Pith. sign in

REVIEW 4 major objections 6 minor 45 references

Deep Learning-Based Image Recovery and Pose Estimation for Resident Space Objects

T0 review · 4 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read The paper claims that a U-Net trained on synthetic blurred/sharp pairs, without deconvolution preprocessing, recovers ISS images and estimates pose from ground-telescope photos, cutting mean angular error by 71.9% and image-recovery MSE…

desk verdict A sensible closed-loop study of synthetic RSO image restoration for pose estimation, but the headline transfer claim rests on a single unquantified real image pair. read the letter →

arxiv 2501.13009 v1 pith:BEIU62ZA submitted 2025-01-22 cs.CV cs.LGeess.IV

Louis Aberdeen , Mark Hansen , Melvyn L. Smith , Lyndon Smith This is my paper · ORCID
classification cs.CVcs.LGeess.IV
keywords deeplearningimagerecoveryposeestimationresidentspaceobjectspointspreadfunctionU-Netdeconvolutionsynthetictrainingdata
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 claims that a deep-learning pipeline can recover usable images and estimate the pose of a resident space object from blurred ground-telescope photos, using only a small set of real star images and a publicly available 3D model for training. The authors build a synthetic dataset by rendering the International Space Station in a physics-based renderer and blurring the result with an effective point spread function constructed from 1,499 star images taken by the same optical system. They then train a U-Net to undo the blur and a ResNet50 regressor to estimate orientation. Their reported result is that U-Net image recovery alone, without pre-deconvolution, gives the best pose accuracy, reducing mean angular error by 71.9% and image-recovery mean squared error by 97.28%. If this transfer works on real imagery, it would let space-domain-awareness providers train pose estimators without large labeled datasets of actual spacecraft images.

What carries the argument

The central object is the effective point spread function (ePSF), a pixel-oversampled model of the telescope's optical response built from 1,499 bright, isolated star images using an iterative star-based construction method. The ePSF encodes atmospheric and optical blur; convolving it over renderer-produced ISS images, then adding bloom and noise, turns a physics-based 3D render into synthetic 'observed' imagery. A U-Net with encoder-decoder paths and skip connections is trained on low-quality/high-quality pairs to invert that degradation, and a ResNet50-based regressor predicts a $3\times 3$ rotation matrix, with SVD orthogonalization to enforce a proper rotation. Three Krylov-subspace deconvolution methods, all regularized by the discrepancy principle, are used as an alternative or preprocessing front end, and their comparison is what makes the U-Net-only result meaningful.

What would settle it

Take real ISS observations with independently known attitude, derived from orbit geometry or a second camera, and compare the model's predicted rotation against it; if the angular error on real images stays near the 1.460-radian level of the no-recovery control while synthetic test error is 0.414 radians, the synthetic-to-real transfer fails. A second targeted check would rebuild the ePSF from stars photographed immediately before or after the ISS pass and test whether pose error drops, which would directly test the timing mismatch in the current ePSF.

Watch

Extended reading notes

Core claim

On its own terms, the paper establishes that a U-Net trained on synthetic blurred/sharp pairs can serve as the image-recovery front end for spacecraft pose estimation, and that it outperforms the deconvolution-then-U-Net alternative. Using the ISS as a test case, the authors show that deconvolution with Tikhonov-regularized projection methods (Arnoldi-Tikhonov, Hybrid GMRES, Golub-Kahan-Tikhonov) improves rotation estimation over the blurred control, but that feeding the raw blurred image directly to U-Net is better still: mean rotation error falls to 0.414 radians versus 0.447 for deconvolution plus U-Net and 1.460 radians for the control, with a corresponding reduction in image-recovery MSE. The pose head is a ResNet50 feature extractor with a small fully connected network that outputs a $3\times 3$ rotation matrix later orthogonalized by SVD. The deeper claim is that a limited amount of real imagery of stars, not of the spacecraft itself, is enough to synthesize a training distribution that transfers to real RSO observations.

Load-bearing premise

The load-bearing premise is that the synthetic degradation pipeline—an ePSF built from stars photographed at a different time, camera parameters estimated rather than measured, and extra injected noise—matches the real telescope's blur closely enough that a network trained on synthetic data makes correct pose predictions on real ISS images.

Editorial extensions

If this is right

  • A network trained purely on synthetic ISS images can recover image detail and estimate pose from blurred ground-based imagery, with U-Net alone reducing mean angular error by 71.9% relative to no recovery.
  • Deconvolution as a preprocessing step is not necessary and can slightly hurt: Golub-Kahan-Tikhonov plus U-Net yields mean angular error 0.447 radians, while U-Net alone yields 0.414 radians.
  • Image recovery before pose regression reduces rotation-estimation MSE from 0.4020 for the control to 0.0784 for U-Net, a roughly 80% reduction.
  • Because the synthetic dataset is produced from a 3D model and a star-derived ePSF, the same recipe can be applied to other resident space objects, such as Tiangong or Hubble, provided a 3D model exists.
  • The Euler-angle rotation representation leaves a failure mode near the extremes of the rotation space, with errors approaching $\pi$ radians, a limitation the authors recommend addressing with quaternions or rotation matrices.

Reading between the lines

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

  • A natural test not reported in the paper is to freeze the trained U-Net and pose regressor and evaluate on a different RSO rendered at test time; if the image-recovery front end is object-agnostic, the pose network should degrade gracefully rather than collapse.
  • The 97.28% image-MSE figure is likely dominated by the U-Net's ability to remove the synthetic blur and noise; on real imagery, where the true clean image is unknown, image-recovery quality should be judged by downstream pose accuracy and human interpretability rather than MSE alone.
  • The framework's practical payoff is economic: a small star-image collection plus an off-the-shelf 3D model replaces the costly task of collecting thousands of labeled spacecraft photos, so viability is best measured by the ratio of real star images to achieved pose accuracy, not by absolute accuracy alone.
  • Because the paper increased training noise beyond observed levels, the model may be over-regularized for benign observing conditions; an ensemble of noise levels matched to per-observation seeing could improve transfer.
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

4 major / 6 minor

Summary. The paper presents a framework for generating synthetic training imagery of resident space objects (RSOs) by rendering a 3D model of the International Space Station in Blender, convolving with an empirical point spread function (ePSF) constructed from real star images, and adding bloom and noise. It then compares image-recovery strategies—Golub-Kahan Tikhonov deconvolution, U-Net restoration, and their combination—as preprocessing for a ResNet50-based rotation estimator. On synthetic test images, the authors report that U-Net restoration alone yields the best pose performance, reducing mean squared error in image recovery by 97.28% and average angular error by 71.9% relative to the unrecovered baseline. A single qualitative example of U-Net applied to real ISS imagery is also shown.

Significance. If the synthetic-to-real transfer claim is valid, the framework would address a practical bottleneck in space domain awareness: the scarcity of labeled RSO imagery for training pose-estimation models. The in-domain comparative study is a useful contribution, particularly the clean comparison between direct U-Net recovery and deconvolution-plus-U-Net, and the use of a real ePSF to generate degraded synthetic data is a sensible methodology. However, the headline quantitative claims are not fully supported by the reported tables, and the real-world transfer rests on a single unquantified image pair. The paper is honest about several limitations, but those limitations directly affect the central claim and need to be addressed with additional experiments or more cautious framing.

major comments (4)
  1. [Results, Table 1; Abstract] The 97.28% image-recovery MSE reduction stated in the Abstract is not traceable from the reported tables. Table 1 lists U-Net MSE=8.36 and GK-Tikhonov+U-Net MSE=8.72, but it does not report the MSE of the unrecovered degraded images that would serve as the baseline for the claimed percentage reduction. Please report the baseline MSE and the exact computation used to derive 97.28%, and state whether this is a mean over the test set or over individual images.
  2. [Results, Tables 2-3 and Figure 13] The conclusion that U-Net alone is the best pose-estimation preprocessing rests on small differences: mean angular error 0.414 vs 0.447 and rotation-estimation MSE 0.0784 vs 0.0868 for U-Net versus GK-Tikhonov+U-Net. No repeated-seed variance, confidence intervals, or statistical significance tests are reported, so it is unclear whether these differences are meaningful. Please add repeated training runs with standard deviations or confidence intervals, and, if appropriate, a paired significance test.
  3. [Methodology, 'Modeling optical degradation'; Results, Figure 12] All quantitative results in Tables 1-3 and Figures 8-13 are measured on synthetic test images degraded with the same ePSF and noise model used for training. The only real-data evidence is the single side-by-side example in Figure 12, which has no reported MSE, SSIM, PSNR, or pose error, and no pose ground truth. Consequently, the abstract's transfer claim ('transfers to the target data set') is not quantitatively supported. Either provide quantitative evaluation on real imagery—for example, image-quality metrics on available real image pairs, or a pose estimate with uncertainty—or revise the abstract and conclusion to scope the claim as synthetic evaluation with a qualitative real-image demonstration.
  4. [Methodology, 'Modeling optical degradation' and 'Dataset generation'] The synthetic degradation model relies on camera parameters that were 'estimated to achieve a final image that is as similar as possible' and training noise that was 'increased beyond that observed.' These are free parameters that directly control the difficulty of the image-recovery task and the resulting pose-estimation metrics. The paper acknowledges the temporal mismatch of the ePSF but does not quantify how sensitive the reported improvements are to these choices. Please provide a sensitivity analysis (e.g., varying the noise level and PSF mismatch) or measured noise/PSF characterization to demonstrate that the 97.28% and 71.9% improvements are not artifacts of the chosen degradation parameters.
minor comments (6)
  1. [Figure 12 caption] The caption contains a typo: 'Demonststion' should be 'Demonstration.'
  2. [Results, first paragraph] The sentence 'Results are show in Fig. 8' should read 'Results are shown in Fig. 8.'
  3. [Abstract] The phrase 'imagery, Using the International Space Station' has an incorrect comma and capitalization; it should be 'imagery. Using the International Space Station' or 'imagery, using the International Space Station.'
  4. [Methodology, Deconvolution section] 'Single Value Decomposition' should be 'Singular Value Decomposition.'
  5. [Throughout] The model name is written both as 'ResNet50' and 'Resnet50'; please use one consistent spelling.
  6. [Table 2 heading] The heading 'Error (radians)' is ambiguous because the table reports mean and standard deviation of the angular error; consider 'Rotation estimation error (radians, mean ± std)' for clarity.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the derivation chain is a standard closed-loop synthetic evaluation, with the real-domain gap explicitly acknowledged rather than hidden.

full rationale

The paper's derivation chain is self-contained and non-circular. The pipeline renders the ISS with a physics-based Blender/Cycles model, convolves the rendered images with an effective PSF estimated from 1,499 real star images, adds bloom noise, trains a U-Net for image recovery and a ResNet50 for pose regression, and evaluates on held-out synthetic images generated by the same degradation pipeline. This is a standard closed-loop evaluation, not a fitted-value tautology: the U-Net output is compared against independent ground-truth clean renders, and the pose predictions against known Euler-angle labels, so the reported 97.28% MSE reduction and 71.9% angular-error reduction are measured against targets that are not constructed from the network outputs. The ePSF is estimated from real star imagery and the camera parameters are described as "estimated to achieve a final image that is as similar as possible," but these are inputs to the synthetic data generator, not parameters fitted to the recovery or pose metrics, so no result is forced by definition. The main weakness is the extrapolation from synthetic test data to real ISS imagery: the only real-data evidence is a single unquantified side-by-side image, and the paper itself flags the temporal PSF mismatch with "Further study is needed." That is an acknowledged validation gap, not a circularity. The only self-citation, reference [31] by three of the authors, is used as a general survey citation in the Introduction and is not load-bearing for any central claim. There is no imported uniqueness theorem, no ansatz smuggled in via citation, and no known result renamed as a new one. Accordingly, the appropriate circularity score is 0.

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

The central result depends on modeling choices rather than derived constants. The most important free parameters are the hand-estimated camera parameters and the arbitrary noise augmentation level. The key axioms are the linear shift-invariant blur model, the representativeness of a temporally separated star-derived ePSF, the fidelity of the Blender render, and the validity of synthetic-to-real transfer. No new physical entities are introduced. All quantitative evaluation remains inside the synthetic distribution.

free parameters (4)
  • Estimated camera parameters = Not specified
    Methodology says 'we estimated the camera parameters to achieve a final image that is as similar as possible.' This hand-tuned fit directly controls the synthetic-to-real domain gap and is central to the transfer claim.
  • Noise augmentation level = Not specified
    Methodology says 'We increased the amount of noise in the training data beyond that observed in the actual imagery.' The scaling is chosen by hand with no sensitivity analysis.
  • Deconvolution iteration counts = 20, 20, 10
    Hyperparameters chosen 'to balance performance and quality' for Arnoldi-Tikhonov, Hybrid GMRES, and Golub-Kahan-Tikhonov. No ablation is reported.
  • Training hyperparameters = lr 0.001 for U-Net, lr 0.0001 for ResNet, 6-step lookahead
    Learning rates and lookahead steps are fixed without sensitivity analysis. They are standard choices but still free parameters for replication.
assumptions (4)
  • domain assumption RSO image degradation is linear and shift-invariant, so convolution with a single PSF models the observed imagery.
    Invoked in Methodology via Eq. (3)-(5). The paper acknowledges temporal variations and moving objects can break shift-invariance, so this is an assumption rather than a proven fact.
  • domain assumption The effective PSF built from 1,499 star images taken at a different time represents the PSF for the ISS imagery.
    Used in Methodology's Point Spread Function section. The authors note 'Further study is needed to quantify the difference in performance this causes.'
  • domain assumption The Blender Cycles render of the NASA VTAD ISS 3D model faithfully represents the real ISS appearance under solar illumination.
    The synthetic dataset is generated from this model alone, and no quantitative comparison between rendered and real ISS images is provided.
  • domain assumption Models trained on synthetic blurred ISS images transfer to real ground-based ISS imagery.
    This is the central transfer assumption. The only real-data evidence is one qualitative before-and-after figure with no pose labels, so the assumption is not independently tested.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Deep Learning-Based Image Recovery and Pose Estimation for Resident Space Objects." pith.science (2026). https://pith.science/paper/BEIU62ZA

@misc{pith2026250113009,
  author       = {Pith},
  title        = {Pith review of: Deep Learning-Based Image Recovery and Pose Estimation for Resident Space Objects},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BEIU62ZA}},
  note         = {Machine review of arXiv:2501.13009}
}
read the original abstract

As the density of spacecraft in Earth's orbit increases, their recognition, pose and trajectory identification becomes crucial for averting potential collisions and executing debris removal operations. However, training models able to identify a spacecraft and its pose presents a significant challenge due to a lack of available image data for model training. This paper puts forth an innovative framework for generating realistic synthetic datasets of Resident Space Object (RSO) imagery. Using the International Space Station (ISS) as a test case, it goes on to combine image regression with image restoration methodologies to estimate pose from blurred images. An analysis of the proposed image recovery and regression techniques was undertaken, providing insights into the performance, potential enhancements and limitations when applied to real imagery of RSOs. The image recovery approach investigated involves first applying image deconvolution using an effective point spread function, followed by detail object extraction with a U-Net. Interestingly, using only U-Net for image reconstruction the best pose performance was attained, reducing the average Mean Squared Error in image recovery by 97.28% and the average angular error by 71.9%. The successful application of U-Net image restoration combined with the Resnet50 regression network for pose estimation of the International Space Station demonstrates the value of a diverse set of evaluation tools for effective solutions to real-world problems such as the analysis of distant objects in Earth's orbit.

Figures

Figures reproduced from arXiv: 2501.13009 by the authors.

Figure 2
Figure 2. For a comprehensive overview, we direct readers to [4, [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 4
Figure 4. Left the actual imagery of the ISS, right the emulated imagery of the ISS. Deconvolution for image deblurring We use convolution to model the blurring observed in our RSO imagery, in the process we collect a PSF for our optical sys￾tem. It is natural to investigate how we might use this effective PSF for image recovery by image deconvolution. This essentially attempts to invert the PSF convolution to recover the ori… view at source ↗
Figure 1
Figure 1. A sample of 25 stars collected using the optical [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figures from the paper (7 more)
Figure 6
Figure 6. Figure 6: The convolution block comprises the majority of [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]
Figure 5
Figure 5. Figure 5: U-Net architecture used for image reconstruction [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 9
Figure 9. Figure 9: top left A box plot comparison of the Mean Squared Error of image recovery approaches; top right A box plot comparison of the peak Signal to Noise Ratio of image recovery approaches; bottom A box plot comparison of the Structural Similarity Index Measure of image recov…
Figure 8
Figure 8. Figure 8: top left A box plot comparison of the Mean Squared Error of deconvolution algorithms; top right A box plot comparison of the peak Signal to Noise Ratio of decon￾volution algorithms; bottom A box plot comparison of the Structural Similarity Index Measure of deconvolutio…
Figure 12
Figure 12. Figure 12: Demonststion of U-Net applied to real RSO im [PITH_FULL_IMAGE:figures/full_fig_p007_12.png]
Figure 10
Figure 10. Figure 10: Demonstration of U-Net applied to simulated [PITH_FULL_IMAGE:figures/full_fig_p007_10.png]
Figure 11
Figure 11. Figure 11: Demonstration of U-Net applied to simulated [PITH_FULL_IMAGE:figures/full_fig_p007_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

45 extracted references · 25 canonical work pages

  1. [1]

    ESA’S ANNUAL SPACE ENVIRONMENT REPORT

    European Space Agency. ESA’S ANNUAL SPACE ENVIRONMENT REPORT . 2024. url: https : / / www . sdo . esoc . esa . int / environment _ report / Space_Environment_Report_latest.pdf

    work page 2024

  2. [2]

    Deep learning-based galaxy im- age deconvolution

    Utsav Akhaury et al. “Deep learning-based galaxy im- age deconvolution”. In: Frontiers in Astronomy and Space Sciences 9 (Nov. 2022). issn: 2296-987X. doi: 10.3389/fspas.2022.1001043 . url: http://dx. doi.org/10.3389/fspas.2022.1001043

    arXiv 2022

  3. [3]

    Empirical Models for the WFC3/IR PSF

    Jay Anderson. Empirical Models for the WFC3/IR PSF. Instrument Science Report WFC3 2016-12, 42 pages. Mar. 2016

    work page 2016

  4. [4]

    Toward High- Precision Astrometry with WFPC2. I. Deriving an Accurate Point-Spread Function

    Jay Anderson and Ivan R. King. “Toward High- Precision Astrometry with WFPC2. I. Deriving an Accurate Point-Spread Function”. In: Publications of the Astronomical Society of the Pacific 112.776 (Oct. 2000), pp. 1360–1382. issn: 1538-3873. doi: 10.1086/ 316632. url: http://dx.doi.org/10.1086/316632

    doi:10.1086/316632

  5. [5]

    International Space Station 3D Model

    NASA Visualization Technology Applications and Development (VTAD). International Space Station 3D Model. Apr. 2019. url: https://science.nasa. gov / resource / international - space - station - 3d-model/

    work page 2019

  6. [6]

    Generalized PSF modeling for optimized quantitation in PET imaging

    Saeed Ashrafinia et al. “Generalized PSF modeling for optimized quantitation in PET imaging”. In: Physics in Medicine & Biology 62.12 (May 2017), p. 5149. doi: 10.1088/1361- 6560/aa6911 . url: https:// dx.doi.org/10.1088/1361-6560/aa6911

    doi:10.1088/1361- 2017

  7. [7]

    Real-Time, Flight-Ready, Non- Cooperative Spacecraft Pose Estimation Using Monocular Imagery

    Kevin Black et al. Real-Time, Flight-Ready, Non- Cooperative Spacecraft Pose Estimation Using Monocular Imagery . 2021. arXiv: 2101 . 09553 [cs.CV]. url: https : / / arxiv . org / abs / 2101 . 09553

    work page 2021

  8. [8]

    Spatial resolu- tion in optical transition radiation beam diagnostics

    M. Castellano and V. A. Verzilov. “Spatial resolu- tion in optical transition radiation beam diagnostics”. In: Phys. Rev. ST Accel. Beams 1 (6 Oct. 1998), p. 062801. doi: 10 . 1103 / PhysRevSTAB . 1 . 062801. url: https : / / link . aps . org / doi / 10 . 1103 / PhysRevSTAB.1.062801

    work page 1998

Show all 45 references
  1. [9]

    Learning to See in the Dark

    Chen Chen et al. Learning to See in the Dark . 2018. arXiv: 1805.01934 [cs.CV]. url: https://arxiv. org/abs/1805.01934

    2018 arXiv

  2. [10]

    Simulation of RSO Images for Space Situation Awareness (SSA) Using Parallel Pro- cessing

    Ryan Clark et al. “Simulation of RSO Images for Space Situation Awareness (SSA) Using Parallel Pro- cessing”. In: Sensors 21.23 (2021). issn: 1424-8220. doi: 10.3390/s21237868. url: https://www.mdpi. com/1424-8220/21/23/7868

    2021 doi

  3. [11]

    Blender - a 3D modelling and rendering package

    Blender Online Community. Blender - a 3D modelling and rendering package. Blender Foundation. Stichting Blender Foundation, Amsterdam, 2018. url: http: //www.blender.org

    2018

  4. [12]

    Wavelet Transforms and Their Applications

    Lokenath Debnath and Jean-Pierre Antoine. “Wavelet Transforms and Their Applications”. In: Physics Today - PHYS TODAY 56 (Apr. 2003), pp. 68–68. doi: 10.1063/1.1580056

    2003 doi

  5. [13]

    ImageNet: A large-scale hierarchical image database

    Jia Deng et al. “ImageNet: A large-scale hierarchical image database”. In: 2009 IEEE Conference on Com- puter Vision and Pattern Recognition. 2009, pp. 248–

    2009

  6. [14]

    Goodfellow et al

    Ian J. Goodfellow et al. Generative Adversarial Net- works. 2014. arXiv: 1406 . 2661 [stat.ML]. url: https://arxiv.org/abs/1406.2661

    2014 arXiv

  7. [15]

    Deep Residual Learning for Image Recognition

    Kaiming He et al. Deep Residual Learning for Image Recognition. 2015. arXiv: 1512.03385 [cs.CV]. url: https://arxiv.org/abs/1512.03385. 8

    2015 arXiv

  8. [16]

    Solar Irradiance Capturing in Cloudy Sky Days–A Convolutional Neural Network Based Image Regression Approach

    Huaiguang Jiang et al. “Solar Irradiance Capturing in Cloudy Sky Days–A Convolutional Neural Network Based Image Regression Approach”. In: IEEE Access 8 (2020), pp. 22235–22248. doi: 10 . 1109 / ACCESS . 2020.2969549

    2020

  9. [17]

    Hailstone

    Surya Kamal and Richard K. Hailstone. Point- Spread Function of the Optics in Scanning Elec- tron Microscopes . 2024. arXiv: 2407 . 01439 [cond-mat.mtrl-sci]. url: https://arxiv.org/ abs/2407.01439

    2024 arXiv

  10. [18]

    Satellite Pose Estimation Chal- lenge: Dataset, Competition Design, and Results

    Mate Kisantal et al. “Satellite Pose Estimation Chal- lenge: Dataset, Competition Design, and Results”. In: IEEE Transactions on Aerospace and Electronic Systems 56.5 (Oct. 2020), pp. 4083–4098. issn: 2371-

    2020

  11. [19]

    20 years of Hubble Space Telescope optical modeling us- ing Tiny Tim

    John E. Krist, Richard N. Hook, and Felix Stoehr. “20 years of Hubble Space Telescope optical modeling us- ing Tiny Tim”. In: Optical Modeling and Performance Predictions V. Ed. by Mark A. Kahan. Vol. 8127. In- ternational Society for Optics and Photonics. SPIE, 2011, 81270J. ...

    2011 doi

  12. [20]

    Photo-Realistic Single Im- age Super-Resolution Using a Generative Adversar- ial Network

    Christian Ledig et al. Photo-Realistic Single Im- age Super-Resolution Using a Generative Adversar- ial Network. 2017. arXiv: 1609.04802 [cs.CV]. url: https://arxiv.org/abs/1609.04802

    2017 arXiv

  13. [21]

    Tikhonov type regularization methods: History and recent progress

    Frank Lenzen and Otmar Scherzer. “Tikhonov type regularization methods: History and recent progress”. In: Proceeding Eccomas 2004 (Jan. 2004)

    2004

  14. [22]

    An Analysis of SVD for Deep Ro- tation Estimation

    Jake Levinson et al. An Analysis of SVD for Deep Ro- tation Estimation. 2020. arXiv: 2006.14616 [cs.CV]. url: https://arxiv.org/abs/2006.14616

    2020 arXiv

  15. [23]

    SPEED+: Next-Generation Dataset for Spacecraft Pose Estimation across Do- main Gap

    Tae Ha Park et al. “SPEED+: Next-Generation Dataset for Spacecraft Pose Estimation across Do- main Gap”. In: 2022 IEEE Aerospace Confer- ence (AERO) . IEEE, Mar. 2022. doi: 10 . 1109 / aero53065 . 2022 . 9843439. url: http : / / dx . doi . org/10.1109/AERO53065.2022.9843439

    2022

  16. [24]

    TRIPs-Py: Techniques for Regu- larization of Inverse Problems in Python

    Mirjeta Pasha et al. TRIPs-Py: Techniques for Regu- larization of Inverse Problems in Python. 2024. arXiv: 2402.17603 [math.NA]. url: https://arxiv.org/ abs/2402.17603

    2024 arXiv

  17. [25]

    U-Net: Convolutional Networks for Biomedical Im- age Segmentation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U-Net: Convolutional Networks for Biomedical Im- age Segmentation. 2015. arXiv: 1505.04597 [cs.CV]. url: https://arxiv.org/abs/1505.04597

    2015 arXiv

  18. [26]

    Lunar surface image restoration using U-net based deep neural networks

    Hiya Roy et al. Lunar surface image restoration using U-net based deep neural networks. 2019. arXiv: 1904. 06683 [cs.CV]. url: https : / / arxiv . org / abs / 1904.06683

    2019 arXiv

  19. [27]

    Generative Ad- versarial Networks (GANs Survey): Challenges, So- lutions, and Future Directions

    Divya Saxena and Jiannong Cao. Generative Ad- versarial Networks (GANs Survey): Challenges, So- lutions, and Future Directions . 2023. arXiv: 2005 . 00065 [cs.LG]. url: https : / / arxiv . org / abs / 2005.00065

    2023 arXiv

  20. [28]

    Transfer Learning and its appli- cation in Computer Vision: A Review

    Chandeep Sharma. “Transfer Learning and its appli- cation in Computer Vision: A Review”. In: Mar. 2022

    2022

  21. [29]

    Pose Estima- tion for Non-Cooperative Rendezvous Using Neural Networks

    Sumant Sharma and Simone D’Amico. Pose Estima- tion for Non-Cooperative Rendezvous Using Neural Networks. 2019. arXiv: 1906 . 09868 [cs.CV]. url: https://arxiv.org/abs/1906.09868

    2019 arXiv

  22. [30]

    Very Deep Convolutional Networks for Large-Scale Image Recog- nition

    Karen Simonyan and Andrew Zisserman. Very Deep Convolutional Networks for Large-Scale Image Recog- nition. 2015. arXiv: 1409.1556 [cs.CV]. url: https: //arxiv.org/abs/1409.1556

    2015 arXiv

  23. [31]

    The quiet revolution in machine vision - a state-of-the-art survey paper, including historical re- view, perspectives, and future directions

    Melvyn L. Smith, Lyndon N. Smith, and Mark F. Hansen. “The quiet revolution in machine vision - a state-of-the-art survey paper, including historical re- view, perspectives, and future directions”. In: Com- puters in Industry 130 (2021), p. 103472. issn: 0166-

    2021

  24. [32]

    Space-track.org: Website for space object tracking data. 2024. url: https://www.space-track.org

    2024

  25. [33]

    Mingxing Tan and Quoc V. Le. EfficientNet: Rethink- ing Model Scaling for Convolutional Neural Networks

  26. [34]

    ESRGAN: Enhanced Super- Resolution Generative Adversarial Networks

    Xintao Wang et al. ESRGAN: Enhanced Super- Resolution Generative Adversarial Networks . 2018. arXiv: 1809.00219 [cs.CV]. url: https://arxiv. org/abs/1809.00219

    2018 arXiv

  27. [35]

    Real-ESRGAN: Training Real- World Blind Super-Resolution with Pure Synthetic Data

    Xintao Wang et al. Real-ESRGAN: Training Real- World Blind Super-Resolution with Pure Synthetic Data. 2021. arXiv: 2107 . 10833 [eess.IV]. url: https://arxiv.org/abs/2107.10833

    2021 arXiv

  28. [36]

    Image quality assessment: from error visibility to structural similarity

    Zhou Wang et al. “Image quality assessment: from error visibility to structural similarity”. In: IEEE Transactions on Image Processing 13.4 (2004), pp. 600–612. doi: 10.1109/TIP.2003.819861

    2004

  29. [37]

    Theoretical and Numerical Study of Tikhonov’s Regularization and Morozov’s Discrepancy Principle

    MaryGeorge L. Whitney. “Theoretical and Numerical Study of Tikhonov’s Regularization and Morozov’s Discrepancy Principle.” In: Georgia State University (2009). url: https://doi.org/10.57709/1234875

    2009 doi

  30. [38]

    Pixel-wise regression using U-Net and its application on pansharpening

    Wei Yao et al. “Pixel-wise regression using U-Net and its application on pansharpening”. In: Neurocomput- ing 312 (2018), pp. 364–371. issn: 0925-2312. doi: https : / / doi . org / 10 . 1016 / j . neucom . 2018 . 05 . 103. url: https : / / www . sciencedirect . com / science/a...

    2018

  31. [39]

    Direct estimation of fetal head cir- cumference from ultrasound images based on regres- sion CNN

    Jing Zhang et al. “Direct estimation of fetal head cir- cumference from ultrasound images based on regres- sion CNN”. In: Proceedings of the Third Conference on Medical Imaging with Deep Learning . Ed. by Tal Arbel et al. Vol. 121. Proceedings of Machine Learn- ing Research. P...

    2020

  32. [40]

    A CNN-based regression framework for estimating coal ash content on mi- croscopic images

    Kanghui Zhang et al. “A CNN-based regression framework for estimating coal ash content on mi- croscopic images”. In: Measurement 189 (2022), p. 110589. issn: 0263-2241. doi: https : / / doi . org / 10 . 1016 / j . measurement . 2021 . 110589. url: https://www.sciencedirect.com...

    2022

  33. [41]

    A Comprehensive Survey on Transfer Learning

    Fuzhen Zhuang et al. A Comprehensive Survey on Transfer Learning. 2020. arXiv: 1911.02685 [cs.LG]. url: https://arxiv.org/abs/1911.02685. 10

    2020 arXiv

  34. [255]

    doi: 10.1109/CVPR.2009.5206848

    2009

  35. [2020]

    11946 [cs.LG]

    arXiv: 1905 . 11946 [cs.LG]. url: https : / / arxiv.org/abs/1905.11946

    1905 arXiv

  36. [3615]

    2021.103472

    doi: https://doi.org/10.1016/j.compind. 2021.103472 . url: https://www.sciencedirect. com/science/article/pii/S0166361521000798

    2021

  37. [9877]

    url: http: //dx.doi.org/10.1109/TAES.2020.2989063

    doi: 10.1109/taes.2020.2989063. url: http: //dx.doi.org/10.1109/TAES.2020.2989063

    2020

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.