Pith. sign in

REVIEW 4 major objections 5 minor 7 references

Semi-automated image analysis of Cellulose Nanofibrils using Machine learning segmentation and Morphological thinning

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read FACT, a semi-automated image-analysis pipeline, measures the width distribution of branched and entangled cellulose nanofibrils from negative-contrast SEM images, producing thousands of width measurements per image with results comparable…

desk verdict A genuinely useful, honest methods paper for semi-automated CNF width measurement, but the high-branching validation does not support the 'comparable to manual' claim as stated. read the letter →

arxiv 2509.06618 v1 pith:U4ANRT4A submitted 2025-09-08 physics.bio-ph

classification physics.bio-ph
keywords cellulosenanofibrilsnegative-contrastSEMimagesegmentationmachinelearningmorphologicalthinningskeletonizationwidthdistributionsemi-automatedanalysis
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

FACT is a semi-automated image-analysis pipeline that measures the widths of cellulose nanofibrils from negative-contrast SEM images, including hierarchical, branched, and entangled networks that existing tools and manual measurement handle poorly. The paper claims that machine-learning segmentation plus morphological thinning produces a one-pixel-wide skeleton of the fibril network, and reading the local distance to the fibril edge along that skeleton yields width distributions comparable to skilled manual measurements. If correct, the method removes analyst bias while measuring every fibril many times along its length, and once the segmentation model is trained, each image can be analyzed in under five minutes. Validation uses idealized branched geometries, standard wire micrographs, and two published CNF data sets with contrasting branching levels.

What carries the argument

The load-bearing mechanism is the skeleton-distance width identity: after thinning, local fibril width at a skeleton pixel is twice the distance-transform value at that pixel, so $\text{width} = 2d$ where $d$ is the Euclidean distance to the nearest background pixel. The thinning operator erodes a binary object until a one-pixel-wide skeleton remains, and the distance transform records, for each foreground pixel, its distance to the nearest background pixel. Two refinement filters protect this measurement: skeleton segment trimming removes pixels near junction points where branch encroachment distorts width values, and skeleton segment filtering discards short spurious segments that arise from edge defects and have a steep slope in distance-transform value along the segment.

What would settle it

Take nanofibers or wires whose widths are already known from transmission electron microscopy or lithographic dimensions, image them with the same NegC-SEM protocol, run FACT, and compare each reported width distribution with the known values. If the FACT means are off by more than the two-pixel resolution limit, or if widths systematically inflate near branch points and junctions, then the central comparability claim fails.

Watch

Extended reading notes

Core claim

The central claim is that a fully specified image-analysis pipeline, FACT, can measure the width distribution of hierarchical, branched, and entangled cellulose nanofibrils from NegC-SEM images, and that the resulting width statistics are comparable to manual measurements. After machine-learning segmentation converts each micrograph to a binary foreground/background image, morphological thinning produces a one-pixel-wide skeleton; the distance from each skeleton pixel to the nearest background pixel, doubled, gives the local fibril width. The authors demonstrate the claim on simulated branch hierarchies, on optical micrographs of uniform wires, and on two published CNF data sets, one lightly branched and one heavily branched and networked. They report that FACT resolved the distinct widths of simulated branch levels and gave mean widths close to manual measurements on real CNF images, while producing thousands of width measurements per image instead of one per fibril.

Load-bearing premise

FACT assumes that after machine-learning segmentation and thinning, the one-pixel line left through each fibril is exactly its centerline, so that twice the distance from that line to the nearest background pixel equals the true local width.

Editorial extensions

If this is right

  • Width statistics can be collected from entire branched CNF networks, with every fibril measured many times along its length, eliminating analyst bias in fibril selection and placement.
  • Once a U-Net segmentation model is trained, each NegC-SEM image can be analyzed in under five minutes, making batch comparisons across CNF grades practical.
  • Width resolution is limited to two pixels, so reliable measurement requires fibrils to span at least five pixels across; digital zoom cannot substitute for higher magnification or higher pixel density.
  • Skeleton segment lengths and junction counts should not be used as direct measures of fibril length or branching density on real CNFs, because segmentation edge defects create artificial junctions.

Reading between the lines

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

  • A natural next test is applying FACT to the same CNF suspension imaged at several magnifications and checking whether the width distribution becomes magnification-independent after weighting by pixel counts; if not, the remaining differences quantify residual segmentation bias.
  • The skeleton-as-medial-axis assumption could be checked quantitatively by comparing FACT widths against TEM widths of the same fibrils or against synthetic images with known ground-truth widths, especially near branch points.
  • The orientation correction suggests a further step: converting FACT's length-weighted width histogram into a number-weighted fibril width distribution, which would make it directly comparable to manual measurements.
  • Because FACT separates segmentation from measurement, the framework could be extended to other fibrous materials and imaging modalities by retraining the segmentation model while keeping the thinning and distance-transform measurement intact.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper presents FACT (Fibril Analysis for Cellulose Technology), a semi-automated image analysis framework that measures fibril width distributions from negative-contrast scanning electron microscopy (NegC-SEM) images of cellulose nanofibrils. The pipeline consists of image acquisition, machine-learning segmentation (Weka or U-Net), morphological thinning to produce skeletons, skeleton refinement via segment trimming (SST) and filtering (SSF), and width measurement from the distance transform of the binary segmentation. The authors validate FACT on simulated branched structures, optical micrographs of wires, low-branching CNFs from Beaumont et al., and high-branching CNFs from Ringania et al., comparing FACT results against manual width measurements. The paper claims that FACT can measure width distributions of hierarchical, branched, and entangled CNF structures and that once the U-Net model is trained, each image can be analyzed in under 5 minutes.

Significance. If the central claims hold, FACT would be a useful, open-source tool for objective, high-throughput width characterization of branched and entangled CNF networks, addressing a recognized need in the CNF community. The paper's strengths include systematic validation on idealized binary geometries (Figs. 5, 6 and supporting figures), quantitative validation on constant-diameter wires (Fig. 7), and good agreement with manual measurements for low-branching CNFs at multiple magnifications (Figs. 9, 10), where mean differences are within pixel resolution. The public availability of the code on Zenodo and the clear description of the skeleton-refinement steps are also positive features. However, the evidence for the high-branching CNF case, which is the most consequential claim, is weaker, as detailed in the major comments.

major comments (4)
  1. [High-Level Branching CNFs, Fig. 13] The comparison in Fig. 13 does not support the abstract's claim that "FACT width results were comparable with manual measurements" for high-branching CNFs. The FACT histogram (Fig. 13c, mean 510 nm, n = 104,891 pixels) is pixel/length-weighted, whereas the manual histogram (Fig. 13d, mean 320 nm, n = 978 fibrils) is per-fibril weighted, and the authors themselves state that "FACT is not entirely commensurate with the manual measurement approach." The 190 nm mean offset is attributed to this sampling difference, which is plausible, but the manual data therefore cannot independently confirm the accuracy of FACT's width values on branched networks. I recommend one of the following: (a) reweight FACT's pixel-based distribution to a per-fibril basis using skeleton-segment lengths, which would allow a direct statistical comparison; or (b) validate on realistic synthetic branched phantoms that include segmentation noise, edge dilation, and overlap ambiguity with known ground-truth widths; or (c) include internal calibration features of known diameter in the SEM images. Without such a test, the high-branching validation remains inconclusive.
  2. [Image Segmentation (U-Net training)] The U-Net segmentation accuracy is not quantified. Section "Image Segmentation" states that Weka segmentation on eight images was used as "ground truth images" to train the U-Net, but the Weka segmentations are themselves described as "too noisy," producing jagged edges. No IoU, Dice, or pixel-wise accuracy is reported on held-out real images or against independently labeled regions. Since the width values in FACT are directly derived from the segmented binary image, errors in segmentation propagate directly into width bias and skeleton artifacts. I recommend reporting segmentation metrics on a held-out subset of real NegC-SEM images, at minimum, and ideally comparing the U-Net output against manually corrected segmentations rather than against the same noisy Weka output used for training.
  3. [Measurement & Analysis (width definition)] The width measurement assumes that, after thinning and refinement, the skeleton coincides with the medial axis of each fibril so that twice the distance-transform value equals the local fibril width. The manuscript documents multiple violations of this assumption: skeleton encroachment at branch junctions (Fig. S9), spurious segments from edge defects (Fig. 4), and skeleton mismatch in regions of extensive bundling (Fig. 11e). The paper also states that "a robust approach to quantifying the fraction of unwanted skeleton segments remains unclear." The simulated validations (Figs. 5, 6) use clean binary phantoms without segmentation noise or edge dilation, so they do not bound the magnitude of these errors on real images. I recommend adding a sensitivity analysis that quantifies how many skeleton pixels are affected by encroachment and edge defects in the real CNF images (e.g., by manual labeling or by deliberately introducing controlled edge noise in phantoms) and reporting the resulting effect on the width distribution.
  4. [Skeleton Refining (SST and SSF parameters)] The SST and SSF parameters are set by visual inspection and trial and error, with values that differ between image sets and even between figures (e.g., Fig. S11 caption states SST=10% and SSF=20°, while Fig. 13 and the text for high-branching CNFs state SST=25% and SSF=25°). This introduces analyst-dependent choices into the pipeline, which partially undermines the claim of removing analyst bias. The paper should either provide a principled, automated criterion for selecting these parameters or explicitly characterize how sensitive the reported width means and distributions are to the chosen SST/SSF values across a plausible range.
minor comments (5)
  1. [Abstract] There is a typo in the abstract: "mfeasurement" should be "measurement."
  2. [Fig. S11 caption] The SST/SSF values in the Fig. S11 caption (SST=10%, SSF=20°) are inconsistent with the values reported for the same high-branching images in the main text and Fig. 13 (SST=25%, SSF=25°); please correct the inconsistency.
  3. [Fig. 13 caption] The Fig. 13 caption contains a typo: "ixels along these lines" should be "Pixels along these lines."
  4. [References] Several reference entries contain typos: "disertation" should be "dissertation" in Ringania (2023), and "Instutue" should be "Institute."
  5. [Code availability] The Zenodo link is given as a placeholder in the Conclusion and Declarations; please include the actual DOI or URL before publication so that the reproducibility claim can be verified.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: FACT width values are direct distance-transform readings validated against known-width phantoms and manual measurements.

full rationale

The derivation chain is self-contained. In 'Measurement & Analysis', FACT defines fibril width as twice the Euclidean distance from the thinned skeleton to the nearest background pixel, and each reported histogram is obtained by reading distance-transform values at skeleton coordinates. No width value is produced by fitting, parameter inversion, or by a self-citation. The idealized-structure validation uses phantoms with known prescribed widths (e.g., the five-level branched structure with widths 4, 9, 18, 36, and 62 pixels), and the wire validation compares FACT against independent manual ImageJ measurements; these are external anchors for the width-reading pipeline. The orientation correction is a calibration of skeleton pixel-count resampling, and SST/SSF are user-selected trimming/filter thresholds set by visual inspection; neither changes the distance-transform width values, so no fitted input is renamed as a prediction. The high-branch comparison uses manual measurements from the authors' prior Ringania et al. study, which is self-citation in the validation set, but those manual readings are independent single-per-fibril measurements and the paper explicitly states the two histograms are not commensurate (length-weighted pixel counts versus per-fibril counts). That candor, plus the use of independent phantom and wire benchmarks, makes the comparison a stated limitation rather than a constructed equivalence. The lack of held-out segmentation metrics and the Weka-derived U-Net ground truth are accuracy and robustness concerns, not circularity: FACT's output is not equal to its input by construction. Overall, no load-bearing circular step was found.

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

The central claim rests on domain assumptions about microscopy representativeness, the validity of skeleton distance transforms as width measures, and the reliability of Weka-generated ground truth; the free parameters are user-tuned segmentation and skeleton-refinement settings that vary by image set.

free parameters (4)
  • SST fraction (skeleton segment trimming) = 10%, 30% or 50% for simulations; 10% for low-branching CNFs; 25% for high-branching CNFs
    Chosen by visual inspection and trial and error. It removes skeleton pixels near junctions and therefore controls which pixels enter the width histogram.
  • SSF slope threshold (skeleton segment filtering) = 10 degrees for simulations; 20 degrees for low-branching CNFs; 25 degrees for high-branching CNFs
    The absolute slope of the standardized distance-transform-value versus position is used to discard segments that do not follow fibril centerlines; the cutoff is set by visual inspection and trial and error.
  • Gaussian blur kernel size = 0 or 2
    Applied to filtered binary images to smooth jagged edges before thinning; the user selects the kernel, and the choice changes edge position and measured widths.
  • Orientation correction factor = Fitted curve in Fig. S5
    A function fitted to normalized skeleton pixel counts versus object orientation is used to resample width statistics, correcting undercounting of off-axis fibrils; it is a calibration fit, not a first-principles correction.
assumptions (4)
  • domain assumption The skeleton produced by morphological thinning lies on the medial axis of each fibril, so twice the distance transform at a skeleton pixel equals the local fibril width.
    Stated in Measurement & Analysis as the definition of width; Fig. S9 and Fig. 4 show the skeleton does not respect branch boundaries near junctions, so this holds only approximately.
  • domain assumption Weka segmentations thresholded at 0.5 are correct enough to serve as ground truth for U-Net training and evaluation.
    In Image Segmentation, Weka output is treated as ground truth for eight training images; no independent segmentation accuracy metric is reported.
  • domain assumption The analyzed NegC-SEM images are representative of the CNF width distribution.
    The Results state correctness relies on images representing the material; the paper uses few images from prior studies and notes multiple magnifications are needed for representativeness.
  • domain assumption Manual ImageJ measurements are a valid reference for FACT width results.
    Manual measurements are the benchmark for both low- and high-branching CNFs, yet the paper motivates FACT by criticizing manual measurement for analyst bias and fatigue, so this reference is not an independent gold standard.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Semi-automated image analysis of Cellulose Nanofibrils using Machine learning segmentation and Morphological thinning." pith.science (2026). https://pith.science/paper/U4ANRT4A

@misc{pith2026250906618,
  author       = {Pith},
  title        = {Pith review of: Semi-automated image analysis of Cellulose Nanofibrils using Machine learning segmentation and Morphological thinning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/U4ANRT4A}},
  note         = {Machine review of arXiv:2509.06618}
}
read the original abstract

Reliable and rapid morphology measurement of cellulose nanofibrils (CNFs) with a high level of branching and entanglement is crucial for quality control, grade definition, and investigating morphology-performance relationships in various applications. An image analysis framework, Fibril Analysis for Cellulose Technology (FACT), which utilizes machine learning (ML) segmentation and morphological thinning, was developed to measure the fibril width distribution of cellulose nanofibers (CNFs) from negative contrast scanning electron microscopy (NegC-SEM) images. The high-contrast and wide magnification range of NegC-SEM imaging enabled the capture of micro- and nanoscopic hierarchical branching structures of CNFs. Two ML approaches [Weka and U-Net] were used to create detailed binary segmentation of grayscale NegC-SEM images, critical for the width analysis. Morphological thinning was applied to the binary image to produce a 1-pixel-wide skeleton of the CNF fibril structure. Subsequently, the distance between the skeleton and the original fibril edge was used to calculate fibril width. The FACT framework was optimized and validated with idealized geometric and hierarchical branched structures. FACT effectively performed segmentation, skeletonization, and fibril width measurement of these CNF morphologies. FACT width results were comparable with manual measurements. In the manual method, a single measurement is made per fibril. In contrast, FACT simultaneously makes multiple measurements along each fibril within the entire CNF branched network structure. The advantage of FACT is that complicated branching and network CNF structures can be measured without imparting any analyst bias in fibril selection and measurement. Additionally, once the ML model is trained, each image can be analyzed in under 5 minutes.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

7 extracted references · 7 canonical work pages

  1. [2]

    fibril collective

    The thinning operator is applied to the filte red binary image. Then , the coordinates of the 9 resulting skeleton pixels are used to extract their corresponding values from the distance transform image of the unfiltered binary. Skeleton Refining: The sensitivity of the thinning operator to sharp en foreground edges sometimes resulted in undesirable skele...

  2. [4]

    skeleton refining, and 5) image analysis. 5 Fig. 1 Flowchart of the FACT automated image analysis process. Where SST = skeleton segment trimming, and SSF = skeleton segment filtering NegC-SEM Image Acquisition: Image acquisition is critically important for the FACT approach to identify CNF branching and provide accurate width measurements of individual fi...

  3. [5]

    ground truth images

    were prepared by first diluting the starting 3 wt.% solid aqueous suspension to 0.001 wt.% using DI water and subsequently mixed using a vortex mixer ( VWR Analog Vortex Mixer No. 10153-838, speed 7, 30 s). A micro-pipette was used to deposit four 2 µL droplets of the diluted suspension onto a 1 cm x 1 cm silicon wafer (resistivity: 1-30 Ohm-cm, P-type wi...

  4. [2019]

    Perspectives on Cellulose Nanofibril Size Measurement Using Scann ing Electron Microscopy

    and ImageJ (Schneider et al. 2012; Ang et al. 2020) has been widely used for CNF image analysis. However, object selection and measurement are susceptible to user bias and fatigue, which contribute to the discrepancies in dimensional data of CNFs available in the literature. Semi-automated image analysis programs designed for the analysis of high aspect r...

  5. [2022]

    The ‘fines’ percentage, defined by the ISO specification (ISO 2014), is the percentage of fibrils with a length less than 200 µm

    study, were produced by the Process Development Center at the University of Maine [lot # U-103, 90% fines]. The ‘fines’ percentage, defined by the ISO specification (ISO 2014), is the percentage of fibrils with a length less than 200 µm . These CNFs were prepared by mechanical fibrillation of wood pulp fibers (Masuko MKZB15 -50J super mass colloider ) wit...

  6. [2810]

    Open Physics 10(1):181-188

    https://doi.org/10.1007/s10570-025-06458-2 27 Nečas D, Klapetek P (2012) Gwyddion: an open-source software for SPM data analysis. Open Physics 10(1):181-188. https://doi.org/10.2478/s11534-011-0096-2 Persson NE, Rafshoon J, Naghshpour K et al (2017) High-throughput image analysis of fibrillar materials: a case study on polymer nanofiber packing, alignment...

  7. [4590]

    https://doi.org/10.15376/BIORES.15.2.ZAMBRANO 28 Supplementary Material Semi-automated image analysis of Cellulose Nanofibrils using Machine learning segmentation and Morphological thinning Carlos Baez1, Udita Ringania2, Saad Bhamla2, Robert J. Moon1 1 The Forest Products Laboratory, USDA Forest Service, Madison, WI 53726 2 Chemical and Biomolecular Engin...

Pith tools

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