Pith. sign in

REVIEW 4 major objections 5 minor 33 references

A generative approach for lensless imaging in low-light conditions

T0 review · 4 major / 5 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 A two-stage pipeline combining a learnable Wiener filter with a wavelet-domain diffusion model can reconstruct clear, bright images from low-light lensless measurements.

desk verdict A genuinely useful two-stage lensless low-light reconstruction method, with real hardware experiments, but the measured-data validation may have reused training images and the 'outperforms all metrics' claim is contradicted by the paper's own tables. read the letter →

arxiv 2501.03511 v1 pith:DNKSM4WT submitted 2025-01-07 eess.IV

Ziyang Liu , Tianjiao Zeng , Xu Zhan , Xiaoling Zhang , Edmund Y. Lam This is my paper · ORCID
classification eess.IV
keywords lenslessimaginglow-lightconditionaldiffusionmodelWienerfilterwavelettransformcomputationalimagereconstructiongenerative
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 is trying to establish that lensless cameras—which capture light through a mask instead of a lens—can still produce high-quality images in dim environments, a regime where existing reconstruction methods fail. The proposed answer is a two-stage pipeline: a physics-based learnable Wiener filter first turns raw noisy measurements into a rough, noisy image, then a conditional diffusion model operating in the wavelet domain removes the noise, restores brightness, and fills in missing detail. The authors argue that this combination of model-driven first stage and generative second stage is what makes the method robust under photon-starved conditions, and they support it with simulations and a self-built camera experiment. If true, it would make lightweight, lensless imagers practical in low-light settings such as space or night-time inspection.

What carries the argument

The load-bearing mechanism is the two-stage reconstruction with four linked components. First, the learnable Wiener filter implements the pseudo-inverse of the forward model $\mathbf{b} = \mathbf{H}\mathbf{x}$ in the frequency domain, with the PSF $\mathbf{h}$ initialized from calibration and treated as learnable, producing an initial image $\hat{x}$. Second, the Haar discrete wavelet transform splits this image into low-frequency (LL) and high-frequency (LH, HL, HH) sub-bands, reducing spatial resolution fourfold and separating brightness/noise from texture. Third, a conditional diffusion model, trained with both forward diffusion and reverse denoising (bidirectional training), is conditioned on the LL sub-band to generate a clean, bright low-frequency component. Fourth, a depthwise-separable convolutional network with cross-attention between the high-frequency sub-bands restores fine details. The loss combines noise-prediction MSE, reconstruction MAE+SSIM+LPIPS, and a high-frequency MSE+TV term.

What would settle it

Measure the reconstruction quality of a model trained with the paper's simulated noise when the real camera is operated at a different temperature, with a different mask-sensor distance, or with a scene containing depth variation; if the PSNR/SSIM margin over simple Wiener deconvolution shrinks to roughly zero, the calibrated-PSF and simulated-noise assumptions are the load-bearing parts.

Watch

Extended reading notes

Core claim

The central claim is that low-light lensless reconstruction is best tackled by separating the problem into two guided steps rather than trying to invert the noisy forward model in one go. The first step uses the known point-spread function to project measurements into the range space of the adjoint operator via a learnable Wiener filter, yielding an initial reconstruction that is noisy and dim but encodes the image content. The second step refines this in the wavelet domain: a conditional diffusion model cleans the low-frequency sub-band (where brightness and noise live) while a depthwise-separable convolutional network with cross-attention restores high-frequency texture. The authors report that this method outperforms both traditional optimization (ADMM) and learned baselines (U-Net, FlatNet, MWDN, DeepLIR) under simulated and real low-light conditions, with particular gains in brightness, noise suppression, and perceptual quality.

Load-bearing premise

The method assumes that the calibrated point-spread function and the simulated camera-noise model match the real sensor's behaviour; if those disagree with the actual optics and electronics, both the Wiener first stage and the diffusion model trained on simulated data will degrade.

Editorial extensions

If this is right

  • If the method is correct, a lensless camera with a simple amplitude mask and a modest CMOS sensor can operate at exposure times of 0.3 to 0.7 seconds and still yield images close to ground truth in brightness and detail.
  • The two-stage design implies that the same diffusion refinement can be reused for different mask designs or sensor noise levels, as long as the first-stage Wiener filter is re-calibrated to the new PSF.
  • The wavelet-domain decomposition reduces memory and computation enough that the diffusion process runs at roughly 0.4 seconds per image, making the approach plausible for field or near-real-time deployments.
  • The method's robustness across exposure times suggests that the generative second stage, not the raw measurement SNR, is the main driver of final image quality.

Reading between the lines

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

  • A natural extension is to replace the fixed calibrated PSF with a PSF that is jointly estimated or made uncertainty-aware; if the first stage becomes robust to PSF error, the same pipeline could apply to out-of-focus or thermal-drift conditions without retraining.
  • The wavelet-domain conditioning could be generalized to other inverse problems with strong noise, such as low-light phase retrieval or holography, since the brightness/noise separation is not specific to lensless optics.
  • The paper's simulated dataset only covers images displayed on an LCD screen; a harder test would be natural scenes with depth variation, where the shift-invariant PSF assumption is violated, and the method may need a spatially-variant extension.
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 / 5 minor

Summary. The paper proposes a two-stage reconstruction method for lensless imaging in low-light conditions. The first stage applies a learnable Wiener filter based on a calibrated point-spread function to obtain an initial noisy reconstruction; the second stage refines it with a conditional diffusion model operating in a wavelet domain, separating low-frequency and high-frequency processing. The method is evaluated on simulated data derived from the LOLv2 dataset and on real measurements from a self-built lensless camera, with comparisons against ADMM, U-Net, FlatNet, MWDN, and DeepLIR. The authors claim substantial improvements in brightness, noise suppression, and sharpness, and state that the proposed method outperforms all baselines across all metrics.

Significance. If the claims are sustained, the work is a useful contribution to low-light lensless imaging: it combines a physics-based first stage with a generative refinement stage, includes a systematic camera noise model, and validates on a real hardware prototype. The explicit treatment of low-light noise, the wavelet-domain diffusion design, and the comparison against several recent baselines are strengths. However, the evaluation is currently undermined by an unsupported 'all metrics' claim, missing statistical characterization, and an unclarified relationship between the training set and the real-world test targets. The absence of code or data makes these points unverifiable from the manuscript alone.

major comments (4)
  1. [§4.4, Tables 1 and 3] The text states that the proposed method 'outperforms the others across all metrics,' but the authors' own tables contradict this. In Table 1, MWDN achieves higher SSIM (0.6115) than the proposed method (0.5719); in Table 3, MWDN again has higher SSIM (0.5275 vs. 0.4951). The claim must be revised to identify the specific metrics on which the method improves, and the SSIM deficit should be discussed rather than ignored.
  2. [§4.1 and §4.5] The real-world validation does not establish that the measured target images are disjoint from the training set. Section 4.1 states that 1000 LOLv2 pairs are split into 900 training and 100 test pairs for simulation, and that the real-world validation uses the LOLv2 Synthetic subset. Section 4.5 then says target images are switched 'until all target images in the original dataset have been traversed' without specifying that these targets are exactly the 100 held-out test images or that they are not among the 900 training images. If training clean images were among the projected targets, the conditional diffusion model has already seen the ground-truth content, which would inflate the real-world results in Table 4 and Table 5 and make the comparison against non-trained baselines unfair. The manuscript must specify the exact target set and confirm its disjointness from the training set.
  3. [Tables 1, 3, 4, and 5] All quantitative results are reported as averages without error bars, confidence intervals, or the number of images per entry. Several between-method differences are small (e.g., PSNR differences around 0.2 dB), and without statistical characterization it is not possible to determine whether these differences are meaningful. Please provide per-image score distributions, error bars, or a statistical significance test, and state how many images each average covers.
  4. [§3.1, Eq. (6), and §4.5] The method's first stage relies on the calibrated PSF in the forward model b = Hx, and the paper itself notes that model mismatch is a known issue. The experiments do not quantify sensitivity to PSF calibration error or to deviations of the real camera noise from the simulation parameters in Table 2. Because the diffusion model is trained on simulated data generated with this forward model, a robustness test with perturbed PSFs or with an independent noise parameter setting would strengthen the claim that the method works under realistic model mismatch.
minor comments (5)
  1. [§3.1] The heading contains a typo: 'Fisrt Stage' should be 'First Stage'.
  2. [§3.2.1, Eq. (12)] The displayed formula has malformed parentheses: 'p_θ(xt−1)| xt, s)' should likely be 'p_θ(xt−1 | xt, s)'.
  3. [§4.2, §4.4, and §4.5] The sensor specifications are inconsistent across sections: §4.2 reports an IMX219 sensor with 1.12 µm pixel size and 2028×1520×4 Bayer input, §4.4 reports a pixel pitch of 0.014 mm, and §4.5 reports a raw resolution of 4056×3040. Please clarify the actual sensor geometry and the relationship between these numbers.
  4. [Tables 3, 4, and 5] All three table captions say 'simulation test set,' but Table 4 reports measured real-world data and Table 5 reports varying-exposure measured data. The captions should be corrected.
  5. [Table 1] The U-Net row consists of dashes with no explanation. If U-Net failed to train or was excluded, this should be stated explicitly; otherwise the reader cannot interpret the comparison.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity; the core pipeline is a supervised physics-guided reconstruction validated on measured data, with only contextual self-citation and a mild simulation self-consistency that does not make the central claim circular.

full rationale

The paper's derivation chain is not circular. The first stage uses a learnable Wiener filter based on the calibrated PSF (Eq. 8), and the second stage is a conditional diffusion model trained with supervised pairs where the condition is the initial reconstruction s (Eq. 12) and the losses (Eqs. 14-16) compare predictions to ground truth. No equation identifies the final reconstruction with a fitted parameter or with the input by construction. The self-citation of the authors' prior work [19] is contextual, motivating model-mismatch handling, but it is not load-bearing for the proposed architecture or its evaluation. The simulation experiments use the same camera-noise model for training and simulated testing, which is a mild self-consistency, but the measured experiments in Section 4.5 provide independent real-world validation. The manuscript leaves unspecified whether the real-world projected target images are exactly the held-out 100 test images or include the 900 training images; if the latter, this would be a data-hygiene concern, but the text does not establish that overlap, and data leakage is not a derivation-level circularity. Under the no-speculation rule, that ambiguity does not raise the circularity score. The internal inconsistency that the text claims 'outperforms all other methods across all metrics' despite Tables 1 and 3 showing MWDN with higher SSIM is a correctness/consistency issue, not circularity.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The central claim rests on the accuracy of the forward model and noise model, and on hand-chosen Wiener filter lambda. No new physical entities are introduced. The axioms listed are the main unverified premises on which the method and its evaluation depend.

free parameters (2)
  • lambda (Wiener filter noise factor) = 50000 (initial), 80000 (noisier scenes)
    Hand-set hyperparameter controlling noise suppression in the first-stage reconstruction. The paper says it is 'fixed in the experiment' then 'increased to 80000 for noisier scenes' (Section 4.2). It is chosen by the authors, not fit to data.
  • Camera simulation noise parameters (Table 2) = max intensity 1000, quantum efficiency 0.7, read noise std 2.63, ADU 0.23, baseline ADU 4.48, bits 8
    Assumed camera characteristics used to generate the simulated low-light lensless dataset for training and evaluation. These values affect the training distribution and are not derived from a measurement procedure described in the paper.
assumptions (5)
  • domain assumption Forward model b = Hx with H the calibrated PSF convolution (Eq. 6, Section 3.1)
    Assumes linear shift-invariance and accurate PSF calibration. The paper acknowledges model mismatch is a known issue (ref [19]) that degrades reconstruction quality.
  • domain assumption Camera noise model in Section 2 (Poisson shot noise, Gaussian readout noise, quantization noise, Eqs. 1-5)
    Assumes the real sensor noise combines these components with parameters from Table 2. The fidelity of this model to actual hardware is load-bearing for the simulated training data.
  • standard math Diffusion forward process is Gaussian (Eq. 9, from Ho et al. [27])
    Standard denoising diffusion probabilistic model assumption; the paper relies on this for the conditional diffusion module.
  • standard math Haar wavelet transform is invertible and separates low and high frequency content (Eq. 13)
    Used to decompose the initial reconstruction into LL and HF sub-bands for separate processing. A standard mathematical property.
  • domain assumption The diffusion model trained on simulated low-light data generalizes to real low-light measurements
    The paper trains on simulated data generated from the Section 2 noise model and tests on real camera data. Generalization depends on the simulation matching real noise characteristics and scene statistics.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A generative approach for lensless imaging in low-light conditions." pith.science (2026). https://pith.science/paper/DNKSM4WT

@misc{pith2026250103511,
  author       = {Pith},
  title        = {Pith review of: A generative approach for lensless imaging in low-light conditions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DNKSM4WT}},
  note         = {Machine review of arXiv:2501.03511}
}
read the original abstract

Lensless imaging offers a lightweight, compact alternative to traditional lens-based systems, ideal for exploration in space-constrained environments. However, the absence of a focusing lens and limited lighting in such environments often result in low-light conditions, where the measurements suffer from complex noise interference due to insufficient capture of photons. This study presents a robust reconstruction method for high-quality imaging in low-light scenarios, employing two complementary perspectives: model-driven and data-driven. First, we apply a physic-model-driven perspective to reconstruct in the range space of the pseudo-inverse of the measurement model as a first guidance to extract information in the noisy measurements. Then, we integrate a generative-model based perspective to suppress residual noises as the second guidance to suppress noises in the initial noisy results. Specifically, a learnable Wiener filter-based module generates an initial noisy reconstruction. Then, for fast and, more importantly, stable generation of the clear image from the noisy version, we implement a modified conditional generative diffusion module. This module converts the raw image into the latent wavelet domain for efficiency and uses modified bidirectional training processes for stabilization. Simulations and real-world experiments demonstrate substantial improvements in overall visual quality, advancing lensless imaging in challenging low-light environments.

Figures

Figures reproduced from arXiv: 2501.03511 by the authors.

Figure 1
Figure 1. The measurement process in low-light conditions, illustrating the mixture of [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Overview of the reconstruction pipeline for the proposed framework. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. The test set results for the simulated dataset, from top to bottom, are the [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Reconstruct and enhance results in a simulated dataset with camera noise, from [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]
Figure 5
Figure 5. Figure 5: Our self-built lensless imaging system. 4.5. Measured Reconstruction This section validates the proposed method through measured experiments. As shown in [PITH_FULL_IMAGE:figures/full_fig_p014_5.png]
Figure 6
Figure 6. Figure 6: The test set results of the self-built measured dataset, from top to bottom, are [PITH_FULL_IMAGE:figures/full_fig_p015_6.png]
Figure 7
Figure 7. Figure 7: Reconstruct and enhance results for measured datasets acquired at varying [PITH_FULL_IMAGE:figures/full_fig_p016_7.png]
Figure 8
Figure 8. Figure 8: Reconstruction results on the diffusercam dataset with added camera noise in [PITH_FULL_IMAGE:figures/full_fig_p018_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

33 extracted references · 32 canonical work pages

  1. [1]

    High-quality computational imaging through simple lenses,

    F. Heide, M. Rouf, M. B. Hullin,et al., “High-quality computational imaging through simple lenses,” ACM Trans. on Graph. (TOG)32, 1–14 (2013)

    work page 2013

  2. [2]

    Lensless camera: Unraveling the breakthroughs and prospects,

    S. Li, Y. Gao, J. Wu,et al., “Lensless camera: Unraveling the breakthroughs and prospects,” Fundam. Res. (2024)

    work page 2024

  3. [3]

    Single-shot inline holography using a physics-aware diffusion model,

    Y. Zhang, X. Liu, and E. Y. Lam, “Single-shot inline holography using a physics-aware diffusion model,” Opt. Express 32, 10444–10460 (2024)

    work page 2024

  4. [4]

    On the use of deep learning for phase recovery,

    K. Wang, L. Song, C. Wang,et al., “On the use of deep learning for phase recovery,” Light. Sci. & Appl.13, 4 (2024)

    work page 2024

  5. [5]

    Iterative phase retrieval with a sensor mask,

    L. Song and E. Y. Lam, “Iterative phase retrieval with a sensor mask,” Opt. Express30, 25788–25802 (2022)

    work page 2022

  6. [6]

    Enhancingimage resolutionofconfocalfluorescencemicroscopywithdeeplearning,

    B.Huang, J.Li, B.Yao, et al., “Enhancingimage resolutionofconfocalfluorescencemicroscopywithdeeplearning,” PhotoniX 4, 2 (2023)

    work page 2023

  7. [7]

    Millisecond autofocusing microscopy using neuromorphic event sensing,

    Z. Ge, H. Wei, F. Xu,et al., “Millisecond autofocusing microscopy using neuromorphic event sensing,” Opt. Lasers Eng. 160, 107247 (2023)

    work page 2023

  8. [8]

    Deep unrolled low-rank tensor completion for high dynamic range imaging,

    T. T. N. Mai, E. Y. Lam, and C. Lee, “Deep unrolled low-rank tensor completion for high dynamic range imaging,” IEEE Trans. on Image Process.31, 5774–5787 (2022)

    work page 2022

Show all 33 references
  1. [9]

    Full scene underwater imaging with polarization and an untrained network,

    Y. Zhu, T. Zeng, K. Liu,et al., “Full scene underwater imaging with polarization and an untrained network,” Opt. Express 29, 41865–41881 (2021)

    2021

  2. [10]

    Flatcam: Thin, lensless cameras using coded aperture and computation,

    M. S. Asif, A. Ayremlou, A. Sankaranarayanan,et al., “Flatcam: Thin, lensless cameras using coded aperture and computation,” IEEE Trans. on Comput. Imaging3, 384–397 (2016)

    2016

  3. [11]

    Phlatcam: Designed phase-mask based thin lensless camera,

    V. Boominathan, J. K. Adams, J. T. Robinson, and A. Veeraraghavan, “Phlatcam: Designed phase-mask based thin lensless camera,” IEEE transactions on pattern analysis machine intelligence42, 1618–1629 (2020)

    2020

  4. [12]

    Recent advances in lensless imaging,

    V. Boominathan, J. T. Robinson, L. Waller, and A. Veeraraghavan, “Recent advances in lensless imaging,” Optica9, 1–16 (2021)

    2021

  5. [13]

    Autofocusing of fresnel zone aperture lensless imaging for qr code recognition,

    F. Liu, J. Wu, and L. Cao, “Autofocusing of fresnel zone aperture lensless imaging for qr code recognition,” Opt. Express 31, 15889–15903 (2023)

    2023

  6. [14]

    Assessment of lateral resolution of single random phase encoded lensless imaging systems,

    S. Goswami, P. Wani, G. Gupta, and B. Javidi, “Assessment of lateral resolution of single random phase encoded lensless imaging systems,” Opt. Express31, 11213–11226 (2023)

    2023

  7. [15]

    Dual-branch fusion model for lensless imaging,

    Y. Zhang, Z. Wu, Y. Xu, and J. Huangfu, “Dual-branch fusion model for lensless imaging,” Opt. Express31, 19463–19477 (2023)

    2023

  8. [16]

    Incoherent reconstruction-free object recognition with mask-based lensless optics and the transformer,

    X. Pan, X. Chen, T. Nakamura, and M. Yamaguchi, “Incoherent reconstruction-free object recognition with mask-based lensless optics and the transformer,” Opt. Express29, 37962–37978 (2021)

    2021

  9. [17]

    Learned reconstructions for practical mask-based lensless imaging,

    K. Monakhova, J. Yurtsever, G. Kuo,et al., “Learned reconstructions for practical mask-based lensless imaging,” Opt. express27, 28075–28090 (2019)

    2019

  10. [18]

    Flatnet: Towards photorealistic scene reconstruction from lensless measurements,

    S. S. Khan, V. Sundar, V. Boominathan,et al., “Flatnet: Towards photorealistic scene reconstruction from lensless measurements,” IEEE Trans. on Pattern Anal. Mach. Intell.44, 1934–1948 (2020)

    2020

  11. [19]

    Robust reconstruction with deep learning to handle model mismatch in lensless imaging,

    T. Zeng and E. Y. Lam, “Robust reconstruction with deep learning to handle model mismatch in lensless imaging,” IEEE Trans. on Comput. Imaging7, 1080–1092 (2021)

    2021

  12. [20]

    Unrolledprimal-dualnetworksforlenslesscameras,

    O.Kingshott,N.Antipa,E.Bostan,andK.Akşit,“Unrolledprimal-dualnetworksforlenslesscameras,”Opt.Express 30, 46324–46335 (2022)

    2022

  13. [21]

    Mwdns: reconstructioninmulti-scalefeaturespacesforlenslessimaging,

    Y.Li, Z.Li, K.Chen, et al., “Mwdns: reconstructioninmulti-scalefeaturespacesforlenslessimaging,” Opt.Express 31, 39088–39101 (2023)

    2023

  14. [22]

    Robust unrolled network for lensless imaging with enhanced resistance to model mismatch and noise,

    H. Qian, H. Ling, and X. Lu, “Robust unrolled network for lensless imaging with enhanced resistance to model mismatch and noise,” Opt. Express32, 30267–30283 (2024)

    2024

  15. [23]

    Phocolens: Photorealistic and consistent reconstruction in lensless imaging,

    X. Cai, Z. You, H. Zhang,et al., “Phocolens: Photorealistic and consistent reconstruction in lensless imaging,” arXiv preprint arXiv:2409.17996 (2024)

    2024 arXiv

  16. [24]

    Thin on-sensor nanophotonic array cameras,

    P. Chakravarthula, J. Sun, X. Li,et al., “Thin on-sensor nanophotonic array cameras,” ACM Trans. on Graph. (TOG) 42, 1–18 (2023)

    2023

  17. [25]

    Let there be light: Robust lensless imaging under external illumination with deep learning,

    E. Bezzam, S. Peters, and M. Vetterli, “Let there be light: Robust lensless imaging under external illumination with deep learning,” arXiv preprint arXiv:2409.16766 (2024)

    2024 arXiv

  18. [26]

    Robustness of single random phase encoding lensless imaging with camera noise,

    S. Goswami, G. Krishnan, and B. Javidi, “Robustness of single random phase encoding lensless imaging with camera noise,” Opt. Express32, 4916–4930 (2024)

    2024

  19. [27]

    Denoisingdiffusionprobabilisticmodels,

    J.Ho,A.Jain,andP.Abbeel,“Denoisingdiffusionprobabilisticmodels,”Adv.neuralinformationprocessingsystems 33, 6840–6851 (2020)

    2020

  20. [28]

    Low-light image enhancement with wavelet-based diffusion models,

    H. Jiang, A. Luo, H. Fan,et al., “Low-light image enhancement with wavelet-based diffusion models,” ACM Trans. on Graph. (TOG)42, 1–14 (2023)

    2023

  21. [29]

    Image quality assessment: from error visibility to structural similarity,

    Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, “Image quality assessment: from error visibility to structural similarity,” IEEE transactions on image processing13, 600–612 (2004)

    2004

  22. [30]

    The unreasonable effectiveness of deep features as a perceptual metric,

    R. Zhang, P. Isola, A. A. Efros,et al., “The unreasonable effectiveness of deep features as a perceptual metric,” in Proceedings of the IEEE conference on computer vision and pattern recognition, (2018), pp. 586–595

    2018

  23. [31]

    Sparse gradient regularized deep retinex network for robust low-light image enhancement,

    W. Yang, W. Wang, H. Huang,et al., “Sparse gradient regularized deep retinex network for robust low-light image enhancement,” IEEE Trans. on Image Process.30, 2072–2086 (2021)

    2021

  24. [32]

    Lenslesspicam: A hardware and software platform for lensless computational imaging with a raspberry pi,

    E. Bezzam, S. Kashani, M. Vetterli, and M. Simeoni, “Lenslesspicam: A hardware and software platform for lensless computational imaging with a raspberry pi,” arXiv preprint arXiv:2206.01430 (2022)

    2022 arXiv

  25. [33]

    Deeplir: Attention-based approach for mask-based lensless image reconstruction,

    A. Poudel and U. Nakarmi, “Deeplir: Attention-based approach for mask-based lensless image reconstruction,” in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, (2024), pp. 431–439

    2024

Pith tools

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