Exploring the role of sample thickness for hyperspectral microscopy tissue discrimination through Monte Carlo simulations

1); Amsterdam; Aquaculture Research (Nofima); Behdad Dashtbozorg (2); Department of Surgery; Fisheries; Gustavo M. Callico (1) ((1) Institute For Applied Microelectronics (Iuma); Henricus J.c.m. Sterenborg (2); Laura Quintana-quintana (1); Mark Witteveen (2)

REVIEW 3 major objections 5 minor 45 references

Exploring the role of sample thickness for hyperspectral microscopy tissue discrimination through Monte Carlo simulations

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Standard 5-micron slices carry almost no spectral contrast for unstained hyperspectral microscopy.

desk verdict Useful literature audit plus a plausible but model-bound thickness prediction; the quantitative optima are not yet trustworthy. read the letter →

arxiv 2507.21675 v1 pith:XINFKJMJ submitted 2025-07-29 physics.optics physics.bio-ph

Laura Quintana-quintana (1) , Mark Witteveen (2) , Behdad Dashtbozorg (2) , Samuel Ortega (3 , 1) , Theo J.m. Ruers (2) , Henricus J.c.m. Sterenborg (2) , Gustavo M. Callico (1) ((1) Institute For Applied Microelectronics (Iuma)

show 10 more authors

University Of Las Palmas De Gran Canaria (Ulpgc) Spain (2) Image-guided Surgery Department Of Surgery The Netherlands Cancer Institute Amsterdam Netherlands (3) Norwegian Institute Of Food Fisheries Aquaculture Research (Nofima) Troms{o} Norway)

This is my paper · ORCID

classification physics.opticsphysics.bio-ph

keywords hyperspectralmicroscopymultispectralimagingMonteCarlosimulationtissuethicknessunstainedspectraldiscriminationlabel-freehistologylighttransport

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 argues that in unstained multispectral and hyperspectral (MS/HS) microscopy, the thickness of a tissue section is not a neutral detail but a first-order determinant of whether normal and tumor tissue can be told apart. Using GPU-accelerated Monte Carlo simulations of breast, colorectal, liver, and lung tissue, it finds that the conventional $2$--$10\,\mu$m sections transmit light almost unmodified, so spectra from different pathologies are statistically indistinguishable. Discrimination improves as thickness rises, with an optimum near $500\,\mu$m for several tissues (and $1000\,\mu$m for breast), at the cost of mean transmittance dropping below half. Since 40% of published MS/HS microscopy studies do not even report thickness and 98% of those that do use $2$--$10\,\mu$m sections, these results imply the field may currently be operating close to the sensitivity floor for label-free contrast.

What carries the argument

The load-bearing instrument is a GPU-accelerated Monte Carlo photon-transport simulation (MCX) configured as a transmittance microscope with K\"ohler illumination over $400$--$1000$ nm in $105$ bands. Tissue is modeled as a homogeneous single-layer slab with a fixed refractive index and a two-term Henyey-Greenstein phase function; absorption comes from a weighted sum of oxy- and deoxy-hemoglobin, water, and fat, and scattering from a Rayleigh--Mie combination. Voxel size is coarsened from $1\,\mu$m to $100\,\mu$m as thickness grows from $5$ to $2000\,\mu$m. The raw output is flat-field corrected, augmented with wavelength miscalibration, intensity fluctuations, and 35 dB Gaussian noise, and then scored with five spectral metrics (Euclidean distance, SAM, NS3, SID, SID-SAM); the maximum threshold-based accuracy separating inter-class from intra-class distances defines the discriminative power of each thickness.

What would settle it

Measure unstained serial sections from the same tissue block at $5$, $20$, $50$, $200$, and $500\,\mu$m on a single hyperspectral microscope, apply identical preprocessing and a classification model, and test two predictions: first, that $5\,\mu$m accuracy sits near chance (0.5) while $200$--$500\,\mu$m improves it; second, that the ranking of thicknesses matches the simulated order for that tissue. A single tissue in which $5\,\mu$m outperforms $50\,\mu$m, or in which the simulated optimum does not correspond to the experimental optimum, would refute the paper's claim that thickness is the dominant driver of unstained spectral contrast.

Watch

Extended reading notes

Core claim

The central claim is that, for unstained tissue, spectral discrimination between normal and lesioned samples in MS/HS microscopy is a direct function of section thickness, and the conventional histology thickness of $2$--$10\,\mu$m is near the worst possible choice. In the authors' homogeneous-slab model with refractive index $1.35$ and bulk optical properties from the literature, a $5\,\mu$m section produces flat transmittance spectra with no absorption or scattering signature, and the best achievable classification accuracy between normal and tumor is about $0.5$ for every metric tested. At $50\,\mu$m differences begin to appear; at $200$--$500\,\mu$m the separation between inter-class and intra-class spectral distances is largest, although transmittance falls below $0.5$; and beyond roughly $1000$--$2000\,\mu$m attenuation flattens the spectra again. The paper concludes that thickness should be optimized per tissue type and treated as a tunable sample-preparation parameter for label-free MS/HS microscopy.

Load-bearing premise

The load-bearing premise is that a tissue section behaves optically like a homogeneous slab with uniform refractive index and bulk optical properties from in vivo measurements, so a $5\,\mu$m slice is simply too short to scatter or absorb light; if real thin sections retain structural contrast from nuclei, collagen, or refractive-index gradients, the zero-contrast result would be weakened and the optimal thickness could shift.

Editorial extensions

If this is right

If the result holds, unstained MS/HS microscopy using $2$--$10\,\mu$m sections is working near the sensitivity floor, and many published classifications may rest on residual staining, fixation, or system artifacts rather than intrinsic tissue spectra.
Thickness becomes a tunable preparation parameter: cutting sections of several hundred micrometers could improve label-free pathology, provided illumination and detector sensitivity compensate for the >50% transmittance loss.
The optimal thickness is tissue-dependent, so protocols should be organ-specific rather than relying on a single standard section thickness.
The same simulation pipeline can be reused to screen other tissue types, wavelength ranges, or sensor configurations for their best thickness before committing to experiments.
Reporting thickness in future MS/HS microscopy studies becomes a required experimental parameter, not an optional detail.

Reading between the lines

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

Because the model treats tissue as a homogeneous slab with no cellular or nuclear structure, the '5 $\mu$m has zero contrast' result is partly a modeling artifact: real unstained thin sections contain refractive-index fluctuations and organelles that may add contrast the simulation cannot produce.
A direct experimental check would be to cut serial sections from one block at 5, 50, 200, and 500 $\mu$m, image the same unstained tissue on the same HS microscope, and see whether classification accuracy follows the simulated ranking.
The optimal-thickness curve could shift in the near-infrared, where scattering is weaker; the paper's 400--1000 nm range leaves the NIR extrapolation untested.
If thin sections truly lack intrinsic contrast, then existing unstained MS/HS microscopy datasets could be re-analyzed to see whether reported accuracies depend more on preparation artifacts than on endogenous chromophores, a testable consequence for published studies.

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.

Desk Editor's Note

read the letter

Two things worth knowing. First, the literature audit is the most useful part: 40% of MS/HS microscopy papers don't report section thickness, and of those that do, 98% use 2–10 µm. That gives the paper a real empirical hook. Second, the simulation result—thickening unstained sections from 5 µm to hundreds of µm improves spectral discrimination before attenuation dominates—is plausible, but it is a model prediction, not an established tissue fact.

The paper does real work. It uses a standard GPU Monte Carlo code (MCX), runs 67,200 simulations with 10^8 photons each, and takes optical properties from independent literature sources. The multi-metric evaluation with majority voting is a reasonable way to avoid metric bias. The Discussion and Conclusions are also honest: the authors explicitly acknowledge the homogeneous-slab limitation and call for empirical validation. That is more than many simulation papers do.

The soft spots, in order of importance. First, the homogeneous single-layer slab with n=1.35 almost guarantees that 5 µm sections transmit unmodified light. Real thin sections contain nuclei, organelles, collagen, and refractive-index discontinuities that can produce wavelength-dependent contrast even below one mean free path. The authors do acknowledge this in Section 5, but the Abstract’s “thin samples reduce differentiation” is stated without that qualification. So the ~500 µm optimum should be read as conditional on structure-free tissue. Second, the discrimination metric optimizes the decision threshold on the same data (Section 3.6). That inflates absolute accuracy, and without confidence intervals or cross-validation the specific optimum thicknesses are shaky; the general trend is probably robust, the exact numbers are not. Third, the voxel schedule changes with thickness (1 µm for thin samples, 100 µm for thick ones), a minor confound even for homogeneous tissue. Fourth, data and code are only “available upon request.” For a pure simulation paper, shipping code and data would make the results independently checkable. Citation patterns are fine; self-citations are to the authors’ instrument papers and are appropriate.

Who is this for: anyone working on label-free MS/HS microscopy and sample preparation. It is a useful benchmark and a reminder to report thickness. It deserves a serious referee, but I would send it back for revision: add a structured-tissue simulation, even a simple two-compartment model, or soften the general claim; fix the threshold-overfitting issue with held-out thresholds or proper cross-validation; and release the code and data.

Referee Report

3 major / 5 minor

Summary. The manuscript combines a literature survey of tissue thickness in multispectral/hyperspectral microscopy with Monte Carlo light-transport simulations of unstained normal versus lesioned breast, colorectal, liver, and lung tissue at eight thicknesses. It computes five spectral dissimilarity metrics and uses the accuracy at an optimized decision threshold to claim that thin (5 µm) sections provide no label-free discrimination, that contrast improves with thickness, and that the optimum discrimination occurs around 200–1000 µm, with 500 µm highlighted, at the cost of reduced transmittance.

Significance. The paper addresses a genuinely under-reported experimental variable, and the literature update is a useful contribution. The simulation pipeline is described in enough detail to be reproducible, with explicit simulation counts, GPU-based MCX, and a documented data-augmentation procedure. If the central trend survived a structured-tissue model and an unbiased evaluation protocol, it would give practical guidance for label-free MS/HS microscopy and motivate rethinking standard 2–10 µm sections. However, as presented the quantitative claims are conditional on model homogeneity and on an evaluation method that selects the decision threshold on the same data used to report accuracy.

major comments (3)

[Section 3.6, Eq. (14)] The accuracy metric is obtained by searching over all possible thresholds on the same data used to report performance, and the reported value is the maximum accuracy over those thresholds. This is an in-sample optimization, so the discriminative-power values in Fig. 8 are upper bounds rather than unbiased estimates. With only ten random draws per tissue/pathology combination (Section 3.3.2), the intra- and inter-class distance histograms are also noisy, and no confidence intervals are provided. Please select the threshold on a training subset and evaluate on a held-out subset, or report AUC with bootstrap confidence intervals.
[Section 3.3.1 and Section 5] The simulated volume is a homogeneous single-layer slab with uniform refractive index n=1.35 and bulk optical properties. At 5 µm the optical depth is only a small fraction of a mean free path, so near-unity transmittance and chance-level accuracy at 5 µm are largely enforced by the absence of sub-resolution structure (nuclei, organelles, refractive-index fluctuations) rather than by a measured property of real thin sections. The limitation is acknowledged in Section 5, but the Abstract and Conclusions present the thickness trend and the roughly 500 µm optimum as general guidance. Please add a structured-tissue simulation (for example, a two-component nuclei/ECM model or measured refractive-index maps) or explicitly reframe the claim as applying only to homogeneous tissue until empirical validation is available.
[Section 3.3.1] The voxel size is coupled to thickness: 1 µm for 5/20 µm, 10 µm for 50–200 µm, and 100 µm for 500–2000 µm. This couples the thickness axis to spatial discretization, and no resolution-convergence check is reported. Without evidence that the transmittance values and metric trends are insensitive to voxel size (or a consistent voxelization), the quantitative comparison across thicknesses is not fully supported. Please include a convergence study at fixed thickness or re-run the key cases with matched voxel sizes.

minor comments (5)

[Section 1] The definition of the mean free path as l_s = 1/μa + 1/μs is not the standard definition; the usual quantity is l = 1/(μa+μs) or the scattering mean free path l_s = 1/μs. Please correct or justify this expression.
[Section 5] The Discussion states that 60% of documents did not report thickness, while the Abstract and Section 2 state 40%; one of these statements is inconsistent and should be corrected.
[Section 3.2, Eq. (2)] Equation (2) appears to contain a typographical error: the first absorption term is rendered as B B μ_a,oxy rather than B×S×μ_a,oxy. Please check the formula and its notation.
[Fig. 8] Fig. 8 would benefit from error bars or shaded confidence bands; as plotted, the reader cannot assess the variability of the accuracy estimates across the ten tissue draws.
[Section 3.4] The data-augmentation parameters (SNR = 35 dB, wavelength shift ±4.8 nm, intensity fluctuations ±5%) are presented without justification or sensitivity analysis; a sentence explaining how these values were chosen would help.

Circularity Check

1 steps flagged · score 3.0 of 10

Forward MC simulation is essentially self-contained; one in-sample threshold-optimization step makes the reported per-thickness 'discriminative power' partially self-referential.

fitted input called prediction [Section 3.6, Quantitative Thickness Evaluation (after Eq. 14; see also Fig. 6)]
"The accuracy was calculated across all possible 𝑖𝑖ℎs for a single set of histograms, covering the lowest to the highest obtained metric scores, with an increment of one bin at a time. The highest accuracy value obtained across all thresholds was chosen to be the discriminative power of that thickness."

The 'discriminative power' reported for each thickness is not an independent estimate: the threshold is fit to the very same inter-class and intra-class distance histograms on which accuracy is then computed, and the maximum over all thresholds is retained. No held-out split, cross-validation, or distributional model is used. Therefore the accuracy values are in-sample training maxima, and comparing these maxima across thicknesses to select the 'optimal discrimination thickness' is partly self-referential: a thickness can appear best because a post hoc threshold happens to separate that particular simulated dataset.

full rationale

The paper's core derivation is a forward Monte Carlo light-transport simulation. Tissue absorption and scattering parameters are taken from independent literature sources, the simulation propagates photons through a defined slab, and the resulting transmittance spectra are compared with standard spectral metrics. No parameter is fitted to the target claim, and the 'thin samples reduce differentiation' trend is a direct consequence of the simulated optical depth: at 5 um the Beer-Lambert attenuation is minimal, so the spectra are nearly featureless regardless of tissue type. This is a legitimate forward result under the stated assumptions rather than a circular one. The homogeneity of the simulated volume is an important modeling limitation for external validity, but it is not circularity: the paper explicitly acknowledges that structural heterogeneity is not modeled, and this affects whether the conclusion transfers to real thin sections, not whether the simulation's output follows from its inputs. The one genuine circularity-adjacent issue is in Section 3.6: the accuracy used as 'discriminative power' is obtained by choosing the threshold that maximizes accuracy on the same data. This makes the quantitative discrimination scores partially self-referential. Citations to the authors' previous work provide review methodology and microscope configuration, but they are not load-bearing for the central transport result; the optical properties come from external experimental studies. Overall, the qualitative central claim is independently grounded, while the quantitative optimal-thickness comparison should be interpreted with caution. Score 3 reflects this partial, non-central circularity rather than a fully circular derivation.

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

The central claim rests on literature-derived optical properties, a homogeneous slab geometry, MCX numerical accuracy, and hand-chosen simulation and evaluation parameters. The most important non-validated input is the homogeneous tissue model with fixed refractive index; it is an ad hoc simplification that may determine the thin-sample result. No new entities are invented.

free parameters (4)

Discrimination threshold (th) = Varies per tissue, thickness, and metric (e.g., Euclidean distance threshold ~0.05 at 20 µm and ~0.07 at 200 µm for…
The reported accuracy is the maximum over all thresholds for the same data (Section 3.6); it is a fitted quantity, not an out-of-sample prediction.
Voxel size schedule = 1 µm for 5-20 µm, 10 µm for 50-200 µm, and 100 µm for 500-2000 µm thicknesses
Chosen by hand in Section 3.3.1 to keep voxel counts manageable; the 100 µm voxels for thick samples coarsen the simulation and may affect transmittance estimates.
Number of random tissue draws per class = 10
Section 3.3.2; the small sample drives histogram noise and threshold overfitting.
Data augmentation hyperparameters = Wavelength shift up to ±4.8 nm, intensity change up to ±5% of mean, SNR 35 dB, 521 bands after interpolation
Hand-chosen following reference [46] to mimic real sensors; the absolute discrimination levels depend on these choices.

assumptions (4)

domain assumption The optical properties of normal and lesioned tissues (absorption from chromophore concentrations, scattering from Rayleigh/Mie model, two-term Henyey-Greenstein phase function, refractive index 1.35) from cited literature are accurate for the simulated scenarios.
Enters Section 3.2 and Table 1; if these coefficients are wrong (especially for ex vivo processed tissue), the simulated spectra and discrimination results change.
ad hoc to paper Tissue can be modeled as a homogeneous single-layer slab without morphological heterogeneity.
Section 3.3.1; this simplification removes structural contrast and may force the 5 µm result.
standard math MCX (Monte Carlo eXtreme) correctly solves the light transport equation for the described geometry and voxelization.
Section 3.3.5; the simulation results depend on the numerical accuracy of MCX.
domain assumption The literature search and eligibility criteria of Ortega et al. [2] are appropriate for the thickness survey.
Section 2; the 40% non-reporting and 98% 2-10 µm statistics depend on this.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Exploring the role of sample thickness for hyperspectral microscopy tissue discrimination through Monte Carlo simulations." pith.science (2026). https://pith.science/paper/XINFKJMJ

@misc{pith2026250721675,
  author       = {Pith},
  title        = {Pith review of: Exploring the role of sample thickness for hyperspectral microscopy tissue discrimination through Monte Carlo simulations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XINFKJMJ}},
  note         = {Machine review of arXiv:2507.21675}
}

read the original abstract

Recent advancements in multispectral (MS) and hyperspectral (HS) microscopy have focused on sensor and system improvements, yet sample processing remains overlooked. We conducted an analysis of the literature, revealing that 40 percent of studies do not report sample thickness. Among those that did report it, the vast majority, 98 percent, used 2 to 10 micrometer samples. This study investigates the impact of unstained sample thickness on MS/HS image quality through light transport simulations. Monte Carlo simulations were conducted on various tissue types (i.e., breast, colorectal, liver, and lung). The simulations revealed that thin samples reduce tissue differentiation, while higher thicknesses (approximately 500 micrometers) improve discrimination, though at the cost of reduced light intensity. These findings highlight the need to study and optimize sample thickness for enhanced tissue characterization and diagnostic accuracy in MS/HS microscopy.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

45 extracted references · 45 canonical work pages

[1]

Medical hyperspectral imaging: a review,

G. Lu and B. Fei, "Medical hyperspectral imaging: a review," J Biomed Opt 19(1), 010901 (2014)

work page 2014
[2]

Hyperspectral and multispectral imaging in digital and computational pathology: a systematic review,

S. Ortega, M. Halicek, H. Fabelo, G. M. Callico, and B. Fei, "Hyperspectral and multispectral imaging in digital and computational pathology: a systematic review," Biomed Opt Express 11(6), 3195 (2020)

work page 2020
[4]

Novel Methodology for Alzheimer’s Disease Biomarker Identification in Plasma using Hyperspectral Microscopy,

H. Fabelo, R. Leon, S. Ortega, F. Balea-Fernandez, C. Bilbao, G. M. Callico, and A. Wagner, "Novel Methodology for Alzheimer’s Disease Biomarker Identification in Plasma using Hyperspectral Microscopy," in 2020 XXXV Conference on Design of Circuits and Integrated Systems (DCIS) (2020)

work page 2020
[5]

Automatic detection of head and neck squamous cell carcinoma on histologic slides using hyperspectral microscopic imaging,

L. Ma, J. V. Little, A. Y. Chen, L. Myers, B. D. Sumer, and B. Fei, "Automatic detection of head and neck squamous cell carcinoma on histologic slides using hyperspectral microscopic imaging," J Biomed Opt 27(04), (2022)

work page 2022
[6]

Leukocyte classification based on spatial and spectral features of microscopic hyperspectral images,

Y. Duan, J. Wang, M. Hu, M. Zhou, Q. Li, L. Sun, S. Qiu, and Y. Wang, "Leukocyte classification based on spatial and spectral features of microscopic hyperspectral images," Opt Laser Technol 112, 530–538 (2019)

work page 2019
[7]

Hyperspectral microscopy- applications of hyperspectral imaging techniques in different fields of science: a review of recent advances,

K. S. Banu, M. Lerma, S. U. Ahmed, and J. L. Gardea-Torresdey, "Hyperspectral microscopy- applications of hyperspectral imaging techniques in different fields of science: a review of recent advances," Appl Spectrosc Rev 1–24 (2023)

work page 2023
[8]

Hyperspectral Push- Broom Microscope Development and Characterization,

S. Ortega, R. Guerra, M. Diaz, H. Fabelo, S. Lopez, G. M. Callico, and R. Sarmiento, "Hyperspectral Push- Broom Microscope Development and Characterization," IEEE Access 7, 122473–122491 (2019)

work page 2019
[9]

Design and Validation of a Custom-Made Hyperspectral Microscope Imaging System for Biomedical Applications,

J. Stergar, R. Hren, and M. Milanič, "Design and Validation of a Custom-Made Hyperspectral Microscope Imaging System for Biomedical Applications," Sensors 23(5), 2374 (2023)

work page 2023

Show all 45 references

[10]

An Atlas of Comparative Vertebrate Histology (Elsevier, 2018)

2018
[12]

An autofocus algorithm considering wavelength changes for large scale microscopic hyperspectral pathological imaging system,

Q. Zhang, Y. Wang, Q. Li, X. Tao, X. Zhou, Y. Zhang, and G. Liu, "An autofocus algorithm considering wavelength changes for large scale microscopic hyperspectral pathological imaging system," J Biophotonics (2022)

2022
[14]

An overview of pre-processing methods available for hyperspectral imaging applications,

D. Cozzolino, P. J. Williams, and L. C. Hoffman, "An overview of pre-processing methods available for hyperspectral imaging applications," Microchemical Journal 193, 109129 (2023)

2023
[15]

Diffusion Theory of Light Transport,

W. M. Star, "Diffusion Theory of Light Transport," in Optical-Thermal Response of Laser-Irradiated Tissue (Springer US, 1995), pp. 131–206

1995
[16]

The microwave and infrared spectra and structure of hydrothiophosphoryl difluoride,

C. R. Nave and J. Sheridan, "The microwave and infrared spectra and structure of hydrothiophosphoryl difluoride," J Mol Struct 15(3), 391–398 (1973)

1973
[17]

Tissue Processing and Hematoxylin and Eosin Staining,

A. T. Feldman and D. Wolfe, "Tissue Processing and Hematoxylin and Eosin Staining," in (2014), pp. 31–43

2014
[18]

Histological Stains: A Literature Review and Case Study,

H. A. Alturkistani, F. M. Tashkandi, and Z. M. Mohammedsaleh, "Histological Stains: A Literature Review and Case Study," Glob J Health Sci 8(3), 72 (2015)

2015
[19]

Monte Carlo Modeling of Light Transport in Tissues,

S. L. Jacques and L. Wang, "Monte Carlo Modeling of Light Transport in Tissues," in Optical-Thermal Response of Laser-Irradiated Tissue (Springer US, 1995), pp. 73–100

1995
[21]

R. Lott, E. Janet Tunnicliffe, J. S. C. Sheppard, M. Hladik, K. Nasim, T. Zeitner, S. K. Haas, and Saeid Movahedi-Lankaran, Practical Guide to Specimen Handling in Surgical Pathology (College of American Pathologists, 2015)

2015
[22]

Molecular Spectral Imaging System for Quantitative Immunohistochemical Analysis of Early Diabetic Retinopathy,

Q. Li, J. Zhang, Y. Wang, and G. Xu, "Molecular Spectral Imaging System for Quantitative Immunohistochemical Analysis of Early Diabetic Retinopathy," Appl Spectrosc 63(12), 1336–1342 (2009)

2009
[23]

New microscopic pushbroom hyperspectral imaging system for application in diabetic retinopathy research,

Q. Li, Y. Xue, G. Xiao, and J. Zhang, "New microscopic pushbroom hyperspectral imaging system for application in diabetic retinopathy research," J Biomed Opt 12(6), 064011 (2007)

2007
[24]

Quantitative Analysis of Protective Effect of Erythropoietin on Diabetic Retinal Cells Using Molecular Hyperspectral Imaging Technology,

Qingli Li, Yiting Wang, Jingfa Zhang, Guotong Xu, and Yongqi Xue, "Quantitative Analysis of Protective Effect of Erythropoietin on Diabetic Retinal Cells Using Molecular Hyperspectral Imaging Technology," IEEE Trans Biomed Eng 57(7), 1699–1706 (2010)

2010
[25]

Microscopic hyperspectral imaging studies of normal and diabetic retina of rats,

Q. Li, Y. Xue, J. Zhang, and G. Xiao, "Microscopic hyperspectral imaging studies of normal and diabetic retina of rats," Sci China C Life Sci 51(9), 789–794 (2008)

2008
[26]

Optical density- based image analysis method for the evaluation of hematoxylin and eosin staining precision,

E. Chlipala, C. M. Bendzinski, K. Chu, J. I. Johnson, M. Brous, K. Copeland, and B. Bolon, "Optical density- based image analysis method for the evaluation of hematoxylin and eosin staining precision," J Histotechnol 43(1), 29–37 (2020)

2020
[27]

Intraoperative Assessment of Tumor Margins in Tissue Sections with Hyperspectral Imaging and Machine Learning,

D. Pertzborn, H.-N. Nguyen, K. Hüttmann, J. Prengel, G. Ernst, O. Guntinas-Lichius, F. von Eggeling, and F. Hoffmann, "Intraoperative Assessment of Tumor Margins in Tissue Sections with Hyperspectral Imaging and Machine Learning," Cancers (Basel) 15(1), 213 (2022)

2022
[28]

Blur-Specific No-Reference Image Quality Assesment for Microscopic Hyperspectral Image Focus Quantification,

L. Quintana, S. Ortega, H. Fabelo, and G. M. Callico, "Blur-Specific No-Reference Image Quality Assesment for Microscopic Hyperspectral Image Focus Quantification," in 2021 11th Workshop on Hyperspectral Imaging and Signal Processing: Evolution in Remote Sensing (WHISPERS) (IE...

2021
[29]

Instrumentation Evaluation for Hyperspectral Microscopy Targeting Enhanced Medical Histology,

L. Quintana, S. Ortega, R. Leon, H. Fabelo, G. M. Callico, C. Lopez, M. Lejeune, and R. Bosch, "Instrumentation Evaluation for Hyperspectral Microscopy Targeting Enhanced Medical Histology," in 2021 XXXVI Conference on Design of Circuits and Integrated Systems (DCIS) (IEEE, 20...

2021
[30]

Optical properties of biological tissues: a review,

S. L. Jacques, "Optical properties of biological tissues: a review," Phys Med Biol 58(11), R37–R61 (2013)

2013
[31]

Using DRS during breast conserving surgery: identifying robust optical parameters and influence of inter-patient variation,

L. L. de Boer, B. H. W. Hendriks, F. van Duijnhoven, M.-J. T. F. D. V. Peeters-Baas, K. Van de Vijver, C. E. Loo, K. Jóźwiak, H. J. C. M. Sterenborg, and T. J. M. Ruers, "Using DRS during breast conserving surgery: identifying robust optical parameters and influence of inter-p...

2016
[32]

Spectral sensing for tissue diagnosis during lung biopsy procedures: The importance of an adequate internal reference and real-time feedback,

J. W. Spliethoff, L. L. de Boer, M. A. J. Meier, W. Prevoo, J. de Jong, T. M. Bydlon, H. J. C. M. Sterenborg, J. A. Burgers, B. H. W. Hendriks, and T. J. M. Ruers, "Spectral sensing for tissue diagnosis during lung biopsy procedures: The importance of an adequate internal refe...

2016
[33]

Tissue biomolecular and microstructure profiles in optical colorectal cancer delineation,

M. S. Nogueira, M. Raju, J. Gunther, S. Maryam, M. Amissah, H. Lu, S. Killeen, M. O’Riordain, and S. Andersson-Engels, "Tissue biomolecular and microstructure profiles in optical colorectal cancer delineation," J Phys D Appl Phys 54(45), 454002 (2021)

2021
[34]

Intraoperative liver steatosis characterization using diffuse reflectance spectroscopy,

N. Reistad, J. H. Nilsson, M. Bergenfeldt, P. Rissler, and C. Sturesson, "Intraoperative liver steatosis characterization using diffuse reflectance spectroscopy," HPB 21(2), 175–180 (2019)

2019
[35]

Measurement of optical transport properties of normal and malignant human breast tissue,

N. Ghosh, S. K. Mohanty, S. K. Majumder, and P. K. Gupta, "Measurement of optical transport properties of normal and malignant human breast tissue," Appl Opt 40(1), 176 (2001)

2001
[37]

Optical properties of human colon tissues in the 350 – 2500 nm spectral range,

A. N. Bashkatov, E. A. Genina, V. I. Kochubey, V. S. Rubtsov, E. A. Kolesnikova, and V. V Tuchin, "Optical properties of human colon tissues in the 350 – 2500 nm spectral range," Quantum Elec (Woodbury) 44(8), 779–784 (2014)

2014
[38]

Estimation of anisotropy coefficient of swine pancreas, liver and muscle at 1064 nm based on goniometric technique,

P. Saccomandi, V. Vogel, B. Bazrafshan, J. Maurer, E. Schena, T. J. Vogl, S. Silvestri, and W. Mäntele, "Estimation of anisotropy coefficient of swine pancreas, liver and muscle at 1064 nm based on goniometric technique," J Biophotonics 8(5), 422–428 (2015)

2015
[39]

Extinction and absorption coefficients and scattering phase functions of human tissues in vitro,

R. Marchesini, A. Bertoni, S. Andreola, E. Melloni, and A. E. Sichirollo, "Extinction and absorption coefficients and scattering phase functions of human tissues in vitro," Appl Opt 28(12), 2318 (1989)

1989
[40]

Hyperspectral imaging: calibration problems and solutions,

P. Geladi, J. Burger, and T. Lestander, "Hyperspectral imaging: calibration problems and solutions," Chemometrics and Intelligent Laboratory Systems 72(2), 209–217 (2004)

2004
[41]

Chang, Hyperspectral Imaging (Springer US, 2003)

C.-I. Chang, Hyperspectral Imaging (Springer US, 2003)

2003
[42]

Graphics-processing-unit-accelerated Monte Carlo simulation of polarized light in complex three-dimensional media,

S. Yan, S. L. Jacques, J. C. Ramella-Roman, and Q. Fang, "Graphics-processing-unit-accelerated Monte Carlo simulation of polarized light in complex three-dimensional media," J Biomed Opt 27(08), (2022)

2022
[43]

Monte Carlo Simulation of Photon Migration in 3D Turbid Media Accelerated by Graphics Processing Units,

Q. Fang and D. A. Boas, "Monte Carlo Simulation of Photon Migration in 3D Turbid Media Accelerated by Graphics Processing Units," Opt Express 17(22), 20178 (2009)

2009
[44]

MCX Cloud—a modern, scalable, high-performance and in-browser Monte Carlo simulation platform with cloud computing,

Q. Fang and S. Yan, "MCX Cloud—a modern, scalable, high-performance and in-browser Monte Carlo simulation platform with cloud computing," J Biomed Opt 27(08), (2022)

2022
[45]

Hybrid mesh and voxel based Monte Carlo algorithm for accurate and efficient photon transport modeling in complex bio-tissues,

S. Yan and Q. Fang, "Hybrid mesh and voxel based Monte Carlo algorithm for accurate and efficient photon transport modeling in complex bio-tissues," Biomed Opt Express 11(11), 6262 (2020)

2020
[46]

Designing a use-error robust machine learning model for quantitative analysis of diffuse reflectance spectra,

A. Scarbrough, K. Chen, and B. Yu, "Designing a use-error robust machine learning model for quantitative analysis of diffuse reflectance spectra," J Biomed Opt 29(01), (2024)

2024
[47]

The spectral image processing system (SIPS)—interactive visualization and analysis of imaging spectrometer data,

F. A. Kruse, A. B. Lefkoff, J. W. Boardman, K. B. Heidebrecht, A. T. Shapiro, P. J. Barloon, and A. F. H. Goetz, "The spectral image processing system (SIPS)—interactive visualization and analysis of imaging spectrometer data," Remote Sens Environ 44(2–3), 145–163 (1993)

1993
[48]

An information-theoretic approach to spectral variability, similarity, and discrimination for hyperspectral image analysis,

Chein-I Chang, "An information-theoretic approach to spectral variability, similarity, and discrimination for hyperspectral image analysis," IEEE Trans Inf Theory 46(5), 1927–1932 (2000)

2000
[49]

New hyperspectral discrimination measure for spectral characterization,

C.-I. Chang, "New hyperspectral discrimination measure for spectral characterization," Optical Engineering 43(8), 1777 (2004)

2004
[50]

Light transport in tissue,

A. E. Profio, "Light transport in tissue," Appl Opt 28(12), 2216 (1989)

1989

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [1]

Medical hyperspectral imaging: a review,

G. Lu and B. Fei, "Medical hyperspectral imaging: a review," J Biomed Opt 19(1), 010901 (2014)

work page 2014

[2] [2]

Hyperspectral and multispectral imaging in digital and computational pathology: a systematic review,

S. Ortega, M. Halicek, H. Fabelo, G. M. Callico, and B. Fei, "Hyperspectral and multispectral imaging in digital and computational pathology: a systematic review," Biomed Opt Express 11(6), 3195 (2020)

work page 2020

[3] [4]

Novel Methodology for Alzheimer’s Disease Biomarker Identification in Plasma using Hyperspectral Microscopy,

H. Fabelo, R. Leon, S. Ortega, F. Balea-Fernandez, C. Bilbao, G. M. Callico, and A. Wagner, "Novel Methodology for Alzheimer’s Disease Biomarker Identification in Plasma using Hyperspectral Microscopy," in 2020 XXXV Conference on Design of Circuits and Integrated Systems (DCIS) (2020)

work page 2020

[4] [5]

Automatic detection of head and neck squamous cell carcinoma on histologic slides using hyperspectral microscopic imaging,

L. Ma, J. V. Little, A. Y. Chen, L. Myers, B. D. Sumer, and B. Fei, "Automatic detection of head and neck squamous cell carcinoma on histologic slides using hyperspectral microscopic imaging," J Biomed Opt 27(04), (2022)

work page 2022

[5] [6]

Leukocyte classification based on spatial and spectral features of microscopic hyperspectral images,

Y. Duan, J. Wang, M. Hu, M. Zhou, Q. Li, L. Sun, S. Qiu, and Y. Wang, "Leukocyte classification based on spatial and spectral features of microscopic hyperspectral images," Opt Laser Technol 112, 530–538 (2019)

work page 2019

[6] [7]

Hyperspectral microscopy- applications of hyperspectral imaging techniques in different fields of science: a review of recent advances,

K. S. Banu, M. Lerma, S. U. Ahmed, and J. L. Gardea-Torresdey, "Hyperspectral microscopy- applications of hyperspectral imaging techniques in different fields of science: a review of recent advances," Appl Spectrosc Rev 1–24 (2023)

work page 2023

[7] [8]

Hyperspectral Push- Broom Microscope Development and Characterization,

S. Ortega, R. Guerra, M. Diaz, H. Fabelo, S. Lopez, G. M. Callico, and R. Sarmiento, "Hyperspectral Push- Broom Microscope Development and Characterization," IEEE Access 7, 122473–122491 (2019)

work page 2019

[8] [9]

Design and Validation of a Custom-Made Hyperspectral Microscope Imaging System for Biomedical Applications,

J. Stergar, R. Hren, and M. Milanič, "Design and Validation of a Custom-Made Hyperspectral Microscope Imaging System for Biomedical Applications," Sensors 23(5), 2374 (2023)

work page 2023