Pith. sign in

REVIEW 3 major objections 6 minor 1 cited by

Optical Convolutional Spectrometer

T0 review · 3 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Tunable periodic filters turn a spectrum scan into a circular convolution, so recovery is one Fourier-domain division.

desk verdict A genuinely new miniaturized spectrometer architecture with convincing core experiments; publish after fixing Eq. (3), quantifying the rigid-shift assumption, and clarifying biomarker train/test splits. read the letter →

arxiv 2502.08749 v1 pith:MYM7R7IF submitted 2025-02-12 physics.optics

Chunhui Yao , Jie Ma , Ningning Wang , Peng Bao , Wei Zhuo , Tao Zhang , Wanlu Zhang , Kangning Xu
show 7 more authors
Ting Yan Liang Ming Yuxiao Ye Tawfique Hasan Ian White Richard Penty Qixiang Cheng
This is my paper · ORCID
classification physics.optics
keywords opticalspectrometerconvolutiontheoremcircularMach-Zehnderinterferometersiliconnitridephotonicsnear-infraredspectroscopycomputationalnon-invasivebiomarkersensing
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

This paper introduces a new class of spectrometer in which a cascade of periodic spectral filters, scanned by proportionally tuned phases, performs circular convolution on the incident spectrum directly in the optical hardware. Because the recorded photocurrent sequence is the circular convolution of the input spectrum with the flipped system response, the spectrum is recovered by one division in Fourier space, $x[\nu] = \mathrm{IDFT}(\mathrm{DFT}(p[t])/\mathrm{DFT}(r_{\mathrm{sys}}^*[\nu]))$. The authors argue that this linear scheme escapes the usual bandwidth-resolution trade-off of miniaturized spectrometers: the bandwidth is set by the composite free spectral range, the resolution is set by the highest non-zero Fourier component of the system response and grows exponentially with the number of stages, and measurement noise enters linearly rather than being amplified by a nonlinear reconstruction algorithm. A packaged silicon-nitride demonstration reports a 2400 cm$^{-1}$ bandwidth, a 5.4 cm$^{-1}$ resolution, sub-0.4 s sampling, and roughly \$10 cost, and is used for solid-sample classification, concentration quantification, and non-invasive biomarker sensing. If the convolutional model holds under calibration, miniaturized spectrometers could move from coarse peak identification to metrological and wearable use.

What carries the argument

The load-bearing object is the circular convolution theorem applied in the spectral domain. Periodic spectral filters (unbalanced Mach-Zehnder interferometers with arm-length differences $\Delta L_i$) are cascaded so that their individual responses multiply into a single periodic mask whose period is the composite free spectral range, $\mathrm{FSR}_{\mathrm{composite}}=\mathrm{LCM}[\mathrm{FSR}_i]$. Proportional phase tuning of each stage shifts this mask by a common wavenumber step $\Delta\nu$, so the sequence of photodetector readings in time is the circular convolution of the incident spectrum with the flipped mask. The inverse operation is one complex division in the DFT domain, Eq. (5), and the resolution is read off from the highest non-zero Fourier component of the mask, Eq. (6). The same structure yields the paper's two scaling claims: additional cascaded stages add Fourier components approximately exponentially, and, because the mask is periodic, the core can operate in any cycle of the composite FSR, so bandwidth is not tied to the sharpness of each individual filter.

What would settle it

Launch a single narrow laser line at a known wavenumber, scan the phase vector over one full composite FSR, and compare the DFT of the measured $p[t]$ with the predicted product $\mathrm{DFT}(r_{\mathrm{sys}}^*[\nu])\,X[k]$; if the ratio $P[k]/(R_{\mathrm{sys}}^*[k])$ is not flat across $k$, or if recovered single-line spectra show ghost peaks or baseline ripple above the quoted noise, the rigid-shift convolution model is falsified.

Watch

Extended reading notes

Core claim

The central claim is that the convolution theorem can be transplanted from digital signal processing directly into the optical hardware of a spectrometer. A cascade of unbalanced Mach-Zehnder interferometers has an overlaid response $r_{\mathrm{sys}}(\nu,t)$ that is periodic with a composite free spectral range; when the per-stage phase shifts obey $\bmod(\Delta\varphi_i(t),2\pi) \propto \Delta L_i$, the whole waveform shifts rigidly in wavenumber, $r_{\mathrm{sys}}(\nu,t_0+\Delta t)=r_{\mathrm{sys}}(\nu+\Delta\nu,t_0)$. The detector output over one scan is therefore $p[t]=\sum_{\nu} x[\nu]\, r_{\mathrm{sys}}^*[(-\nu+\delta\nu\,t)_N]$, a circular convolution whose DFT satisfies $P[k]=R_{\mathrm{sys}}^*[k]X[k]$, so the input spectrum is recovered by Eq. (5), and the resolution is $\mathrm{FSR}_{\mathrm{composite}}/\max\{k \mid R_{\mathrm{sys}}^*[k]>0\}$. The paper reports experimental confirmation on a four-stage silicon-nitride chip: circular shifting in steps of 2.1 cm$^{-1}$, recovery of single through quad-peak and randomly shaped spectra with relative errors between 0.022 and 0.044, and a dual-peak resolution of 5.8 cm$^{-1}$ against the theoretical 5.4 cm$^{-1}$.

Load-bearing premise

The retrieval is exact only if the cascade's overlaid response shifts as a rigid waveform under applied phase, with a uniform wavenumber shift across the operating band and across every composite-FSR cycle; if group-index dispersion or unequal thermo-optic phase efficiency breaks that rigidity, Eq. (5) is approximate and the recovered spectra acquire a bias that calibration must remove.

Editorial extensions

If this is right

  • Bandwidth and resolution become independently engineerable: bandwidth is set by the composite free spectral range, while resolution is set by the number and FSR ratios of the cascaded stages.
  • Adding stages increases resolution roughly exponentially for a fixed composite FSR; the paper simulates a change from about 30 cm⁻¹ at three stages to below 0.5 cm⁻¹ at seven stages.
  • Noise is linearly imposed on the recovered spectrum, so standard low-pass filtering can suppress it, and thermal drift appears as a common circular shift that can be corrected algorithmically over at least -20 to 80 °C.
  • The same convolution core can operate inside any cycle of its composite FSR, so covering a wider band means adding more cycles; the demonstrated chip spans 5900 to 8300 cm⁻¹ with six superluminescent-diode sources.
  • A packaged microcontroller-based device acquires and processes one spectrum in under 0.4 s at roughly 10 USD, enabling the real-time classification, concentration, and biomarker measurements reported.

Reading between the lines

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

  • If rigid-shift calibration holds at other wavelengths, the same architecture should port to other platforms and bands, and a natural extension is demonstrating contiguous recovery across adjacent composite-FSR cycles with one continuous broadband source rather than six stitched SLDs.
  • The biomarker numbers are regression results on a finite participant set, not device-level guarantees; the paper itself notes that the glucose models are not yet transferable to new participants, so the clinical claims should be evaluated separately from the spectrometer's spectral-recovery claim.
  • A direct test of the convolution model's exactness is to scan a single narrow line across multiple composite-FSR cycles and check for aliasing or phase-dependent bias; the two chip variants with 420 and 840 cm⁻¹ composite FSR make this test feasible.
  • Because resolution scales with stage count, a similar package with more MZI stages could plausibly reach sub-cm⁻¹ resolution, provided the added calibration complexity and phase-shifter power do not erase the cost and speed advantages.
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 / 6 minor

Summary. The paper proposes a new class of miniaturized spectrometer, the convolutional spectrometer (ConvSpec), in which a cascade of phase-tunable periodic filters (unbalanced MZIs) is used to execute circular convolution in the spectral domain. The measured photodetector output p[t] is claimed to be the circular convolution of the input spectrum x[ν] with the system response r_sys[ν]; the input is recovered by inverse DFT after division by the DFT of the flipped system response, as in Eq. (5). A packaged SiN device with four MZI stages is reported to achieve a 2400 cm^-1 bandwidth, 5.4 cm^-1 resolution, sub-0.4 s sampling, and roughly $10 cost. The authors demonstrate retrieval of arbitrary spectra generated by a waveshaper, classification of solid samples (100% success), concentration measurements of solutions, and non-invasive sensing of skin moisture, blood alcohol, blood lactate, and blood glucose in human participants.

Significance. If the central convolution claim is sound, the paper introduces a genuinely new spectrometer category that trades the usual hardware complexity for a simple, linear computational reconstruction. The experimental work has notable strengths: the system response is calibrated independently with a commercial spectrum analyzer, test inputs are generated by an independent waveshaper, and the resolution is checked against resolved dual-peak spacing (5.8 cm^-1 versus a theoretical 5.4 cm^-1) rather than fitted. The device is fully packaged with a cost breakdown, and the authors provide public demo code. The application breadth, including biomarker sensing, is substantial. However, the central derivation contains a sign inconsistency, and the rigid-shift assumption that underpins the convolution model is not fully verified across the operational bandwidth; these issues must be resolved before the claims can be accepted as they stand.

major comments (3)
  1. [Eq. (3)-(5)] The two lines of Eq. (3) are not equivalent under the stated definition of r_sys* as the flipped sequence. With r_sys*[ν] = r_sys[−ν], the second line gives r_sys*[(−ν + δν t)_N] = r_sys[(ν − δν t)_N], which differs from the first line's r_sys[(ν + δν t)_N]. The standard circular convolution identity p[t] = Σ x[ν] r_sys[(t − ν)_N] leads to P[k] = X[k] R_sys[k]; as written, the DFT relation in Eq. (4) and the retrieval in Eq. (5) do not follow from the displayed p[t] expression. This is a load-bearing issue because Eq. (5) is the core mathematical claim of the paper. The authors should correct the indexing convention and verify that the convolution theorem is indeed being applied.
  2. [Methods, Eq. (10); main text after Eq. (5)] Equation (10) establishes the rigid-shift property only when the phase increments satisfy the proportionality condition with a common effective index n_eff across all stages and the operating band. The paper acknowledges that dispersion stretches FSR periods and states that this 'can be compensated by mathematically amending Eq. (5)', but no amended equation or quantitative estimate of the residual non-rigid shift error is provided. The retrieval experiments in Fig. 2 are performed within a single 420 cm^-1 composite FSR, whereas the full-band applications use a chip with a doubled FSR (840 cm^-1) and six SLDs. The residual shift error must be quantified relative to the shift step (2.1 cm^-1) and the resolution (5.4 cm^-1) for the full-band operation to support the convolutional retrieval claim.
  3. [Abstract; Fig. 2; Table 1] The abstract and Table 1 claim a 2400 cm^-1 bandwidth and list the convolutional spectrometer's bandwidth as 'unlimited' (FSR per operation). The experimental validation of retrieval accuracy (Fig. 2f-j) is limited to a single composite FSR (6250-6670 cm^-1). The full-band spectra in the applications are obtained with the FSR-doubled chip, but no retrieval-accuracy experiment using known input spectra is reported for that configuration. The 'unlimited' bandwidth is a theoretical property of circular convolution, not a demonstrated feature. The authors should either provide full-band retrieval validation or clearly label the bandwidth claim as a theoretical property.
minor comments (6)
  1. [Eq. (3)] The summation limits are written as ν from ν_0 to ν_{N-1}; it would be clearer to use integer indices 0 to N-1.
  2. [Eq. (9)] The expression '1/DFR1' appears to be a typo for '1/FSR1'.
  3. [Table 1 footnote] The phrase 'respectively, respectively' is duplicated in the footnote.
  4. [Methods: Data processing and modeling] The sentence 'we also we preprocess' contains a grammatical error and should be corrected.
  5. [Fig. 3 and Supplementary Section 11] The classification results are based on 80 repeated measurements per sample with a random 2:1 split; because repeated spectra from the same sample are likely highly correlated, the test accuracy may be optimistic. Reporting leave-one-sample-out or leave-one-batch-out cross-validation would better support the 100% success rate claim.
  6. [Fig. 4 and Methods: Data processing and modeling] The biomarker models are evaluated on a single SPXY split; reporting repeated random splits or bootstrap confidence intervals would better quantify the uncertainty in the reported MAE/RMSE values.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the convolution retrieval is calibrated against an external spectrum analyzer and validated with independently generated test spectra.

full rationale

The paper's derivation chain is self-contained and does not reduce to its inputs. The retrieval formula (Eq. 5) follows from applying the textbook discrete circular convolution theorem (ref 30) to Eq. 3, and the kernel r_sys is measured independently with a commercial benchtop spectrum analyzer rather than being defined by the retrieved spectra. Eq. 10 is an algebraic consequence of Eqs. 7-9: if the phase increments obey mod{Δφ_i} ∝ ΔL_i, each MZI term shifts by the same Δν = Δφ_i/(2π n_eff ΔL_i), so the product shifts rigidly; this is a stated operating condition, not an assumption that smuggles in the result. The resolution figure (Eq. 6) is a definitional relation based on the measured kernel's DFT support, but the dual-peak resolving-power test in Fig. 2h is a separate experimental measurement that confirms the definition rather than being fitted to it. The waveshaper-generated inputs and the external reference spectra for solids, liquids, and biomarkers provide independent validation. The same-author citations (refs 15, 20, 25, 28) are used only as examples of prior reconstructive spectrometer work and background claims, not to justify the convolution theorem or to forbid alternative designs; hence they are not load-bearing. Overall, no circular step was found.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central spectrometer math has no fitted constants: the convolution model is derived from the MZI transfer function and the convolution theorem, with r_sys calibrated externally. The free parameters are calibration coefficients and ML hyperparameters used in the application demonstrations. The axioms are the standard physics and math assumptions: product of stage responses, rigid phase-proportional shift, single-FSR confinement, and DFT validity. No new physical entities are introduced.

free parameters (2)
  • Per-stage thermo-optic phase calibration coefficients = Not reported
    Eq. (9) requires each MZI stage's phase shift to be proportional to its arm length difference; the paper states each stage is calibrated (Supplementary Fig. S7) but gives no values or residuals. The rigid-shift property of Eq. (10), and therefore the convolution model, depends on these coefficients.
  • Machine-learning hyperparameters (kNN k, SVR kernel parameters, DFNN architecture and dropout) = Tuned on training data; values not listed
    The reported classification, concentration, and biomarker accuracies are produced by these models; they are fit to reference data and are not outputs of the convolution theorem.
assumptions (4)
  • domain assumption Cascaded MZI transmission equals the product of individual stage responses, each modeled by Eq. (7) with a single cosine term.
    Used in Eqs. (8) and (10); ignores higher-order coupling and fabrication deviations.
  • domain assumption Proportional phase modulation produces a rigid spectral shift of the overlaid response, valid across the operational band.
    Eqs. (9)-(10) require dispersionless n_eff and exact proportionality; the paper defers dispersion compensation to Supplementary Section 3.
  • domain assumption The incident spectrum is confined to one composite FSR so circular convolution is the correct inversion model.
    Eq. (2) integrates only over one FSR_composite; energy outside aliases into the recovered spectrum.
  • standard math DFT and IDFT, together with the convolution theorem, hold for the discretized sequences.
    Eqs. (4)-(5) rest on this textbook result (Ref. 30).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Optical Convolutional Spectrometer." pith.science (2026). https://pith.science/paper/MYM7R7IF

@misc{pith2026250208749,
  author       = {Pith},
  title        = {Pith review of: Optical Convolutional Spectrometer},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MYM7R7IF}},
  note         = {Machine review of arXiv:2502.08749}
}
abstract

Optical spectrometers are fundamental across numerous disciplines in science and technology. However, miniaturized versions, while essential for in situ measurements, are often restricted to coarse identification of signature peaks and inadequate for metrological purposes. Here, we introduce a new class of spectrometer, leveraging the convolution theorem as its mathematical foundation. Our convolutional spectrometer offers unmatched performance for miniaturized systems and distinct structural and computational simplicity, featuring a centimeter-scale footprint for the fully packaged unit, low cost (~$10) and a 2400 cm-1 (approximately 500 nm) bandwidth. We achieve excellent precision in resolving complex spectra with sub-second sampling and processing time, demonstrating a wide range of applications from industrial and agricultural analysis to healthcare monitoring. Specifically, our spectrometer system classifies diverse solid samples, including plastics, pharmaceuticals, coffee, flour and tea, with 100% success rate, and quantifies concentrations of aqueous and organic solutions with detection accuracy surpassing commercial benchtop spectrometers. We also realize the non-invasive sensing of human biomarkers, such as skin moisture (mean absolute error; MAE = 2.49%), blood alcohol (1.70 mg/dL), blood lactate (0.81 mmol/L), and blood glucose (0.36 mmol/L), highlighting the potential of this new class of spectrometers for low-cost, high-precision, portable/wearable spectral metrology.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. 3D surface profiling via photonic integrated geometric sensor

    physics.optics 2025-06 conditional novelty 6.0 of 10

    A photonic integrated circuit with 729 spectral sampling states, decoded by a 1D-CNN, predicts surface thickness with about 18.5 um average error on smooth and structured samples.

Reference graph

Works this paper leans on

46 extracted references · 44 canonical work pages · cited by 1 Pith paper

  1. [1]

    Department of Engineering, University of Cambridge, UK

    work page

  2. [2]

    GlitterinTech Limited, Xuzhou, China

    work page

  3. [3]

    Abstract Optical spectrometers are fundamental across numerous disciplines in science and technology

    University of Bath, Bath, UK *qc223@cam.ac.uk; †These two authors contribute equally to this work. Abstract Optical spectrometers are fundamental across numerous disciplines in science and technology. However, miniaturized versions, while essential for in situ measurements, are often restricted to coarse identification of signature peaks and inadequate fo...

    work page

  4. [4]

    Crocombe, R. A. Portable spectroscopy. Applied Spectroscopy 72, 1701–1751 (2018)

    work page 2018

  5. [5]

    & Bawendi, M

    Bao, J. & Bawendi, M. G. A colloidal quantum dot spectrometer. Nature 523, 67–70 (2015)

    work page 2015

  6. [6]

    Yang, Z. et al. Single-nanowire spectrometers. Science 365, 1017–1020 (2019)

    work page 2019

  7. [7]

    Yoon, H. H. et al. Miniaturized spectrometers with a tunable van der Waals junction. Science 378, 296–299 (2022)

    work page 2022

  8. [8]

    P., Mattley, Y

    Bacon, C. P., Mattley, Y . & DeFrece, R. Miniature spectroscopic instrumentation: applications to biology and chemistry. Review of Scientific instruments 75, 1–16 (2004)

    work page 2004

Show all 46 references
  1. [9]

    & Hasan, T

    Yang, Z., Albrow-Owen, T., Cai, W. & Hasan, T. Miniaturization of optical spectrometers. Science 371, eabe0722 (2021)

    2021

  2. [10]

    Applications of Infrared Spectroscopy in Biochemistry, Biology, and Medicine

    Parker, F. Applications of Infrared Spectroscopy in Biochemistry, Biology, and Medicine. (Springer Science & Business Media, 2012)

    2012

  3. [11]

    Near infrared spectroscopy: A mature analytical technique with new perspectives – A review

    Pasquini, C. Near infrared spectroscopy: A mature analytical technique with new perspectives – A review. Analytica Chimica Acta 1026, 8–36 (2018)

    2018

  4. [12]

    Martin, F. L. et al. Distinguishing cell types or populations based on the computational analysis of their infrared spectra. Nat Protoc 5, 1748–1760 (2010)

    2010

  5. [13]

    Baker, M. J. et al. Using Fourier transform IR spectroscopy to analyze biological materials. Nat Protoc 9, 1771–1791 (2014)

    2014

  6. [14]

    B., Grabska, J

    Beć, K. B., Grabska, J. & Huck, C. W. Near-Infrared Spectroscopy in Bio-Applications. Molecules 25, 2948 (2020)

    2020

  7. [15]

    Chitnis, D. et al. Towards a wearable near infrared spectroscopic probe for monitoring concentrations of multiple chromophores in biological tissue in vivo. Review of Scientific Instruments 87, 065112 (2016)

    2016

  8. [16]

    Li, A. et al. Advances in cost-effective integrated spectrometers. Light Sci Appl 11, 174 (2022)

    2022

  9. [17]

    M., Delbeck, S

    Heise, H. M., Delbeck, S. & Marbach, R. Noninvasive Monitoring of Glucose Using Near-Infrared Reflection Spectroscopy of Skin—Constraints and Effective Novel Strategy in Multivariate Calibration. Biosensors 11, 64 (2021)

    2021

  10. [18]

    Lafrance, D., Lands, L. C. & Burns, D. H. Measurement of lactate in whole human blood with near- infrared transmission spectroscopy. Talanta 60, 635–641 (2003)

    2003

  11. [19]

    Yao, C. et al. Chip-scale sensor for spectroscopic metrology. Nat Commun 15, 10305 (2024)

    2024

  12. [20]

    Yao, C. et al. Integrated reconstructive spectrometer with programmable photonic circuits. Nat Commun 14, 6376 (2023)

    2023

  13. [21]

    Shekhar, S. et al. Roadmapping the next generation of silicon photonics. Nat Commun 15, 751 (2024)

    2024

  14. [22]

    Compact spectrometer based on a disordered photonic chip

    Redding, B. Compact spectrometer based on a disordered photonic chip. Nat Photonics 7, 6 (2013)

    2013

  15. [23]

    Kita, D. M. et al. High-performance and scalable on -chip digital Fourier transform spectroscopy. Nat Commun 9, 4405 (2018)

    2018

  16. [24]

    H., Sun, H., Chew, J

    Guan, Q., Lim, Z. H., Sun, H., Chew, J. X. Y . & Zhou, G. Review of Miniaturized Computational Spectrometers. Sensors 23, 8768 (2023)

    2023

  17. [25]

    Cheben, P. et al. A high-resolution silicon -on-insulator arrayed waveguide grating microspectrometer with sub-micrometer aperture waveguides. Optics express 15, 2299–2306 (2007)

    2007

  18. [26]

    Zhang, L. et al. Ultrahigh-resolution on-chip spectrometer with silicon photonic resonators. Opto - Electronic Advances 5, 210100–1 (2022)

    2022

  19. [27]

    Montesinos-Ballester, M. et al. On-chip Fourier -transform spectrometer based on spatial heterodyning tuned by thermo-optic effect. Sci Rep 9, 14633 (2019)

    2019

  20. [28]

    Yao, C. et al. Broadband picometer -scale resolution on -chip spectrometer with reconfigurable photonics. Light Sci Appl 12, 156 (2023)

    2023

  21. [29]

    Yao, C. et al. Benchmarking Reconstructive Spectrometer with Multiresonant Cavities. ACS Photonics (2024) doi:10.1021/acsphotonics.4c00915

    2024 doi

  22. [30]

    & Tsang, H

    Xu, H., Qin, Y ., Hu, G. & Tsang, H. K. Cavity -enhanced scalable integrated temporal random - speckle spectrometry. Optica, 10, 1177–1188 (2023)

    2023

  23. [31]

    Wan, Y ., Fan, X. & He, Z. Review on Speckle-Based Spectrum Analyzer. Photonic Sens 11, 187– 202 (2021)

    2021

  24. [32]

    & Kuchta, K

    Alassali, A., Fiore, S. & Kuchta, K. Assessment of plastic waste materials degradation through near infrared spectroscopy. Waste Management 82, 71–81 (2018)

    2018

  25. [33]

    & Liu, S

    Chen, C., Gu, H. & Liu, S. Ultra-simplified diffraction-based computational spectrometer. Light Sci Appl 13, 9 (2024)

    2024

  26. [34]

    Digital Signal Processing Algorithms: Number Theory, Convolution, Fast Fourier Transforms, and Applications

    Krishna, H. Digital Signal Processing Algorithms: Number Theory, Convolution, Fast Fourier Transforms, and Applications. (Routledge, New York, 2017)

    2017

  27. [35]

    Vaseghi, S. V . Advanced Digital Signal Processing and Noise Reduction. (John Wiley & Sons, 2008)

    2008

  28. [36]

    Ge, Q. et al. Evaluation and Validation on Sensitivity of Near-Infrared Diffuse Reflectance in Non- Invasive Human Blood Glucose Measurement. Sensors 24, 5879 (2024)

    2024

  29. [37]

    Infrared Spectroscopy for Food Quality Analysis and Control

    Sun, D.-W. Infrared Spectroscopy for Food Quality Analysis and Control. (Academic Press, 2009)

    2009

  30. [38]

    & Bello, V

    Bodo, E., Merlo, S. & Bello, V . Spectral Fingerprint Investigation in the near Infra -Red to Distinguish Harmful Ethylene Glycol from Isopropanol in a Microchannel. Sensors 22, 459 (2022)

    2022

  31. [39]

    & Al-Naib, I

    Alsunaidi, B., Althobaiti, M., Tamal, M., Albaker, W. & Al-Naib, I. A Review of Non -Invasive Optical Systems for Continuous Blood Glucose Monitoring. Sensors 21, 6820 (2021)

    2021

  32. [40]

    Villena Gonzales, W., Mobashsher, A. T. & Abbosh, A. The Progress of Glucose Monitoring —A Review of Invasive to Minimally and Non-Invasive Techniques, Devices and Sensors. Sensors 19, 800 (2019)

    2019

  33. [41]

    We record the spectra from 27 participants undergoing a standard oral glucose tolerance test (OGTT; see methods)13, as shown in Supplementary Information Fig. S14d. The resulting SVR model achieves an MAE and RMSE of 0.363 mmol L-1 and 0.510 mmol L-1, respectively (Fig. 4j). A...

  34. [42]

    Heise, H. M. Non-Invasive Monitoring of Metabolites Using Near Infrared Spectroscopy: State of the Art. Hormone and Metabolic Research 28, 527–534 (2007)

    2007

  35. [43]

    Huber, M. et al. Stability of person-specific blood-based infrared molecular fingerprints opens up prospects for health monitoring. Nat Commun 12, 1511 (2021)

    2021

  36. [44]

    Khalil, O. S. Non-Invasive Glucose Measurement Technologies: An Update from 1999 to the Dawn of the New Millennium. Diabetes Technology & Therapeutics 6, 660–697 (2004)

    2004

  37. [45]

    J., Chen, C.-J

    Tang, L., Chang, S. J., Chen, C.-J. & Liu, J.-T. Non-Invasive Blood Glucose Monitoring Technology: A Review. Sensors 20, 6925 (2020)

    2020

  38. [46]

    & Karimzadeh, R

    Saghaei, H., Elyasi, P. & Karimzadeh, R. Design, fabrication, and characterization of Mach – Zehnder interferometers. Photonics and Nanostructures - Fundamentals and Applications 37, 100733 (2019). Acknowledgement This research was mainly supported by GlitterinTech Limited, bu...

    2019

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.