Pith. sign in

REVIEW 3 major objections 5 minor 47 references

Frequency Range Boosted Magnetometry Beyond the Spin Coherence Limit via Compressive Sensing

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read Compressive sensing extends free-induction-decay magnetometry's resolvable frequency range from 251 Hz to 3000 Hz while keeping sensitivity at 4 pT/√Hz.

desk verdict The paper's central claim rests on a multi-rate reconstruction whose linear model is not justified by the experimental timing; worth refereeing, but the authors must fix the phase problem and validation gaps. read the letter →

arxiv 2505.05825 v1 pith:CEK2PLXD submitted 2025-05-09 physics.atom-ph physics.optics

classification physics.atom-phphysics.optics PACS 07.55.Ge
keywords compressivesensingfreeinductiondecaymagnetometryspincoherencelimitsub-Nyquistsamplingfrequencyaliasingatomicmagnetometersparsespectrumreconstructionquantum
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

Free-induction-decay magnetometry normally faces a trade-off: a longer coherence time improves sensitivity but, by the sampling theorem, forces a narrower resolvable frequency range and causes frequency aliasing. This paper reports that compressive sensing breaks that trade-off. By recording the same decay at 41 distinct sub-Nyquist repetition rates and jointly reconstructing a sparse spectrum, the resolvable range of an alkali-vapor FID magnetometer is extended from the $251\,\mathrm{Hz}$ limit to $3000\,\mathrm{Hz}$, with alias-induced spurious peaks suppressed. The recovered $800\,\mathrm{Hz}$ test tone appears at the right frequency and, after correcting the magnetometer's low-pass response, at the right amplitude within 0.6%. Sensitivity stays at $4\,\mathrm{pT}/\sqrt{\mathrm{Hz}}$ at $100\,\mathrm{Hz}$, so the frequency extension comes at no measured sensitivity cost.

What carries the argument

The load-bearing mechanism is multi-rate asynchronous sub-Nyquist sampling (MASS): 41 measurement matrices $\Phi_i$ are built from prime sample lengths $M_i$ of order $\sqrt{N_s}$, chosen so different sampling rates fold high-frequency components into distinguishable alias patterns. Concatenating the records gives $Y = \Phi X$, a linear system whose solution $X$ is recovered by a least-squares solver. The scaling law $f_{\mathrm{CS}} \simeq T f_{\mathrm{SCL}} / \tau$ carries the argument: total acquisition time $T$ buys enough independent low-rate projections to replace the coherence-time-limited range with a sparsity-limited range, provided the spectrum has at most $k$ nonzero lines and the number of rates $v$ exceeds $2k-1$.

What would settle it

Apply a field whose two-sided spectrum contains more than 21 nonzero components (e.g., 11 distinct positive tones above 251 Hz plus their negative mirrors), or a single tone that drifts by more than 1 Hz over the 41 s acquisition; if the reconstruction shows spurious peaks or omits true tones at the expected amplitudes, the central claim is refuted.

Watch

Extended reading notes

Core claim

The central claim is that the spin coherence limit is not a hard ceiling for frequency range: a linear-system reconstruction can recover spectral components far above the single-shot sampling rate as long as the underlying spectrum is sparse. The authors demonstrate this with a free-induction-decay $^{87}\mathrm{Rb}$ magnetometer whose coherence time sets the per-shot sampling rate near $500\,\mathrm{Hz}$. They take 41 one-second records, each sampled at a different prime repetition rate between $263\,\mathrm{Hz}$ and $503\,\mathrm{Hz}$, form the vertical stack $Y$ from them, and solve $Y = \Phi X$ for the two-sided spectrum $X$ using a least-squares reconstruction. The reconstructed spectrum reaches $3000\,\mathrm{Hz}$, a factor of 12 beyond the $251\,\mathrm{Hz}$ limit set by the sampling rate, and shows the applied $800\,\mathrm{Hz}$ modulation plus 50 Hz power-line harmonics; the same signal appears as different aliases (12 Hz, 43 Hz, 207 Hz) in individual undersampled records, demonstrating that the joint reconstruction rather than any single record resolves it. Sensitivity is unchanged at $4\,\mathrm{pT}/\sqrt{\mathrm{Hz}}$ at $100\,\mathrm{Hz}$, limited by photon shot noise, and the reconstructed amplitude matches the applied field after a low-pass frequency-response correction.

Load-bearing premise

The reconstruction treats the 41 one-second records as different projections of one fixed, sparse magnetic-field spectrum; the central claim collapses if the field drifts in frequency or contains many closely spaced tones during the 41-second acquisition.

Editorial extensions

If this is right

  • For any FID-based sensor, the resolvable frequency range can be lifted from about $1/(2\tau)$ to roughly $(T/\tau)\cdot f_{\mathrm{SCL}}$, so longer total averaging time directly buys more bandwidth at fixed sensitivity.
  • Frequency-alias artifacts that would otherwise appear inside the original band—here, spurious peaks at 12 Hz, 43 Hz, and 207 Hz from an 800 Hz signal—are removed by the joint reconstruction.
  • A tone above the old limit is recovered with correct frequency and, after low-pass response correction, correct amplitude to within 0.6%.
  • The sensitivity of the CS-enhanced magnetometer remains photon-shot-noise limited at $4\,\mathrm{pT}/\sqrt{\mathrm{Hz}}$ at $100\,\mathrm{Hz}$, meaning the range extension does not degrade sensitivity.
  • The sparsity requirement quantifies the trade-off: with $v$ sampling rates, the bilateral spectrum must contain at most $k=(v+1)/2$ nonzero components for the MASS reconstruction to be valid.

Reading between the lines

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

  • The same multi-rate reconstruction should transfer to other long-coherence FID platforms such as NV-center or noble-gas-spin sensors, where the per-shot sampling rate is even smaller relative to the frequencies of interest; the paper motivates but does not demonstrate this transfer.
  • The formula $f_{\mathrm{CS}} \simeq (T/\tau)f_{\mathrm{SCL}}$ implies the 12-fold gain is not a ceiling: longer total acquisition time $T$ should push the resolvable range higher, provided the spectrum stays sparse and stationary over that longer window.
  • A stress test with a slowly drifting tone or a sparse-but-time-varying spectrum would quantify the failure threshold of the stationarity assumption; the paper does not provide such a robustness curve.
  • Because the method is a post-processing reconstruction rather than a change to the spin system, it could in principle be combined with existing coherence-extension techniques such as spin locking or dynamic decoupling, pushing sensitivity and bandwidth simultaneously.
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

3 major / 5 minor

Summary. The manuscript reports a demonstration of compressive sensing (CS) with multi-rate sub-Nyquist sampling applied to a free-induction-decay (FID) 87Rb magnetometer. The authors model the undersampled DFT spectra from 41 prime repetition rates (263–503 Hz) as linear projections of a common high-resolution spectrum, solve the concatenated system with Lawson–Hanson non-negative least squares, and reconstruct a spectrum extending to 3000 Hz. They identify an 800 Hz applied modulation tone outside the nominal 251 Hz Nyquist range, suppress alias peaks seen in individual undersampled spectra, and report unchanged photon-shot-noise-limited sensitivity of 4 pT/√Hz at 100 Hz. A scaled frequency-response correction recovers the applied 800 Hz amplitude to within 0.6%, and a separate linearity check at 137 Hz gives slope 1.006. The paper claims a factor-of-12 extension of the resolvable frequency range and a general route beyond the 'spin coherence limit.'

Significance. The core idea — using sparse recovery over multiple sub-Nyquist rates to convert long coherence time from a bandwidth limitation into a reconstruction advantage — is potentially useful for atomic magnetometry and other FID-based sensors. The paper includes real experimental data, a concrete 800 Hz recovery above the nominal Nyquist limit, an explicit comparison of aliased spectra before and after reconstruction, and a sensitivity measurement. These strengths make the empirical demonstration worth taking seriously. However, the theoretical scaling law in Eq. (3) is inconsistent with the experimental parameters by about a factor of 40, and the linear measurement model in Eq. (10) requires a common absolute phase reference that is not documented. Both points are load-bearing for the central claim, so the manuscript needs substantive revision before the result can be accepted.

major comments (3)
  1. [Measurement Process and Compressive Sensing, Eq. (3)] Eq. (3) claims f_CS ≃ T/(2τ²), but this scaling is contradicted by the experimental parameters. The conventional limit is quoted as 251 Hz at a pump-probe repetition rate of 503 Hz, which gives τ ≈ 1/(2×251) ≈ 2.0 ms. Inserting T = 1 s and τ ≈ 2.0 ms into Eq. (3) yields f_CS ≈ 125 kHz, not the demonstrated 3000 Hz. The text does not identify the actual coherence time used in the experiment or explain why the realized extension is about 40 times smaller than the predicted limit. The stated design condition M_i ∼ sqrt(N_s) is also not satisfied: for f_CS = 3000 Hz and T = 1 s one has N_s = 6000 and sqrt(N_s) ≈ 77, while the 41 experimental sampling rates are primes from 263 to 503 Hz. Please correct the scaling or provide the actual parameters and an explanation of the discrepancy.
  2. [Experimental Results / Methods Eq. (10)] The construction of Y by vertical concatenation of the 41 DFT spectra is the key step that turns the experiment into a CS problem, but the linear relation Y_i = Φ_i X is not justified in the manuscript. Fig. 4(C,E,G) shows 'amplitude spectra', whereas Eq. (10) is a coherent linear relation between complex spectra. If the concatenated Y are magnitudes, the relation is nonlinear because the magnitude of a sum of aliased components is not the sum of their magnitudes. If Y are complex DFT values, each 1-s record contains an arbitrary initial phase φ0 from Eq. (4) and the start times of records at different sampling rates are not stated to be synchronized to the 800 Hz source; with record-dependent phases the true relation is Y_i = D_i Φ_i X, and the concatenated system does not have the assumed form. The Lawson–Hanson solver is applied without specifying how complex data are handled. Please specify the exact contents of Y and the phase reference, and either establish phase coherence experimentally or reformulate the reconstruction to be invariant to record-dependent phases.
  3. [Reconstruction consistency, Fig. 7] The amplitude-linearity test in Fig. 7 is performed at 137 Hz, which lies inside the conventional 251 Hz Nyquist band. It therefore does not test whether the CS reconstruction preserves amplitudes in the extended range beyond the SCL. The only above-Nyquist amplitude validation is the single 800 Hz point, and that point uses an independently calibrated low-pass response correction. To support the claim that the method recovers the correct amplitude at the correct frequency in the extended band, the authors should provide amplitude validation at several frequencies above 251 Hz, or at least quantify the reconstruction error over the full 3000 Hz range.
minor comments (5)
  1. [Fig. 4 caption] The caption lists panels (B), (C), and (F) where the text refers to (B,D,F) and (C,E,G); the caption should be corrected.
  2. [Methods, Eq. (5)] The DFT formula in Eq. (5) uses the summation index k but the exponent contains t and N_i; the sum should be over the sampled time index consistently.
  3. [References] Reference 15 is malformed ('K. M. Budker, D.'), reference 2 is duplicated as reference 6, and reference 10 concatenates two distinct works that should be separated.
  4. [Experimental Results] The reconstruction dimension N_s, the number of frequency bins, and the sparsity level k are not defined in the experimental section; these parameters should be stated explicitly, including how k = 8 is determined and whether it is known before or after reconstruction.
  5. [Text] There are several typographical errors, including 'detunning' and 'denots', that should be corrected in a final pass.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central CS recovery is benchmarked against an externally applied 800 Hz tone and a DFT-derived measurement matrix, not against fitted or self-referential inputs.

full rationale

The paper's central derivation is self-contained. The measurement matrix Phi is constructed from the DFT aliasing relation (Eqs. 5-14), not fitted to the experimental spectrum; the MASS sub-Nyquist scheme is cited to an external source [40], and the Lawson-Hanson solver [41] is a standard external algorithm. The key experimental evidence is an externally applied 800 Hz magnetic field (Fig. 4A) that appears at 800 Hz after CS reconstruction (Fig. 4H); this is an out-of-sample check, not an input to the reconstruction. The amplitude correction uses a separately measured frequency response (Fig. 6) and a separate linearity calibration at 137 Hz (Fig. 7), so the 0.6% consistency is calibration, not a circular prediction. Equation (3) is a scaling relation derived from T, tau, and the CS requirements M_i ~ T/tau, M_i ~ sqrt(N_s), N_s = 2T f_CS; it is not reverse-engineered from the reported 3000 Hz value. The post-hoc statement that k=8 from the reconstructed spectrum satisfies v > 2k-1 is a consistency check rather than a definitional step; it is not load-bearing for the 800 Hz recovery. The phase-coherence issue raised about Y = Phi X (record-dependent phase in concatenated DFTs) is a validity/correctness concern about the measurement model, not a circularity of the derivation. Minor self-citations (refs 16, 17, 42) are background support and do not carry the central claim.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central claim rests on sparsity and stationarity assumptions that are not tested with multi-tone or broadband inputs, plus design choices (v=41, N_s=6000) whose relation to the theoretical scaling is unexplained.

free parameters (4)
  • Reconstructed frequency range f_CS = 3000 Hz
    The paper demonstrates a range to 3000 Hz but does not explain why this value is chosen or how it relates to the theoretical limit of Eq (3).
  • Number of sampling rates v = 41
    The experiment uses 41 prime sampling rates from 263 Hz to 503 Hz; the specific primes are not listed, and the choice of 41 is motivated by v > 2k-1 with k=8.
  • Reconstruction dimension N_s = 6000 (2T f_CS)
    N_s is set to 2T f_CS with T=1 s and f_CS=3000 Hz; this determines the measurement matrix dimensions and the frequency resolution.
  • Low-pass filter parameters for frequency response = Not reported
    The frequency response in Fig. 6 is fitted with a low-pass transfer function and used to correct the reconstructed amplitude; the fitted parameters are not given.
assumptions (5)
  • domain assumption The magnetic-field signal is stationary over the total acquisition time (41 s).
    The 41 sub-Nyquist records are concatenated to form Y, requiring the spectrum X to be identical across all records. This is not tested.
  • domain assumption The spectrum X is k-sparse, with k=8 in the demonstration.
    The MASS/CS reconstruction relies on sparsity to disambiguate aliased frequency components; the paper sets k=8 after seeing the reconstructed spectrum.
  • standard math The measurement matrix built from prime sampling rates is incoherent enough to allow exact recovery.
    This is the MASS condition from Sun et al. (ref 40), accepted as established prior work.
  • domain assumption The FID signal has the form S(t)=A0 e^{-t/τ} sin(2πΩt + φ0).
    Standard model for free induction decay in atomic magnetometry, cited from ref 6.
  • domain assumption The Lawson-Hanson algorithm solves Y=ΦX to the true spectrum.
    The paper does not justify that least-squares with non-negativity yields the sparse solution when the system may be underdetermined; this is assumed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Frequency Range Boosted Magnetometry Beyond the Spin Coherence Limit via Compressive Sensing." pith.science (2026). https://pith.science/paper/CEK2PLXD

@misc{pith2026250505825,
  author       = {Pith},
  title        = {Pith review of: Frequency Range Boosted Magnetometry Beyond the Spin Coherence Limit via Compressive Sensing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CEK2PLXD}},
  note         = {Machine review of arXiv:2505.05825}
}
read the original abstract

Free induction decay (FID) of spin precession serves as an essential tool for quantum sensing across diverse platforms. While extending spin coherence time remains critical for sensitivity enhancement, the requisite long single-shot acquisitions narrow the resolvable frequency range, establishing a fundamental ``spin coherence limit (SCL)'', according to the Nyquist Sampling Theorem. Besides, conventional spectral analysis for FID measurement suffers from frequency alias, causing signal attenuation and positional errors that compromise the measurement validity. Here, we demonstrate a general frequency-range-extended technique that overcomes SCL by leveraging compressive sensing. By applying this method to the FID magnetometer, we expand the resolvable frequency range significantly from the Nyquist-limited range of 251\,Hz to 3000\,Hz, effectively avoiding frequency alias. Our work paves the way for implementing long-coherence-time spin systems in high-sensitivity, broad-bandwidth, and alias-free magnetic field sensing.

Figures

Figures reproduced from arXiv: 2505.05825 by the authors.

Figure 1
Figure 1. The diagram of frequency alias and signal reconstruction. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. shows trade-off relations between normalized sensitivity of Larmor frequency δΩ0 = √2𝑁𝐹 δΩ = 1/(𝜏𝑇) and frequency range with and without CS. Blue and red lines denote the SCL and the CS-enhanced limit, respectively. 𝑇 is set to be 1 s and 10 s for solid lines and dashed lines, respectively. An extended 𝑇 allows for a greater number of sampling points under a fixed 𝜏, thereby showing a better spectral extension perfo… view at source ↗
Figure 3
Figure 3. Experimental setup. PBS: polarizing beam splitter. BS: beam splitter. AOM: acousto-optic modulator used to pulse the pump and repump laser. AWG: arbitrary waveform generator. λ/2, half￾wave plate. λ/4: quarter-wave plate. BPD: balanced photodetector. Atoms are contained in a vapor cell positioned in the center of the magnetic shield and are pumped and probed by laser beams under a static magnetic field 𝐵0 (along 𝑧̂)… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Wideband spectral reconstruction based on CS. [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Here, pump-probe repetition rate is set to 503 Hz. For conventional FID magnetometry, two aliasing peaks appear at 60 Hz and 206 Hz, with its resolvable frequency range limited to half of the pump-probe repetition rate, which here is 251 Hz . In contrast, the CS-enhanc…
Figure 5
Figure 5. Figure 5: Sensitivity Comparison. (A) Sensitivity of the conventional FID magnetometer and the CS￾enhanced FID magnetometer. (B) An enlarged view of the area enclosed by the dashed line in (A). The sensitivity of both magnetometer configurations remains consistent, with PSN limi…
Figure 6
Figure 6. Figure 6: The frequency response of the magnetometer. T [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: The relationship of the reconstructed magnetic field amplitude Brec and the applied magnetic modulation amplitude Bmod. The circles are experimental data and the solid line is the fit with linear function. Construction of the measurement matrix The DFT process between …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

47 extracted references · 47 canonical work pages

  1. [1]

    Sheng, S

    D. Sheng, S. Li, N. Dural, M. V. Romalis, Subfemtotesla Scalar Atomic Magnetometry Using Multipass Cells. Phys. Rev. Lett. 110, 160802 (2013)

  2. [2]

    Lucivero, W

    V. Lucivero, W. Lee, M. Limes, E. Foley, T. Kornack, M. Romalis, Femtotesla Nearly - Quantum-Noise-Limited Pulsed Gradiometer at Earth-Scale Fields. Phys. Rev. Appl. 18, L021001 (2022)

  3. [3]

    Colangelo, F

    G. Colangelo, F. M. C iurana, L. C. Bianchet, R. J. Sewell, M. W. Mitchell, Simultaneous tracking of spin angle and amplitude beyond classical limits. Nature 543 (7646), 525–528 (2017)

  4. [4]

    Limes, E

    M. Limes, E. Foley, T. Kornack, S. Caliga, S. McBride, A. Braun, W. Lee, V. Lucivero, M. Romalis, Portable Magnetometry for Detection of Biomagnetism in Ambient Environments. Phys. Rev. Appl. 14, 011002 (2020)

  5. [5]

    Jensen, R

    K. Jensen, R. Budvytyte, R. A. Thomas, T. Wang, A. M. Fuchs, M. V. Balabas, G. Vasilakis, L. D. Mosgaard, H. C. Stærkind, J. H. Mü l ler, T. Heimburg, S. -P. Olesen, E. S. Polzik, Non - invasive detection of animal nerve impulses with an atomic magnetometer operating near quantum limited sensitivity. Scientific Reports 6 (1), 29638 (2016)

  6. [6]

    Lucivero , W

    V. Lucivero , W. Lee, M. Limes, E. Foley, T. Kornack, M. Romalis, Femtotesla Nearly - Quantum-Noise-Limited Pulsed Gradiometer at Earth-Scale Fields. Phys. Rev. Appl. 18, L021001 (2022)

  7. [7]

    Bü rgler, T

    B. Bü rgler, T. F. Sjolander, O. Brinza, A. Tallaire, J. Achard, P. Maletinsky, All-optical nuclear quantum sensing using nitrogen-vacancy centers in diamond. npj Quantum Information 9 (1), 56 (2023)

  8. [8]

    Cujia, J

    K. Cujia, J. M. Boss, K. Herb, J. Zopes, C. L. Degen, Tracking the precession of single nuclear spins by weak measurements. Nature 571 (7764), 230–233 (2019)

Show all 47 references
  1. [9]

    Y. S. Greenberg, Application of superconducting quantum interference devices to nuclear magnetic resonance. Rev. Mod. Phys. 70, 175–222 (1998)

  2. [10]

    C. Liu, H. Dong, J. Sang, Submillimeter -resolution magnetic field imaging with digital micromirror device and atomic vapor cell. Applied Physics Letters 119 (11), 114002 (2021). K. Yi, Y. Liu, B. Wang, W. Xiao, D. Sheng, X. Peng, H. Guo, Free -induction-decay ⁴He magnetometer...

  3. [11]

    Rosner, D

    M. Rosner, D. Beck, P. Fierlinger, H. Filter, C. Klau, F. Kuchler, P. Rö sner, M. Sturm, D. Wurm, Z. Sun, A highly drift -stable atomic magnetometer for fundamental physics experiments. Applied Physics Letters 120 (16), 161102 (2022)

  4. [12]

    K. Yi, Y. Liu, B. Wang, W. Xiao, D. Sheng, X. Peng, H. Guo, Free -induction-decay ⁴He magnetometer using a multipass cell. Phys. Rev. Appl. 22, 014084 (2024)

  5. [13]

    M. N. Nabighian, V. Grauch, R. Hansen, T. LaFehr, Y. Li, J. W. Peirce, J. D. Phillips, M. Ruder, The histori cal development of the magnetic method in exploration. Geophysics 70 (6), 33ND–61ND (2005)

  6. [14]

    Jensen, B

    K. Jensen, B. H. Bentzen, E. S. Polzik, Small animal biomagnetism applications, in Flexible High Performance Magnetic Field Sensors: On -Scalp Magnetoencephalogr aphy and Other Applications (Springer), pp. 33–48 (2022)

  7. [15]

    K. M. Budker, D., Sensing: Equation One. Optical Memory and Neural Networks 32 (3), S409–S414 (2023)

  8. [16]

    G. Bao, A. Wickenbrock, S. Rochester, W. Zhang, D. Budker, Suppression of the Nonlinear Zeeman Effect and Heading Error in Earth-Field-Range Alkali-Vapor Magnetometers. Phys. Rev. Lett. 120, 033202 (2018)

  9. [17]

    P. Yang, G. Bao, L. Q. Chen, W. Zhang, Coherence Protection of Electron Spin in Earth-Field Range by All-Optical Dynamic Decoupling. Phys. Rev. Appl. 16, 014045 (2021)

  10. [18]

    M. Xu, M. Jiang, Y. Wang, H. Su, Y. Huang, X. Peng, Cooperative Spin Amplifier for Enhanced Quantum Sensing. Phys. Rev. Lett. 133, 133202 (2024)

  11. [19]

    Y. Yang, W. -T. Luo, J. -L. Zhang, S. -Z. Wang, C. -L. Zou, T. Xia, Z. -T. Lu, Minute -scale schrö dinger-cat state of spin-5/2 atoms. Nature Photonics pp. 1–6 (2024)

  12. [20]

    Jerri, The Shannon sampling theorem —Its various extensions and applications: A tutorial review

    A. Jerri, The Shannon sampling theorem —Its various extensions and applications: A tutorial review. Proceedings of the IEEE 65 (11), 1565–1596 (1977)

  13. [21]

    Candes, J

    E. Candes, J. Romberg, T. Tao, Robust uncertainty principles: exact signal reconstruction from highly incomplete frequency information. IEEE Transactions on Information Theory 52 (2), 489– 509 (2006)

  14. [22]

    Donoho, Compressed sensing

    D. Donoho, Compressed sensing. IEEE Transactions on Info rmation Theory 52 (4), 1289 – 1306 (2006)

  15. [23]

    E. J. Candè s, T. Tao, Near-Optimal Signal Recovery From Random Projections: Universal Encoding Strategies? IEEE Transactions on Information Theory 52 (12), 5406–5425 (2006)

  16. [24]

    E. J. Candes, M. B. Wakin, An Introduction To Compressive Sampling. IEEE Signal Processing Magazine 25 (2), 21–30 (2008)

  17. [25]

    D. L. D onoho, M. Elad, Optimally sparse representation in general (nonorthogonal) dictionaries via ℓ₁ minimization. Proceedings of the National Academy of Sciences 100 (5), 2197– 2202 (2003)

  18. [26]

    Candè s, J

    E. Candè s, J. Romberg, Sparsity and incoherence in compressive sampling. Inverse Problems 23 (3), 969 (2007)

  19. [27]

    Otazo, E

    R. Otazo, E. Candes, D. K. Sodickson, Low -rank plus sparse matrix decomposition for accelerated dynamic MRI with separation of background and dynamic components. Magnetic Resonance in Medicine 73 (3), 1125–1136 (2015)

  20. [28]

    P. Cao, E. X. Wu, Accelerating phase -encoded proton MR spectroscopic imaging by compressed sensing. Journal of Magnetic Resonance Imaging 41 (2), 487–495 (2015)

  21. [29]

    Huang, L

    W. Huang, L. Xiao, H. Liu, Z. Wei, Hyperspectral imagery super -resolution by compressive sensing inspired dictionary learning and spatial-spectral regularization. Sensors 15 (1), 2041–2058 (2015)

  22. [30]

    Lustig, D

    M. Lustig, D. Donoho, J. M. Pauly, Sparse MRI: The application of compressed sensing for rapid MR imaging. Magnetic Resonance in Med icine: An Official Journal of the International Society for Magnetic Resonance in Medicine 58 (6), 1182–1195 (2007)

  23. [31]

    Lustig, D

    M. Lustig, D. L. Donoho, J. M. Santos, J. M. Pauly, Compressed Sensing MRI. IEEE Signal Processing Magazine 25 (2), 72–82 (2008)

  24. [32]

    T. M. Quan, T. Nguyen-Duc, W.-K. Jeong, Compressed Sensing MRI Reconstruction Using a Generative Adversarial Network With a Cyclic Loss. IEEE Transactions on Medical Imaging 37 (6), 1488–1497 (2018)

  25. [33]

    G. Yang, S. Yu, H. Dong, G. Slabaugh, P. L. Dragotti, X. Ye, F. Liu, S. Arridge, J. Keegan, Y. Guo, D. Firmin, DAGAN: Deep De -Aliasing Generative Adversarial Networks for Fast Compressed Sensing MRI Reconstruction. IEEE Transactions on Medical Imaging 37 (6), 1310– 1321 (2018)

  26. [34]

    Giacobello, M

    D. Giacobello, M. G. Christenson, M. N. Murthi, S. H. Jensen, M. Moonen, Retrieving Sparse Patterns Using a Compressed Sensing Framework: Applications to Speech Coding Based on Sparse Linear Prediction. IEEE Signal Processing Letters 17 (1), 103–106 (2010)

  27. [35]

    X.-Y. Liu, Y. Zhu, L. Kong, C. Liu, Y. Gu, A. V. Vasilakos, M. -Y. Wu, CDC: Compressive Data Collection for Wireless Sensor Networks. IEEE Transactions on Parallel and Distributed Systems 26 (8), 2188–2197 (2015)

  28. [36]

    J.-P. Hong, W. Choi, B. D. Rao, Sparsity controlled random multiple access with compressed sensing. IEEE Transactions on Wireless Communications 14 (2), 998–1010 (2014)

  29. [37]

    C. R. Berger, Z. Wang, J. Huang, S. Zhou, Application of compressive sensing to sparse channel estimation. IEEE Communications Magazine 48 (11), 164–174 (2010)

  30. [38]

    T. T. Do, Y. Chen, D. T. Nguyen, N. Nguyen, L. Gan, T. D. Tran, Distributed compressed video sensing, in 2009 16th IEEE International Conference on Image Processing (ICIP) (2009), pp. 1393–1396

  31. [39]

    Hegde, R

    C. Hegde, R. G. Baraniuk, Signal R ecovery on Incoherent Manifolds. IEEE Transactions on Information Theory 58 (12), 7204–7214 (2012)

  32. [40]

    H. Sun, W. -Y. Chiu, J. Jiang, A. Nallanathan, H. V. Poor, Wideband spectrum sensing with sub-Nyquist sampling in cognitive radios. IEEE Transactions on Signal Processing 60 (11), 6068– 6073 (2012)

  33. [41]

    Linear Least Squares with Linear Inequality Constraints

    C. L. Lawson, R. J. Hanson, “Linear Least Squares with Linear Inequality Constraints” in Solving Least Squares Problems (Prentice Hall, Upper Saddle River, NJ, 1974), pp.158–173

  34. [42]

    S. Wu, G. Bao, J. Guo, J . Chen, W. Du, M. Shi, P. Yang, L. Chen, W. Zhang, Quantum magnetic gradiometer with entangled twin light beams. Science Advances 9 (15), eadg1760 (2023)

  35. [43]

    V. Shah, G. Vasilakis, M. V. Romalis, High Bandwidth Atomic Magnetometry with Continuous Quantum Nondemolition Measurements. Phys. Rev. Lett. 104, 013601 (2010)

  36. [44]

    M. V. Balabas, T. Karaulanov, M. P. Ledbetter, D. Budker, Polarized Alkali-Metal Vapor with Minute-Long Transverse Spin-Relaxation Time. Phys. Rev. Lett. 105, 070801 (2010)

  37. [45]

    Shaham, O

    R. Shaham, O. Katz, O. Firstenberg, Strong coupling of alkali -metal spins to noble-gas spins with an hour-long coherence time. Nature Physics 18 (5), 506–510 (2022)

  38. [46]

    Bar-Gill, L

    N. Bar-Gill, L. M. Pham, A. Jarmola, D. Budker, R. L. Walsworth, Solid-state electronic spin coherence time approaching one second. Nature Communications 4 (1), 1743 (2013)

  39. [47]

    Egorov, R

    M. Egorov, R. P. Anderson, V. Ivannikov, B. Opanchuk, P. Drummond, B. V. Hall, A. I. Sidorov, Long-lived periodic revivals of coherence in an interacting Bose –Einstein condensate. Phys. Rev. A 84, 021605 (2011). ACKNOWLEDGMENTS This work is supported by the Innovation Program...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.