pith. sign in

arxiv: 2606.25183 · v1 · pith:YXLWTXSVnew · submitted 2026-06-23 · ⚛️ physics.optics

Low-Cost Continuous-Wave Diffusive Microtomography with Fiber-Scanned White-Light Illumination

Alexander Ingold This is my paper

Pith reviewed 2026-06-25 21:41 UTC · model grok-4.3

classification ⚛️ physics.optics
keywords continuous-wave diffusive tomographylow-cost microscopyfiber-scanned illuminationvolumetric reconstructionbiological samplessmartphone microscopemachine learning optimizationdiffusive light transport
0
0 comments X

The pith

Scanned fiber white-light illumination with a smartphone microscope and machine learning produces three-dimensional reconstructions of biological samples at low cost.

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

The paper demonstrates an ultra-low-cost continuous-wave diffusive tomography system that combines a smartphone microscope, a white LED coupled to an optical fiber, 3D-printed positioners, and a physics-based forward model optimized by machine learning. It shows full-color volumetric reconstructions from cleared poplar sections, scattering phantoms, fungal mycelium near plant roots, and thick branches imaged with a side-emitting fiber. The approach targets diffusive light transport rather than ballistic imaging, enabling internal views without expensive optics or X-ray sources. Results remain qualitative and exploratory yet indicate that inexpensive scanned illumination can yield usable three-dimensional outputs for microscopy experiments.

Core claim

The central claim is that scanned fiber illumination from inexpensive white-light sources, paired with a machine-learning-optimized physics forward model, enables continuous-wave diffusive microtomography capable of full-color volumetric reconstructions in both cleared and scattering biological samples using only a smartphone microscope and 3D-printed hardware.

What carries the argument

The physics-based forward model optimized with machine learning, which models light transport under continuous-wave diffusive conditions to invert scanned fiber illumination data into volumetric reconstructions.

If this is right

  • Scanned fiber illumination can replace more complex light sources in diffusive tomography setups.
  • Machine learning optimization of the forward model improves reconstruction fidelity from low-cost data.
  • The system supports imaging of both optically cleared and naturally scattering biological specimens.
  • Full-color output becomes feasible without additional spectral hardware.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The method could support portable or educational 3D imaging in settings without access to research-grade microscopes.
  • Extensions to other wavelengths or dynamic samples would require only changes to the illumination and model training.
  • Combining the approach with additional smartphone sensors could yield multimodal low-cost tomography.

Load-bearing premise

The physics-based forward model, once optimized by machine learning, sufficiently captures light transport in the tested scattering and cleared samples to yield reliable volumetric outputs.

What would settle it

Quantitative comparison of the reconstructed volumes against known ground-truth structures in the same samples, such as measured feature sizes or densities obtained from higher-resolution reference imaging, showing systematic mismatch.

Figures

Figures reproduced from arXiv: 2606.25183 by Alexander Ingold.

Figure 1
Figure 1. Figure 1: A simulated physics-based tissue model with learned absorbance parameters and a sim [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: A. A 250 µm thick tangential section of poplar branch was soaked for 18 hours in tartrazine. B. Transmission images were captured. C. The simulated light output from the tomo￾graphic reconstruction of the poplar section. D. Absorbance values rendered in a heat map from three views: front, angled, and rear. 7 [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: A. A 2 mm thick section of an optical phantom made of silicon carbide sieved for 25– [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: A. Fungal mycelium growing from Arabidopsis root. B. Transmission images were cap [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: A. A 90 degree emission optical fiber is inserted into 3.5 mm thick poplar branch with [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗
read the original abstract

Tomographic microscopy enables three-dimensional internal imaging but often requires expensive optical or X-ray instrumentation. Here we present an ultra-low-cost continuous-wave diffusive tomography (CWDT) system for biological samples. The system uses a smartphone microscope, a white LED coupled into an optical fiber, 3D-printed micropositioners, and a physics-based forward model optimized with machine learning. We demonstrate full-color volumetric reconstructions from a tartrazine-cleared poplar section, a scattering phantom, fungal mycelium near an Arabidopsis root, and thick poplar branch imaging with an inserted side-emitting fiber. The current results are qualitative and exploratory, but they show that scanned fiber illumination and inexpensive hardware can produce useful three-dimensional reconstruction outputs for low-cost microscopy experiments.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 2 minor

Summary. The manuscript presents an ultra-low-cost continuous-wave diffusive microtomography (CWDT) system for biological samples. It combines a smartphone microscope, white LED coupled to a scanned optical fiber, 3D-printed micropositioners, and a physics-based forward model whose parameters are optimized by machine learning. Qualitative full-color volumetric reconstructions are shown for a tartrazine-cleared poplar section, a scattering phantom, fungal mycelium near an Arabidopsis root, and a thick poplar branch imaged with an inserted side-emitting fiber. The work is explicitly framed as exploratory and qualitative, with the central claim that scanned fiber illumination plus inexpensive hardware can yield useful three-dimensional outputs for low-cost microscopy experiments.

Significance. If the reconstructions prove reliable, the approach could lower barriers to 3D tomographic imaging in biology by replacing costly instrumentation with commodity components and a learned forward model. The explicit use of a physics-based model optimized by ML is a constructive element that may generalize to other diffusive regimes. At present the exploratory framing limits the result to a proof-of-concept demonstration rather than a ready-to-deploy technique.

minor comments (2)
  1. [Abstract] Abstract: the phrase 'useful three-dimensional reconstruction outputs' is not accompanied by any operational criterion (e.g., recovery of known phantom features, consistency across illumination angles, or comparison with a reference modality). Adding a short sentence that defines what 'useful' means in the present qualitative setting would help readers evaluate the displayed volumes.
  2. [Methods (forward-model section)] The manuscript states that the forward model is 'optimized with machine learning' but supplies no information on the loss function, regularization, or whether the fitted parameters are held fixed when reconstructing new targets. A brief methods paragraph clarifying this point would remove ambiguity about potential circularity.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive assessment of the manuscript and for recommending minor revision. The work is explicitly presented as an exploratory, qualitative demonstration, consistent with the referee's characterization. No specific major comments were provided in the report.

Circularity Check

0 steps flagged

No significant circularity

full rationale

The paper explicitly frames its outputs as qualitative and exploratory demonstrations of a low-cost CWDT system using scanned fiber illumination and a physics-based forward model optimized via ML. No derivation chain, equations, or predictions are presented in the available text that reduce by construction to fitted parameters or self-citations; the central claim remains internally consistent with the modest, hardware-focused scope without load-bearing steps that equate outputs to inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review prevents enumeration of specific free parameters or axioms; the central claim rests on an unstated assumption that the physics-based forward model plus ML fitting produces faithful reconstructions without post-hoc tuning that would invalidate the outputs.

pith-pipeline@v0.9.1-grok · 5649 in / 1062 out tokens · 16336 ms · 2026-06-25T21:41:51.705003+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

19 extracted references · 9 canonical work pages

  1. [1]

    Continuous wave-diffuse optical tomography (cw-dot) in human brain mapping: A review

    Shuo Guan, Yuhang Li, Yuanyuan Gao, Yuxi Luo, Hubin Zhao, Dalin Yang, and Rihui Li. Continuous wave-diffuse optical tomography (cw-dot) in human brain mapping: A review. 15 Sensors, 25(7):2040, 2025. doi: 10.3390/s25072040

    work page doi:10.3390/s25072040 2040

  2. [2]

    J. P. Culver, R. Choe, M. J. Holboke, L. Zubkov, T. Durduran, A. Slemp, V. Ntziachristos, B. Chance, and A. G. Yodh. Three-dimensional diffuse optical tomography in the parallel plane transmission geometry: Evaluation of a hybrid frequency domain/continuous wave clinical system for breast imaging.Medical Physics, 30(2):235–247, 2003. doi: 10.1118/1.1534109

    work page doi:10.1118/1.1534109 2003

  3. [3]

    Alfonso Galderisi, Sabrina Brigadoi, Simone Cutini, Sara Basso Moro, Elisabetta Lolli, Feder- ica Meconi, Silvia Benavides-Varela, Eugenio Baraldi, Piero Amodio, Claudio Cobelli, Daniele Trevisanuto, and Roberto Dell’Acqua. Long-term continuous monitoring of the preterm brain with diffuse optical tomography and electroencephalography: A technical note on ...

    work page doi:10.1117/1.nph.3.4.045009 2016

  4. [5]

    J. J. Davenport, M. Hickey, J. P. Phillips, and P. A. Kyriacou. Method for producing angled optical fiber tips in the laboratory.Optical Engineering, 55(2):026120, 2016. doi: 10.1117/1. oe.55.2.026120

    work page doi:10.1117/1 2016

  5. [6]

    Triaxis micropositioner.https://makerworld.com/en/models/ 1256154-triaxis-micropositioner, 2025

    Alexander Ingold. Triaxis micropositioner.https://makerworld.com/en/models/ 1256154-triaxis-micropositioner, 2025. MakerWorld 3D-print model page

    2025

  6. [7]

    Triaxis micropositioner micromanipulator 3-axis stage.https://www

    Alexander Ingold. Triaxis micropositioner micromanipulator 3-axis stage.https://www. thingiverse.com/thing:6976184, 2025. Thingiverse 3D-print model page

    2025

  7. [8]

    Triaxis micropositioner cad.https://cad.onshape

    Alexander Ingold. Triaxis micropositioner cad.https://cad.onshape. com/documents/c31396d958296393a0cbb499/w/10e122b0be94b2b2a2f3787d/e/ 975a9698867d3d77a59824b4, 2025. Editable Onshape CAD document

    2025

  8. [9]

    Swentowsky, and David Jackson

    Penelope Lindsay, Kyle W. Swentowsky, and David Jackson. Cultivating potential: Harnessing plant stem cells for agricultural crop improvement.Molecular Plant, 17(1):50–74, 2024. doi: 10.1016/j.molp.2023.12.014

    work page doi:10.1016/j.molp.2023.12.014 2024

  9. [10]

    Juan Du, Yichen Wang, Wenfan Chen, Mingling Xu, Ruhong Zhou, Huixia Shou, and Jun Chen. High-resolution anatomical and spatial transcriptome analyses reveal two types of meristematic cell pools within the secondary vascular tissue of poplar stem.Molecular Plant, 16(5):809–828, 2023. doi: 10.1016/j.molp.2023.03.005

    work page doi:10.1016/j.molp.2023.03.005 2023

  10. [11]

    Live cell imaging of cellular dynamics in poplar wood using computational cannula microscopy.Applied Optics, 63(28):G47–G53, 2024

    Alexander Ingold, Gayatri Mishra, Reed Sorenson, Andrew Groover, Leslie Sieburth, and Rajesh Menon. Live cell imaging of cellular dynamics in poplar wood using computational cannula microscopy.Applied Optics, 63(28):G47–G53, 2024. doi: 10.1364/AO.523456

    work page doi:10.1364/ao.523456 2024

  11. [12]

    Rommelfanger, Carl H

    Zihao Ou, Yi-Shiou Duh, Nicholas J. Rommelfanger, Carl H. C. Keck, Shan Jiang, Kenneth Brinson, Su Zhao, Elizabeth L. Schmidt, Xiang Wu, Fan Yang, Betty Cai, Han Cui, Wei Qi, Shifu Wu, Adarsh Tantry, Richard Roth, Jun Ding, Xiaoke Chen, Julia A. Kaltschmidt, Mark L. Brongersma, and Guosong Hong. Achieving optical transparency in live animals with absorbin...

    work page doi:10.1126/science.adm6869 2024

  12. [13]

    Miniscope v4.https://open-ephys.org/miniscope-v4, 2024

    Open Ephys. Miniscope v4.https://open-ephys.org/miniscope-v4, 2024

    2024

  13. [14]

    Miniscope v4 assembly wiki.https://github.com/AharoniLab/ Miniscope-v4/wiki/Assembly, 2024

    Aharoni Lab. Miniscope v4 assembly wiki.https://github.com/AharoniLab/ Miniscope-v4/wiki/Assembly, 2024

    2024

  14. [15]

    Autofluorescence in plants.Molecules, 25(10):2393, 2020

    Lloyd Donaldson. Autofluorescence in plants.Molecules, 25(10):2393, 2020. doi: 10.3390/ molecules25102393

    2020

  15. [16]

    Miniscope-daq-qt-software.https://github.com/Aharoni-Lab/ Miniscope-DAQ-QT-Software, 2024

    Aharoni Lab. Miniscope-daq-qt-software.https://github.com/Aharoni-Lab/ Miniscope-DAQ-QT-Software, 2024. GitHub repository

    2024

  16. [17]

    Forster, D

    B. Forster, D. Van De Ville, J. Berent, D. Sage, and M. Unser. Complex wavelets for extended depth-of-field: A new method for the fusion of multichannel microscopy images.Microscopy Research and Technique, 65(1–2):33–42, 2004

    2004

  17. [18]

    3d gaussian splatting for real-time radiance field rendering.ACM Transactions on Graphics, 42(4):139,

    Bernhard Kerbl, Georgios Kopanas, Thomas Leimkuehler, and George Drettakis. 3d gaussian splatting for real-time radiance field rendering.ACM Transactions on Graphics, 42(4):139,

  18. [19]

    doi: 10.1145/3592433

    work page doi:10.1145/3592433

  19. [20]

    torch-splatting: A pure pytorch implementation of 3d gaussian splatting.https: //github.com/hbb1/torch-splatting, 2025

    hbb1. torch-splatting: A pure pytorch implementation of 3d gaussian splatting.https: //github.com/hbb1/torch-splatting, 2025. GitHub repository, MIT license. 17 Figure S2: Raw chicken breast, sectioned with 0–3 mm thickness. A. Sample stained with tartrazine for 15 minutes; grid lines are 0.125 inches. B. Sample stained with tartrazine for 18 hours; grid ...

    2025