Pith. sign in

REVIEW 3 major objections 5 minor 36 references

Loggia dei Lanzi: AI Thermography Enhancement Comparisons through 3D Photogrammetry

T0 review · 3 major / 5 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read AI super-resolution of thermal images does not improve photogrammetric 3D reconstruction of heritage surfaces, and aggressive models degrade it; native imagery is the most accurate.

desk verdict A careful, reproducible negative result for AI upscaling in heritage thermal photogrammetry — though the AI models were tested via a float32 roundtrip, not their native 8-bit input, so the general 'AI SR degrades reconstruction' claim needs that caveat. read the letter →

arxiv 2608.02404 v1 pith:OJPDVH2M submitted 2026-08-03 cs.CV cs.DL

Scott McAvoy , Jonathan Klingspon , George Bent , Dave Pfaff , Aviral Agarwal , Maurizio Seracini , Falko Kuester This is my paper
classification cs.CVcs.DL
keywords thermographyAIsuper-resolutionphotogrammetryculturalheritagethermalimagingStructure-from-Motion3Dreconstructionradiometricfidelity
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 asks whether AI upscaling of thermal images helps or hurts 3D photogrammetric reconstruction of heritage buildings, using a winter thermal survey of Florence's Loggia dei Lanzi as the test case. The authors compared six resolution tiers—native sensor captures, hardware pixel-shifted super-resolution, bicubic interpolation, and three state-of-the-art AI super-resolution models—inside an identical structure-from-motion pipeline with camera positions locked to LiDAR reference geometry. Their central finding is that native imagery produces the most accurate reconstruction by every measure, that hardware super-resolution comes close but adds little, and that the AI methods provide no measurable improvement and by quality metrics degrade the result, with the most aggressive models collapsing reconstruction onto edges. This matters because quantitative heritage thermography aims to measure subsurface structures, and the paper indicates that radiometric fidelity at native resolution outweighs any perceptual sharpness AI upscaling adds.

What carries the argument

The load-bearing mechanism is a controlled structure-from-motion comparison: every resolution tier is processed through an identical photogrammetric network with camera positions locked to a LiDAR reference frame and a fixed, resolution-scaled camera calibration, so image content is the only variable. The decisive instrument is multi-view geometric consistency—reprojection error normalized to physical units via an analytically derived ground sample distance (the wall area each pixel covers), plus resolution-independent reconstruction uncertainty and the fraction of tie points observed in three or more images. These quantities expose whether AI-synthesized pixels triangulate to the same 3D lo

What would settle it

Re-run the identical six-tier comparison with the same AI models applied to 8-bit-quantized thermal frames (their native training range), keeping the 16-bit native path as control; if a model then matches or beats native reconstruction uncertainty and surface completeness, the null result is an artifact of the float-normalized 16-bit roundtrip, not a property of AI upscaling.

Watch

Extended reading notes

Core claim

The paper's central claim is that AI super-resolution, whether from a generic transformer, a thermal-challenge winner, or a diffusion model, yields no measurable improvement in photogrammetric reconstruction over native or interpolated baselines—and by quality metrics actively degrades it. Native imagery produced the most geometrically accurate reconstruction by every measure: lowest normalized reprojection error, highest cross-view consistency, and a sparse cloud essentially free of poorly triangulated points. Hardware pixel-shifted super-resolution was the only enhancement that approached native quality, but its benefit was modest. The most aggressive AI upscaling methods increased point c

Load-bearing premise

The load-bearing premise is that the AI models were fairly tested, but they were run on float-normalized 16-bit radiometric data rather than their native 8-bit training domain; if that mismatch degrades their feature representations, the null result may be an artifact of the integration pipeline rather than a property of the models.

Editorial extensions

If this is right

  • For quantitative heritage thermography, native-resolution 16-bit imagery should be the default; AI upscaling should not be used when measurements are the goal.
  • Hardware pixel-shifted super-resolution adds little geometric value in the field but remains a low-risk addition; it should not substitute for good acquisition conditions.
  • Aggressive 4x and 8x AI upscaling actively degrades sparse-cloud quality and collapses dense reconstruction onto edges, leaving the smooth wall surfaces—where hidden features appear—unreconstructed.
  • Tie-point and dense-point counts are misleading; high-uncertainty or single-depth-map points inflate them, so confidence-filtered quality metrics are essential in reconstruction evaluation.
  • Because AI super-resolution requires converting 16-bit radiometric data to the 8-bit range most models accept, adopting it discards calibrated temperature information that subsurface analysis depends on.

Reading between the lines

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

  • An implication the paper leaves implicit: the same multi-view consistency test could be run on visible-light photogrammetry of low-texture facades; the edge-collapse failure mode may explain mixed results in AI-upscaled architectural scans.
  • A testable extension: feed the AI models 8-bit quantized versions of the thermal frames while keeping the 16-bit-native branch; if an 8-bit-fed model then matches native quality, the null result is an artifact of the float-normalized roundtrip rather than a general property of AI super-resolution.
  • The finding that raw point counts invert reconstruction value suggests previous photogrammetry studies that report only point counts may need re-evaluation with uncertainty-filtered metrics.
  • If hardware microscanning were driven by precise actuator-controlled sub-pixel shifts rather than improvised tripod jiggle, a small genuine benefit might appear; this remains untested and is a natural next experiment.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper compares six resolution tiers for thermal photogrammetric reconstruction of the Loggia dei Lanzi rear wall using a FLIR T1020HD: native 1024×768, UltraMax hardware super-resolution, bicubic upscaling, SwinIR, DifIISR, and TongJi/DRCT. All tiers are processed through an identical Agisoft Metashape workflow with LiDAR-locked camera poses, a fixed resolution-scaled calibration, and a common 149-image subset. Metrics include tie-point counts, GSD-normalized reprojection error (mm), reconstruction uncertainty, cross-view consistency, and dense-cloud confidence/completeness. The central finding is that native imagery yields the most accurate and highest-quality reconstruction; UltraMax is marginally useful with the cleanest enhanced sparse cloud; all evaluated AI super-resolution methods fail to improve reconstruction, and the high-factor generative models actively degrade it, concentrating dense points on high-contrast edges.

Significance. If correct, this is a useful negative result for heritage thermography: quantitative photogrammetric documentation should rely on native radiometric imagery, and AI upscaling should not be assumed beneficial. The study's strengths are the controlled comparison (fixed camera network, fixed calibration, common image subset), the analytical GSD normalization tied to LiDAR geometry, the use of two resolution-independent metrics (reconstruction uncertainty and multi-view consistency), and the public release of datasets and processing scripts. The conclusion is, however, only as strong as the premise that the AI models were evaluated in a regime relevant to their design; the float32-input issue is central to that premise.

major comments (3)
  1. [§7.2, §9, Table 1] The AI models are applied to float32-normalized 16-bit thermal values, not to the 8-bit inputs on which they were trained and benchmarked. The normalization scheme is not specified (global 65535 scaling vs. per-image min-max), and the claimed roundtrip-radiometric-fidelity metrics are never reported. If the models see an almost-black or abnormally low-contrast input, the null result could reflect integration failure rather than model capability. The Discussion's argument that AI SR 'requires reducing 16-bit radiometric data to 8-bit imagery' is internally inconsistent with the float32 pipeline actually used. Please add an evaluation with proper 8-bit quantization (or at minimum report input-distribution statistics and roundtrip errors) and restrict the general conclusion accordingly.
  2. [§7.1, §9] UltraMax was captured with a deliberately loosened tripod head, and no quantitative check is reported that the induced displacement matches the hand-tremor magnitude and distribution assumed by FLIR's reconstruction. The Discussion appropriately cautions about this, but the paper still calls the UltraMax comparison 'the first such independent evaluation' in §4. If the UltraMax result is to be more than a case study, please provide displacement statistics from the acquired sequences or a justification that the induced motion is representative.
  3. [§7.4, §8.1] The GSD normalization assumes a constant camera-to-wall standoff derived from LiDAR, but the paper notes local standoff variation. Since the ranking of the 2× tiers is close (17.0–18.7 mm reprojection error), the normalization could affect these small differences. Report per-station standoff values and the resulting GSD range, and confirm that the small inter-tier differences are not within the uncertainty of this assumption. This is not a fatal issue because the two resolution-independent metrics agree with the ranking, but it needs quantification.
minor comments (5)
  1. [Table 1] The column header 'RU>100≥3 img' appears to combine two separate statistics (fraction of tie points with reconstruction uncertainty above 100 and fraction observed in ≥3 images). Please split these into two columns for clarity.
  2. [§9] The sentence 'obtaining its output requires reducing 16-bit radiometric data to 8-bit imagery' contradicts §7.2, where the authors deliberately avoid 8-bit conversion and use a float32 path. If this is intended as a general property of available models, it should be stated as such and supported by model documentation, not presented as a consequence of the present experiment.
  3. [§1, References] The text says 'Carl Frey demonstrated masterfully in 1885 [1]' but reference [1] is listed as 'K. Frey.' Please reconcile the initials.
  4. [Figure 4] The caption explains the color mapping well, but adding a color legend directly in the figure would improve readability, especially for readers viewing the figure separately from the caption.
  5. [§7.2] The paper mentions that 'roundtrip radiometric fidelity (mean absolute error, RMSE, and maximum error against the native input) was recorded at each stage' but these values are not presented. Reporting them would help assess whether the float32 roundtrip itself introduces radiometric distortion that could influence the photogrammetric comparison.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the central comparison rests on external LiDAR geometry and standard SfM metrics, not on fitted parameters or self-citation chains.

full rationale

The paper's central claim—that native thermal imagery yields the most geometrically accurate photogrammetric reconstruction and that AI super-resolution does not help—is not derived from its own inputs by construction. Camera positions are fixed using an external terrestrial LiDAR reference dataset [29]; the GSD used for resolution normalization is computed analytically from standoff, focal length, and sensor resolution, not fitted to the outcome; and the principal quality metrics (reconstruction uncertainty and multi-view consistency) are argued to be resolution-independent and are not defined in terms of the native tier's results. The fixed camera calibration is solved once on native imagery and scaled by resolution ratio, which is a controlled experimental constraint rather than a fitted prediction. Self-citations to CHEI datasets, tools, and platforms ([28], [29], [31], [33], [35]) are data or tooling citations; none supplies a theorem, fitted parameter, or prior conclusion on which the comparison logically depends. The paper openly acknowledges its main limitations: the improvised UltraMax displacement mechanism, the constant-standoff assumption in the analytic GSD, and the deliberate float32 roundtrip that preserves 16-bit radiometric data instead of converting to the AI models' native 8-bit input. These are external-validity concerns (the null result may partly reflect the integration pipeline), not circular dependencies. The Discussion's additional argument that AI SR is 'arguably negative' because it would require reducing 16-bit data to 8-bit is somewhat in tension with the experimental pipeline, but that is an inference gap, not a definitional or self-referential reduction. Overall, no circular step meets the evidentiary bar: there is no fitted parameter renamed as a prediction, no load-bearing self-citation chain, and no result equivalent to its input by construction.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The central comparison depends on several domain assumptions and one ad-hoc procedural choice, but uses no fitted parameters to derive its conclusion; the RU threshold is a hand-picked cutoff. The main load-bearing assumption is that the 16-bit float32 normalization is a fair way to run 8-bit-trained models.

free parameters (1)
  • RU threshold = 100
    Chosen threshold for classifying tie points as high-uncertainty (reconstruction uncertainty > 100). The ranking is robust across metrics, but the specific proportions depend on this hand-picked value.
assumptions (4)
  • domain assumption 16-bit float32 normalization is a valid way to deploy 8-bit-trained AI SR models (Section 7.2)
    If the models are sensitive to input distribution, the null result may be pipeline-specific.
  • domain assumption Analytic GSD from fixed geometry (8.75 mm/px at native, halving with resolution) is accurate despite local standoff variation (Sections 7.4, 8.1)
    The physical-unit normalization of reprojection error assumes constant camera-to-wall standoff; residual variation is acknowledged as a limitation.
  • standard math Multi-frame microscanning recovers real detail (theory from [14-18])
    The interpretation that UltraMax's lower RU indicates genuine detail recovery relies on this established theory.
  • ad hoc to paper The loosened tripod head reproduces hand-tremor displacement for UltraMax (Sections 7.1, 9)
    The improvised jiggle may not match the displacement magnitude/distribution FLIR's reconstruction expects; the paper itself flags this as a caveat.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Loggia dei Lanzi: AI Thermography Enhancement Comparisons through 3D Photogrammetry." pith.science (2026). https://pith.science/paper/OJPDVH2M

@misc{pith2026260802404,
  author       = {Pith},
  title        = {Pith review of: Loggia dei Lanzi: AI Thermography Enhancement Comparisons through 3D Photogrammetry},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OJPDVH2M}},
  note         = {Machine review of arXiv:2608.02404}
}
read the original abstract

The Loggia dei Lanzi in the Piazza della Signoria is one of Florence's most prominent structures visited by millions every year. Its construction history spans multiple centuries of modification. This paper presents the results of a thermal imaging campaign conducted in December 2025, using a FLIR T1020 HD camera, revealing hidden architectural features including walled-up openings and material transitions beneath the plaster surface. The favorable winter ambient conditions provided a feature-rich benchmark upon which to compare the results of enhancement algorithms and artificial intelligence models. We evaluate the application of AI-based image enhancement to thermal heritage documentation through a comparison of three tiers of image resolution in a photogrammetric Structure-from-Motion (SfM) pipeline: native resolution, FLIR's hardware-based pixel-shifted super-resolution (UltraMax), and state of the art AI-upscaled imagery models. We quantify the effect of each resolution tier on feature detection and tie-point generation, assessing whether the additional detail produced by super-resolution, whether hardware or AI-derived, translates into meaningfully denser and more accurate 3D thermal models. Our results contribute to the emerging intersection of artificial intelligence and heritage thermography by providing a direct comparison of hardware microscanning and AI super-resolution within a thermal photogrammetric workflow for cultural heritage. All datasets are made publicly available and accessible within an interactive 3D archival framework, and integrated into a custom citywide extended reality overlay application.

Figures

Figures reproduced from arXiv: 2608.02404 by the authors.

Figure 1
Figure 1. Image of Loggia in late 19th century (Source: Fratelli Alinari archive). [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Visible light panoramic mosaic (a) and thermographic mosaic (b) showing hidden [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. 3D orthomosaic of all native thermographic inputs, showing the Loggia dei Lanzi back [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: The resulting dense point cloud reconstructions from each method, colored by point [PITH_FULL_IMAGE:figures/full_fig_p014_4.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

36 extracted references · 15 canonical work pages

  1. [1]

    Frey,Die Loggia dei Lanzi zu Florenz: Eine quellenkritische Untersuchung

    K. Frey,Die Loggia dei Lanzi zu Florenz: Eine quellenkritische Untersuchung. Berlin: Wilhelm Hertz, 1885

  2. [2]

    Thermographic imaging in cultural heritage: A short review,

    S. Sfarra, C. Ibarra-Castanedo, P. Theodorakeas, N. P. Avdelidis, S. Perilli, H. Zhang, and X. P. V. Maldague, “Thermographic imaging in cultural heritage: A short review,”Diagnostics, vol. 12, no. 12, p. 3121, 2022.https://doi.org/10.3390/diagnostics12123121

    work page doi:10.3390/diagnostics12123121 2022

  3. [3]

    Infrared thermography (IRT) applications for building diagnostics: A review,

    A. Kylili, P. A. Fokaides, P. Christou, and S. A. Kalogirou, “Infrared thermography (IRT) applications for building diagnostics: A review,”Applied Energy, vol. 134, pp. 531–549, 2014. https://doi.org/10.1016/j.apenergy.2014.08.005

    work page doi:10.1016/j.apenergy.2014.08.005 2014

  4. [4]

    Locating hidden elements in walls of cultural heritage buildings by using infrared thermography,

    B. Lehmann, K. Ghazi Wakili, T. Frank, B. Vera Collado, and C. Tanner, “Locating hidden elements in walls of cultural heritage buildings by using infrared thermography,”Buildings, vol. 9, no. 2, p. 32, 2019.https://doi.org/10.3390/buildings9020032

    work page doi:10.3390/buildings9020032 2019

  5. [5]

    Introduction of deep learning in thermographic monitoring of cultural heritage and improvement by automatic thermogram 16 pre-processing algorithms,

    I. Garrido, S. Lagüela, S. Sfarra, F. J. Madruga, and P. Arias, “Introduction of deep learning in thermographic monitoring of cultural heritage and improvement by automatic thermogram 16 pre-processing algorithms,”Sensors, vol. 21, no. 3, p. 750, 2021.https://doi.org/10.3390/ s21030750

    2021

  6. [6]

    A novel application of deep learning approach over IRT images for the automated detection of rising damp on historical masonries,

    A. Tavukçuoğlu, E. N. Caner-Saltık, and B. A. Güney, “A novel application of deep learning approach over IRT images for the automated detection of rising damp on historical masonries,” Case Studies in Construction Materials, vol. 20, p. e02780, 2024. https://doi.org/10.1016/ j.cscm.2024.e02780

    2024

  7. [7]

    Automated wall moisture detection in heritage sites based on convolutional neural network (CNN) for infrared imagery,

    C.-H. Tai, Y.-C. Lin, and H.-P. Tserng, “Automated wall moisture detection in heritage sites based on convolutional neural network (CNN) for infrared imagery,”Applied Sciences, vol. 15, no. 12, p. 6495, 2025.https://doi.org/10.3390/app15126495

    work page doi:10.3390/app15126495 2025

  8. [8]

    Using thermal infrared image processing technology of computer vision to protect art and cultural heritage,

    R. Zhou, Y. Wang, and H. Li, “Using thermal infrared image processing technology of computer vision to protect art and cultural heritage,”Journal of Radiation Research and Applied Sciences, vol. 18, no. 1, p. 101934, 2025.https://doi.org/10.1016/j.jrras.2025.101934

    arXiv 2025

Show all 36 references
  1. [9]

    The secret lives of paintings,

    M. Seracini, “The secret lives of paintings,” TED Conferences, 2012.https://www.ted.com/ talks/maurizio_seracini_the_secret_lives_of_paintings

    2012

  2. [10]

    ARtifact: Tablet-based augmented reality for interactive analysis of cultural artifacts,

    D. Vanoni, M. Seracini, and F. Kuester, “ARtifact: Tablet-based augmented reality for interactive analysis of cultural artifacts,” inProc. 2012 IEEE International Symposium on Multimedia (ISM ’12), pp. 44–49, IEEE, 2012.https://doi.org/10.1109/ISM.2012.17

    2012 doi

  3. [11]

    Protagonists of their own discovery: Engaging the European public with Florentine her- itage,

    D. Mangano, S. McAvoy, G. Bent, M. Seracini, D. Pfaff, A. Agarwal, and F. Kuester, “Protagonists of their own discovery: Engaging the European public with Florentine her- itage,”International Archives of the Photogrammetry, Remote Sensing and Spatial Infor- mation Sciences, vo...

    2025

  4. [12]

    Automatic façade modelling using point cloud data for energy-efficient retrofitting,

    M. Previtali, L. Barazzetti, R. Brumana, B. Cuca, D. Oreni, F. Roncoroni, and M. Scaioni, “Automatic façade modelling using point cloud data for energy-efficient retrofitting,”Applied Geomatics, vol. 6, pp. 95–113, 2013.https://doi.org/10.1007/s12518-014-0129-9

    2013 doi

  5. [13]

    Efficient osmosis filtering of thermal-quasi reflectography images for cultural heritage,

    S. Parisotto, L. Calatroni, and C. Daffara, “Efficient osmosis filtering of thermal-quasi reflectography images for cultural heritage,” inProc. International Conference on Scale Space and Variational Methods in Computer Vision, pp. 447–458, Springer, 2017. https: //doi.org/10....

    2017 doi

  6. [14]

    Multiframe image restoration and registration,

    R. Y. Tsai and T. S. Huang, “Multiframe image restoration and registration,” inAdvances in Computer Vision and Image Processing, vol. 1, pp. 317–339, JAI Press, 1984

    1984

  7. [15]

    Improving resolution by image registration,

    M. Irani and S. Peleg, “Improving resolution by image registration,”CVGIP: Graphical Models and Image Processing, vol. 53, no. 3, pp. 231–239, 1991.https://doi.org/10.1016/ 1049-9652(91)90045-L

    1991

  8. [16]

    Restoration of a single superresolution image from several blurred, noisy, and undersampled measured images,

    M. Elad and A. Feuer, “Restoration of a single superresolution image from several blurred, noisy, and undersampled measured images,”IEEE Transactions on Image Processing, vol. 6, no. 12, pp. 1646–1658, 1997.https://doi.org/10.1109/83.650118

    1997 doi

  9. [17]

    The perception-distortion tradeoff,

    Y. Blau and T. Michaeli, “The perception-distortion tradeoff,” inProc. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 6228–6237, 2018. https://doi. org/10.1109/CVPR.2018.00652

    2018

  10. [18]

    UltraMax — The ultimate resolution,

    FLIR Systems, “UltraMax — The ultimate resolution,” Technical brief, 2020.https://www. flir.com/discover/rd-science/ultramax--the-ultimate-resolution/ 17

    2020

  11. [19]

    Thermal image super-resolution: A survey,

    Y. He, S. Huang, and T. Wang, “Thermal image super-resolution: A survey,” arXiv preprint, 2023.https://github.com/yongsongH/Infrared_Image_SR_Survey

    2023

  12. [20]

    Infrared image super resolution by combining compressive sensing and deep learning,

    X. Zhang, C. Li, Q. Meng, S. Liu, Y. Zhang, and J. Wang, “Infrared image super resolution by combining compressive sensing and deep learning,”Sensors, vol. 18, no. 8, p. 2587, 2018. https://doi.org/10.3390/s18082587

    2018 doi

  13. [21]

    Swin Transformer: Hierarchical vision transformer using shifted windows,

    Z. Liu, Y. Lin, Y. Cao, H. Hu, Y. Wei, Z. Zhang, S. Lin, and B. Guo, “Swin Transformer: Hierarchical vision transformer using shifted windows,” inProc. IEEE/CVF International Con- ference on Computer Vision, pp. 10012–10022, 2021. https://doi.org/10.1109/ICCV48922. 2021.00986

    2021

  14. [22]

    SwinIR: Image restoration using Swin Transformer,

    J. Liang, J. Cao, G. Sun, K. Zhang, L. Van Gool, and R. Timofte, “SwinIR: Image restoration using Swin Transformer,” inProc. IEEE/CVF International Conference on Computer Vision, pp. 1833–1844, 2021.https://arxiv.org/abs/2108.10257

    2021 arXiv

  15. [23]

    DRCT: Saving image super-resolution away from information bottleneck,

    C.-C. Hsu, C.-M. Lee, and Y.-S. Chou, “DRCT: Saving image super-resolution away from information bottleneck,” inProc. IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, pp. 6133–6142, 2024.https://doi.org/10.1109/CVPRW63382.2024. 00618

    2024

  16. [24]

    Thermal image super-resolution challenge results — PBVS 2025,

    R. E. Rivadeneira, A. D. Sappa, R. Hammoud, J. Rao, H. Zhong, Y. Wang, S. Zhao, Z. Zhong, Y.-H. Li, S. Wang, Q. Shen, H. Wang, and X. Zhang, “Thermal image super-resolution challenge results — PBVS 2025,” inProc. IEEE/CVF Conference on Computer Vision and Pattern Recognition W...

    2025

  17. [25]

    DifIISR: A diffusion model with gradient guidance for infrared image super-resolution,

    X. Li, Z. Wang, Y. Zou, Z. Chen, J. Ma, Z. Jiang, L. Ma, and J. Liu, “DifIISR: A diffusion model with gradient guidance for infrared image super-resolution,” inProc. IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 7534–7544, 2025. https://arxiv.org/abs/ 2503.01187

    2025 arXiv

  18. [26]

    AI upscaling supporting image alignment in photogrammetric reconstruction,

    S. Parrinello, A. Dell’Amico, and F. Galasso, “AI upscaling supporting image alignment in photogrammetric reconstruction,”International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, vol. XLVIII-4/2024, pp. 337–344, 2024. https: //doi.org/10.5...

    2024 doi

  19. [27]

    Deep learning-based single image super-resolution: An investigation for dense scene reconstruction with UAS photogrammetry,

    M. Pashaei, M. J. Starek, H. Kamangir, and J. Berryhill, “Deep learning-based single image super-resolution: An investigation for dense scene reconstruction with UAS photogrammetry,” Remote Sensing, vol. 12, no. 11, p. 1757, 2020.https://doi.org/10.3390/rs12111757

    2020 doi

  20. [28]

    Florence Thermography and XR, Dec 1st – Jan 5th 2025, Florence, Italy,

    S. P. McAvoy, A. Agarwal, J. Klingspon, D. Rissolo, and F. Kuester, “Florence Thermography and XR, Dec 1st – Jan 5th 2025, Florence, Italy,” Field Report, Cultural Heritage Engineering Initiative (CHEI), University of California San Diego, 2026.https://escholarship.org/uc/ ite...

    2025

  21. [29]

    Florence As It Was: Palazzo Vecchio, City Hall, Florence — LiDAR — Terrestrial,

    G. Bent, S. McAvoy, and L. Vigotti, “Florence As It Was: Palazzo Vecchio, City Hall, Florence — LiDAR — Terrestrial,” [Data set], Cultural Heritage Engineering Initiative (CHEI), Open Heritage 3D, 2026.https://doi.org/10.34946/D6W0

    2026 doi

  22. [30]

    Florence, Turin, Milan March 3rd – 7th, 2025 Italy,

    S. P. McAvoy, G. R. Bent, D. Rissolo, and F. Kuester, “Florence, Turin, Milan March 3rd – 7th, 2025 Italy,” Field Report, Cultural Heritage Engineering Initiative (CHEI), University of California San Diego, 2025.https://escholarship.org/uc/item/5ts4r5cx 18

    2025

  23. [31]

    3D Radiometric Thermography Mosaics with Low-Cost Mobile Sensor Stack,

    S. McAvoy, J. Klingspon, A. Tong, E. Lo, N. Hui, M. Seracini, D. Rissolo, N. Driscoll, and F. Kuester, “3D Radiometric Thermography Mosaics with Low-Cost Mobile Sensor Stack,” Remote Sensing, vol. 18, no. 9, p. 1335, 2026.https://doi.org/10.3390/rs18091335

    2026 doi

  24. [32]

    OpenHer- itage3D: An Open Platform for 3D Cultural Heritage Data Sharing and Visualization,

    S. McAvoy, B. Tanduo, F. Chiabrando, A. Spreafico, D. Rissolo, and F. Kuester, “OpenHer- itage3D: An Open Platform for 3D Cultural Heritage Data Sharing and Visualization,” inProc. CIPA Heritage Documentation, 2023

    2023

  25. [33]

    Loggia dei Lanzi thermography 2025 — Florence — thermography, LiDAR — terrestrial,

    S. McAvoy, G. Bent, M. Seracini, F. Kuester, et al., “Loggia dei Lanzi thermography 2025 — Florence — thermography, LiDAR — terrestrial,” [Data set], Open Heritage 3D, 2025. https://doi.org/10.34946/D66S3Q

    2025 doi

  26. [34]

    Claude (Opus 4) [Large Language Model],

    Anthropic, “Claude (Opus 4) [Large Language Model],” 2025.https://claude.ai (accessed May 22, 2026)

    2025

  27. [35]

    3D thermography: Processing scripts for the thermal super-resolution comparison pipeline,

    S. McAvoy, “3D thermography: Processing scripts for the thermal super-resolution comparison pipeline,” GitHub repository, 2026.https://github.com/smcavoy12/3d_thermography_chdc

    2026

  28. [36]

    Advancement of infrared thermography for built heritage,

    E. Rosina, M. Kavuru, and E. I. Parisi, “Advancement of infrared thermography for built heritage,” inDiagnosis of Heritage Buildings by Non-Destructive Techniques, E. Rosina, M. Kavuru, and E. I. Parisi, Eds. Sawston, UK: Woodhead Publishing (Elsevier), 2024, ch. 2, pp. 23–47....

    2024 doi

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.