the N linearly polarized (LP) modes making up the field in the few-mode fiber. The key mechanism is the two-stage anti-noise fast mode decomposition. In a preparatory stage, many specklegrams are decomposed with stochastic parallel gradient descent (SPGD) to construct an eigenmode matrix H_z that accounts for the actual noisy environment. In the running stage, each new specklegram is converted to coefficients by the fast inverse matrix solution X=(H_z)^-1 I, and the initial values are refined with the BFGS optimizer. This reduces a 256x256 pixel intensity pattern to at most 2N-

What would settle it

Measure the same fiber states with a phase-sensitive reference technique that reports the true modal amplitudes and phases, and compare those against the SPGD-BFGS outputs. If the two disagree while the retrieved coefficients still reconstruct the recorded speckle, the claimed accuracy rests on a self-consistent but physically wrong representation, and a model trained on it should fail on unseen optical states.

Watch

Extended reading notes

Core claim

On its own terms, the discovery is that a few-mode fiber's speckle pattern is better represented for machine learning by the underlying 2N-1 modal coefficients (amplitudes and relative phases of its LP eigenmodes) than by raw pixels. These coefficients are retrieved reliably under noise by a two-stage decomposition: an SPGD calibration builds an eigenmode matrix for the noisy environment, then fast matrix inversion plus BFGS refinement solves each specklegram. Trained on these coefficients, a LightGBM model estimates curvature, bending position, bending angle, and torsion with R² values from 0.9933 to 0.9999, reconstructs 2D tactile patterns, and cuts training data about 5x and training time

Load-bearing premise

The load-bearing premise is that the routine that pulls the mode amplitudes and phases out of a speckle image is recovering the true physical state of the light, not just a convenient set of numbers that can re-create the measured speckle; if the calibration pattern library is skewed by noise, every downstream estimate inherits that skew, and matching the measured speckle does not prove the mode weights are right.

Editorial extensions

If this is right

  • A speckle-based fiber sensor can be recalibrated for a new environment in about 40 seconds on ordinary computing hardware rather than roughly ten hours.
  • The same sensor can estimate multiple physical parameters simultaneously — curvature, bending position, bending angle, and torsion — with a mild accuracy trade-off as the parameter count increases.
  • Because modal fields are continuous, the model can interpolate between discrete calibration points, yielding 3x to 7x better resolutions than classification-style speckle sensors in the reported comparisons.
  • The expensive SPGD calibration is paid once; afterward, mode decomposition takes about 0.42 s per sample, so the front end can be reused across sensing tasks without retraining.
  • Two-dimensional tactile imaging works without prior knowledge of the written pattern by classifying the 10x10 pressed positions from modal weights.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A direct cross-validation against phase-resolved reference mode decomposition would separate the gain from true physical information from the gain from a convenient learned coordinate system; the paper's 0.985 correlation is with the measured speckle, not with ground-truth mode weights.
  • The current sensor uses only modal amplitudes for regression, so the N-1 relative phases are an untapped channel that could improve accuracy or add sensing axes.
  • The same decomposition-plus-regression recipe should carry over to other fiber types and to scattering media that admit a low-dimensional modal basis, with decomposition noise setting the practical ceiling.
  • Because the dual-parameter experiment already shows crosstalk, scaling to three or more simultaneous parameters will require either more modes or a model that explicitly separates modal responses.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes a machine-learning pipeline for speckle-based fiber metrology in which pixel-level specklegrams are replaced by 2N−1 modal coefficients (amplitudes and relative phases) recovered by a two-stage anti-noise mode-decomposition method. An SPGD-based preparatory stage builds an eigenmode matrix Hz, and each specklegram is then decomposed by matrix inversion followed by BFGS refinement. The resulting coefficients are used to train a LightGBM regressor for estimating fiber curvature, bending position, bending angle, and torsion angle, and for 2D tactile-pattern reconstruction. The authors report high R² values (0.9987 for curvature, up to 0.9999 for position), a 5× reduction in training data, an 800× reduction in training time, and a 100× reduction in inference time relative to a CNN trained directly on specklegrams.

Significance. If the claims held as stated, the work would offer a practical and low-cost route to multi-dimensional fiber sensing by compressing high-dimensional speckle data into a small set of physically motivated features. The demonstration of tactile pattern reconstruction and simultaneous dual-parameter estimation is interesting, and the authors provide a reasonably detailed experimental description. However, the central premise—that the retrieved coefficients are true physical LP-mode weights—is not validated against any ground truth. The performance comparisons with the CNN baseline are incomplete and partly misleading (the reported inference and training times exclude the mode-decomposition cost). The paper does not ship code, data, or machine-checked proofs, so reproducibility rests on the experimental description alone. The practical idea is plausible, but the current evidence is insufficient to support the strong physics-informed claims.

major comments (4)
  1. [§2.2, Eqs. (2)–(4) and Fig. 2] The decomposition accuracy is validated only by the correlation between measured and reconstructed speckle intensities (0.985). This is a self-consistency check: it shows the coefficients reproduce the measured intensity, not that they equal the true physical mode weights. Since Hz is calibrated from SPGD in a noisy environment, any SPGD bias is inherited by the fast decomposition. The paper needs an independent validation of the modal coefficients, e.g., using mode-selective excitation, an interferometric measurement, or synthetic specklegrams with known coefficients under matched noise statistics. Without this, the claim that the features are physical modal fields—and the associated advantages of continuity, predictability, and low crosstalk—is not established.
  2. [Table 2 and §2.3] The inference and training-time comparisons are misleading. The 0.015 ms inference time for the modal-weight scheme excludes the 0.42 s per-speckle mode decomposition; including that cost, end-to-end inference is about 0.42 s, which is far slower than the 1.4 ms CNN inference. Likewise, the '40 s training time' excludes the mode-decomposition preprocessing; the text later says the total time is 'around 20 minutes', but 4,200 samples at 0.42 s/sample is 29.4 min, not 20 min. The 800×/100× speed-up claims need to be recomputed consistently, and the decomposition time should be included or explicitly separated in the comparisons.
  3. [§2.3, text near Fig. 4] The manuscript states 'we utilize only the amplitude information, which is the modal weight ρ2, for sensing.' This conflicts with the surrounding text and figures, which use multiple modal weights (Fig. 4a, Fig. 7, Fig. 8a) and emphasize diverse responses of different modes. If only ρ2 is used, the multi-mode diversity argument collapses; if all modal weights are used, the sentence is an error. This must be corrected and the exact feature set used in each experiment must be stated unambiguously.
  4. [§2.3, Table 1 and CNN comparison] The CNN baseline's accuracy is never reported. The claim that 'our approach based on modal fields achieves comparable estimation accuracy with only 4,200 samples' requires the same metrics (R², RMSE, error distributions) for the CNN trained on the same dataset, same splits, and same platform. Comparing against literature values (Table 1) is not a controlled comparison because datasets, fiber types, and tasks differ. Without this baseline, the 5× data-reduction claim is unsubstantiated.
minor comments (5)
  1. [§2.2, Eqs. (2)–(4)] The equations are poorly rendered and hard to parse. The definitions of X, the eigenmode matrix H, and the sign convention in Eq. (4) should be rewritten clearly; in particular, the expression for X is garbled and the role of Hz versus H is not explicitly defined.
  2. [Fig. 2 caption] The abbreviations MEA, REC, DIS, COR are used in the caption but not defined in the main text. Please define them or remove the unexplained acronyms.
  3. [Abstract and §2.3] The abstract states 'training time ... reduced by 800 times, from 9 h 45 min to 40 s', while §2.3 says the total time (including mode decomposition) is 'around 20 minutes, reduced by 30 times'. These two statements should be reconciled, and the definition of 'training time' should be spelled out consistently.
  4. [Table 1] The column headers mix 'Resolution' with standard deviations of estimation errors. For a fair comparison, specify how resolution is defined in each cited work and whether the reported values are standard deviations, mean absolute errors, or quantization steps.
  5. [Throughout] The symbol 'R²' is sometimes written as 'R2' or 'R^2'. Also, no error bars or repeated-measurements statistics are given for the R² and RMSE values; please report the number of independent trials and variability.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the metrology targets are external labels, and the modal coefficients are features derived from measured intensity, not from the targets.

full rationale

The claimed derivation chain is: speckle intensity I is mapped to 2N-1 modal coefficients through Eq. (2)-(4) plus BFGS refinement; the coefficients (specifically rho_2) are then used as inputs to a LightGBM regressor whose outputs are experimentally controlled parameters (curvature, bending position, bending angle, torsion). The target labels are set by translation stages/rotators and are never used to define the decomposition or to construct the feature vector. Thus no reported R^2 or RMSE reduces to a fitted value by construction. The only internal consistency check reported for the decomposition is the correlation between measured and reconstructed specklegrams (0.985, Fig. 2a), which validates intensity reconstruction rather than ground-truth mode weights; this is an evidence-strength limitation, not a circular reduction. The paper contains one self-citation (Ref. [5]) in a general wearable-device context, but it is not load-bearing. Even if 'physics-informed' overstates the role of physical laws in the ML step, that is a framing issue rather than a circular derivation. Consequently, no self-definitional, fitted-input-renamed-as-prediction, uniqueness-importation, or ansatz-by-citation circularity is present.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The paper introduces no new physical entities. It relies on standard few-mode fiber physics, an empirically calibrated eigenmode matrix, and learned regression weights. The main assumed inputs are the validity of the mode decomposition under noise and the generalizability of the ML regressor.

free parameters (2)
  • Eigenmode matrix H_z (calibration) = not reported
    Built from SPGD decompositions of noisy specklegrams in the preparatory stage; it is an empirical calibration, not a closed-form expression, and is used in X=(H_z)^-1 I.
  • LightGBM model weights = not reported
    The regression maps modal weights to physical parameters; no equation is derived, the mapping is learned from data.
assumptions (5)
  • domain assumption The optical field in the FMF is a linear superposition of a small number of LP eigenmodes (Eq. 1).
    Invoked in Eqs. 1-4 to represent the field and to define the modal coefficients used as ML features.
  • domain assumption The measured speckle intensity equals the squared modulus of that superposition with negligible polarization or mode-coupling effects.
    The mode decomposition in Section 2.2 assumes the eigenmode matrix H relates intensity to the amplitude-phase vector X through Eq. 2.
  • ad hoc to paper SPGD decompositions in the preparatory stage are accurate enough to calibrate H_z.
    Section 2.2: 'multiple specklegrams are collected and decomposed using the classic SPGD algorithm. The calculated modal coefficients are used to obtain the eigenmode matrix (Hz)'. No ground-truth validation is given in the main text.
  • domain assumption BFGS optimization converges to the physically correct mode coefficients.
    The refinement step in Section 2.2 and Experimental Section assumes the optimization landscape is well-behaved; non-convex phase retrieval can have multiple intensity-equivalent solutions.
  • domain assumption LightGBM can generalize from modal weights to continuous physical parameters.
    The regression is the core of the sensing; the paper relies on standard ML generalization rather than a physical model of the parameter-to-mode mapping.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Harnessing modal fields retrieved from speckle for multi-dimensional metrology." pith.science (2026). https://pith.science/paper/RMZAB23I

@misc{pith2026250903976,
  author       = {Pith},
  title        = {Pith review of: Harnessing modal fields retrieved from speckle for multi-dimensional metrology},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RMZAB23I}},
  note         = {Machine review of arXiv:2509.03976}
}
read the original abstract

Although speckle is a powerful tool for high-precision metrology, large datasets and cumbersome training are always required to learn from the encoded speckle patterns, which is unfavorable for rapid deployment and multi-dimensional metrology. To enable high accuracy and fast training, physics-informed machine learning enforces physical laws to address high-dimensional problems. Here, we harness the modal fields in a few-mode fiber, which follow the law of beam propagation, to enable high-accuracy and fast-training parameter estimation. Anti-noise fast mode decomposition is implemented to retrieve the modal fields from the speckles. The accuracy is enhanced since the modal fields enable parameter estimation at random points in the continuous space-time domain. Artificial tactile perception and multi-dimensional metrology are achieved with high accuracy because the modal fields respond diversely to different parameters. Meanwhile, the number of specklegrams for training is reduced by around 5 times. The training time of machine learning is significantly reduced by 800 times, from 9 hours and 45 minutes to 40 seconds. Therefore, harnessing the modal fields paves a new way for the speckle-based metrology to develop efficient, low-cost, multi-dimensional sensors, making it suitable for intelligent wearable devices, industrial robots and healthcare applications.

Figures

Figures reproduced from arXiv: 2509.03976 by the authors.

Figure 1
Figure 1. Schematic illustration of the speckle-based metrology and sensing using an FMF. The conventional approach is data-driven, while the proposed approach is informed by modal fields. The concept of speckle-based metrology and sensing using an FMF is illustrated in [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. illustrates the experimental evaluation of IMS, SPGD and our method. For 3,500 samples, the average correlations between the measured and the reconstructed specklegrams are 0.801, 0.979 and 0.985, respectively, as shown in Figure 2a. The standard deviations are 0.029, 0.006 and 0.004, respectively. Moreover, Figure 2b illustrates the performance at different SNRs. The increase in noise leads to large errors and a de… view at source ↗
Figure 3
Figure 3. The experimental system and workflow for mode-informed metrology and sensing. VOA: variable optical attenuator; PC: polarization controller; PL: photonic lantern; L: lens. In Figure 4c, the root mean square error (RMSE) on the validation set decreases to 0.032 m⁻¹ after 1,000 training iterations, close to that of the training set. It indicates that the model is effectively trained and possesses strong generalization… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: d. The R² value for the estimated results is 0.9987. Figure 4e illustrates the estimation of step-change curvature, in which the standard deviation of the estimation errors is 0.033 m⁻¹. Moreover, multiple parameters, including bending position, bending angle and torsi…
Figure 5
Figure 5. Figure 5: Bending position, bending angle and torsion angle sensing informed by modal fields. (a)–(c) Relationship between the estimated values and the actual values. (d)–(f) Distribution of estimated parameters for step changes. (g)–(i) Experimental setups for changing the bend…
Figure 6
Figure 6. Figure 6: Tactile sensing informed by modal fields. (a) Experimental setup for tactile sensation. (b) Confusion matrix for the predictions of the 10 × 10 pixel positions. (c) Actual images (TRU) of four different tactile patterns and corresponding reconstructed results (REC). Si…
Figure 7
Figure 7. Figure 7: Importance of multiple modes to fiber curvature, bending position, bending angle and torsion angle sensing in a six-mode optical fiber [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: Dual-parameter estimation with an FMF. (a) Evolution of modal weights (LP02, LP11o and LP21e) with the bending position D and bending angle θ1. (b) Simultaneous dual-parameter estimation. Another attractive feature of the mode-informed speckle sensor is that different …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 59 canonical work pages

  1. [1]

    Lalam, S

    N. Lalam, S. Bukka, H. Bhatta, M. Buric, P. Ohodnicki, R. Wright, Commun. Eng. 2024, 3, 121

    work page 2024

  2. [2]

    M. Chen, Z. Wang, Q. Zhang, Z. Wang, W. Liu, M. Chen, L. Wei, Nat. Commun. 2021, 12, 1416

    work page 2021

  3. [3]

    R. Wang, R. Wang, C. Dou, S. Yang, R. Gnanasambandam, A. Wang, Z. Kong, Nat. Commun. 2024, 15, 7568

    work page 2024

  4. [4]

    Y . Li, Y . Su, Light Sci. Appl. 2024, 13, 79

    work page 2024

  5. [5]

    P. Li, Y . Wang, X. He, Y . Cui, J. Ouyang, J. Ouyang, Z. He, J. Hu, X. Liu, H. Wei, Y . Wang, X. Lu, Q. Ji, X. Cai, L. Liu, C. Hou, N. Zhou, S. Pan, X. Wang, H. Zhou, C.-W. Qiu, Y .-Q. Lu, G. Tao, Light Sci. Appl. 2024, 13, 48

    work page 2024

  6. [6]

    J. Guo, M. Niu, C. Yang, Optica 2017, 4, 1285

    work page 2017

  7. [7]

    Leal-Junior, L

    A. Leal-Junior, L. Avellar, A. Frizera, C. Marques, Sci. Rep. 2020, 10, 13867

    work page 2020

  8. [8]

    Y . Li, Z. Guo, X. Zhao, S. Liu, Z. Chen, W.-F. Dong, S. Wang, Y .-L. Sun, X. Wu, Nat. Commun. 2024, 15, 2906

    work page 2024

Show all 59 references
  1. [9]

    Z. Peng, H. Wen, J. Jian, A. Gribok, M. Wang, S. Huang, H. Liu, Z.-H. Mao, K. P. Chen, Sci. Rep. 2020, 10, 21014

    2020

  2. [10]

    L. Li, F. Yang, Q. Ma, T. Xu, M. Su, Y . Zhao, X. Liu, Optica 2025, 12, 263

    2025

  3. [11]

    H. C. Lee, N. Elder, M. Leal, S. Stantial, E. Vergara Martinez, S. Jos, H. Cho, S. Russo, Nat. Commun. 2024, 15, 8456

    2024

  4. [12]

    D. Feng, W. Zhou, X. Qiao, J. Albert, Sci. Rep. 2015, 5, 17415

    2015

  5. [13]

    Jeong, M

    S.-M. Jeong, M. Son, Y . Kang, J. Yang, T. Lim, S. Ju, Npj Flex. Electron. 2021, 5, 35

    2021

  6. [14]

    Z. Che, X. Wan, J. Xu, C. Duan, T. Zheng, J. Chen, Nat. Commun. 2024, 15, 1873

    2024

  7. [15]

    Lo Presti, F

    D. Lo Presti, F. Santucci, C. Massaroni, D. Formica, R. Setola, E. Schena, Sci. Rep. 2021, 11, 21162

    2021

  8. [16]

    Issatayeva, A

    A. Issatayeva, A. Amantayeva, W. Blanc, D. Tosi, C. Molardi, Sci. Rep. 2021, 11, 8609

    2021

  9. [17]

    Facchin, S

    M. Facchin, S. N. Khan, K. Dholakia, G. D. Bruce, Nat. Rev. Phys. 2024, 6, 500

    2024

  10. [18]

    Lengenfelder, F

    B. Lengenfelder, F. Mehari, M. Hohmann, M. Heinlein, E. Chelales, M. J. Waldner, F. Klämpfl, Z. Zalevsky, M. Schmidt, Sci. Rep. 2019, 9, 1057

    2019

  11. [19]

    Shimadera, K

    S. Shimadera, K. Kitagawa, K. Sagehashi, Y . Miyajima, T. Niiyama, S. Sunada, Sci. Rep. 2022, 12, 13096

    2022

  12. [20]

    C. Lu, J. Su, X. Dong, T. Sun, K. T. V . Grattan, J. Light. Technol. 2018, 36, 2796

    2018

  13. [21]

    Y . Zhao, L. Cai, X.-G. Li, IEEE Trans. Instrum. Meas. 2017, 66, 141

    2017

  14. [22]

    X. Cai, S. Gao, M. Wu, S. Zheng, H. Fu, D. Chen, Opt. Laser Technol. 2024, 168, 109843

    2024

  15. [23]

    R. K. Gupta, G. D. Bruce, S. J. Powis, K. Dholakia, Laser Photonics Rev. 2020, 14, 2000120

    2020

  16. [24]

    Y . Liu, G. Li, Q. Qin, Z. Tan, M. Wang, F. Yan, Opt. Laser Technol. 2020, 131, 106424

    2020

  17. [25]

    G. Li, Y . Liu, Q. Qin, X. Zou, M. Wang, W. Ren, IEEE Sens. J. 2022, 22, 15974

    2022

  18. [26]

    Razmyar, M

    S. Razmyar, M. T. Mostafavi, J. Light. Technol. 2021, 39, 1850

    2021

  19. [27]

    X. Wang, Y . Wang, K. Zhang, K. Althoefer, L. Su, Sci. Rep. 2022, 12, 12684

    2022

  20. [28]

    M. Wei, G. Tang, J. Liu, L. Zhu, J. Liu, C. Huang, J. Zhang, L. Shen, S. Y u, J. Light. Technol. 2021, 39, 6315

    2021

  21. [29]

    K. Sun, Z. Ding, Z. Zhang, Appl. Opt. 2020, 59, 5745

    2020

  22. [30]

    Z. Ding, Z. Zhang, Opt. Laser Technol. 2021, 136, 106760

    2021

  23. [31]

    X. Liu, S. He, J. Kang, B. Liu, C. Zhu, Opt. Express 2024, 32, 31783

    2024

  24. [32]

    Y . Li, Y . Xue, L. Tian, Optica 2018, 5, 1181

    2018

  25. [33]

    Rahmani, D

    B. Rahmani, D. Loterie, G. Konstantinou, D. Psaltis, C. Moser, Light Sci. Appl. 2018, 7, 69

    2018

  26. [34]

    G. E. Karniadakis, I. G. Kevrekidis, L. Lu, P. Perdikaris, S. Wang, L. Yang, Nat. Rev. Phys. 2021, 3, 422

    2021

  27. [35]

    C. Zuo, J. Qian, S. Feng, W. Yin, Y . Li, P. Fan, J. Han, K. Qian, Q. Chen, Light Sci. Appl. 2022, 11, 39

    2022

  28. [36]

    Jiang, D

    X. Jiang, D. Wang, Q. Fan, M. Zhang, C. Lu, A. P. T. Lau, Laser Photonics Rev. 2022, 16, 2100483

    2022

  29. [37]

    Z. Tian, L. Pei, J. Wang, K. Hu, W. Xu, J. Zheng, J. Li, T. Ning, Opt. Express 2022, 30, 39932

    2022

  30. [38]

    X. Zhan, Y . Liu, M. Tang, L. Ma, R. Wang, L. Duan, L. Gan, C. Yang, W. Tong, S. Fu, D. Liu, Z. He, Opt. Express 2018, 26, 15332

    2018

  31. [39]

    Ashry, Y

    I. Ashry, Y . Mao, A. Trichili, B. Wang, T. K. Ng, M.-S. Alouini, B. S. Ooi, IEEE Access 2020, 8, 179592

    2020

  32. [40]

    Demas, S

    J. Demas, S. Ramachandran, Opt. Express 2014, 22, 23043

    2014

  33. [41]

    D. M. Nguyen, S. Blin, T. N. Nguyen, S. D. Le, L. Provino, M. Thual, T. Chartier, Appl. Opt. 2012, 51, 450

    2012

  34. [42]

    Paurisse, L

    M. Paurisse, L. Lévèque, M. Hanna, F. Druon, P. Georges, Opt. Express 2012, 20, 4074

    2012

  35. [43]

    Y . Z. Ma, Y . Sych, G. Onishchukov, S. Ramachandran, U. Peschel, B. Schmauss, G. Leuchs, Appl. Phys. B 2009, 96, 345

    2009

  36. [44]

    Forbes, A

    A. Forbes, A. Dudley, M. McLaren, Adv. Opt. Photonics 2016, 8, 200

    2016

  37. [45]

    Brüning, P

    R. Brüning, P. Gelszinnis, C. Schulze, D. Flamm, M. Duparré, Appl. Opt. 2013, 52, 7769

    2013

  38. [46]

    Shapira, A

    O. Shapira, A. F. Abouraddy, J. D. Joannopoulos, Y . Fink, Phys. Rev. Lett. 2005, 94, 143902

    2005

  39. [47]

    H. Lü, P. Zhou, X. Wang, Z. Jiang, Appl. Opt. 2013, 52, 2905

    2013

  40. [48]

    L. Li, J. Leng, P. Zhou, J. Chen, Opt. Express 2017, 25, 19680

    2017

  41. [49]

    W. Yan, X. Xu, J. Wang, Opt. Express 2019, 27, 13871

    2019

  42. [50]

    Manuylovich, A

    E. Manuylovich, A. Donodin, S. Turitsyn, Opt. Express 2021, 29, 36769

    2021

  43. [51]

    E. S. Manuylovich, V . V . Dvoyrin, S. K. Turitsyn, Nat. Commun. 2020, 11, 5507

    2020

  44. [52]

    P. S. Anisimov, V . V . Zemlyakov, J. Gao, Opt. Express 2022, 30, 8804

    2022

  45. [53]

    Z. Tian, L. Pei, J. Wang, J. Zheng, K. Hu, Y . Chang, Z. Li, J. Li, T. Ning, J. Light. Technol. 2022, 40

    2022

  46. [54]

    Y . An, L. Huang, J. Li, J. Leng, L. Yang, P. Zhou, Opt. Express 2019, 27, 10127

    2019

  47. [55]

    Cheng, X

    K. Cheng, X. Yang, X. Wu, IEEE Trans. Geosci. Remote Sens. 2024, 62, 1

    2024

  48. [56]

    J. Ye, X. Ma, Y . Zhang, J. Xu, H. Zhang, T. Yao, J. Leng, P. Zhou, PhotoniX 2021, 2, 15

    2021

  49. [57]

    L. Wu, X. Y uan, Y . Tang, S. Wageh, O. A. Al-Hartomy, A. G. Al-Sehemi, J. Yang, Y . Xiang, H. Zhang, Y . Qin, PhotoniX 2023, 4, 15

    2023

  50. [58]

    Nocedal, S

    J. Nocedal, S. J. Wright, Numerical Optimization, Springer, New York, 1999

    1999

  51. [59]

    Falciai, C

    R. Falciai, C. Trono, IEEE Sens. J. 2005, 5, 1310

    2005

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.