pith. sign in

arxiv: 1907.07408 · v1 · pith:3NSSICOQnew · submitted 2019-07-17 · 💻 cs.CV

Underexposed Image Correction via Hybrid Priors Navigated Deep Propagation

Risheng Liu , Long Ma , Yuxi Zhang , Xin Fan , Zhongxuan Luo This is my paper

Pith reviewed 2026-05-24 20:38 UTC · model grok-4.3

classification 💻 cs.CV
keywords underexposed image correctionhybrid priorsdeep propagationreflectance and illuminationenergy-inspired modelimage enhancementface detectionhaze removal
0
0 comments X

The pith

An energy-inspired model with hybrid priors uses deep propagation to adjust reflectance and illumination in underexposed images at once.

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

The paper formulates underexposed image correction as an energy-inspired model that draws on both physical principles and data distributions. It designs a propagation procedure navigated by these hybrid priors to move reflectance and illumination components toward better luminance and detail. Experiments compare the results against existing methods on visual quality and objective measures. The same approach is tested on face detection to check natural appearance and on haze removal to check broader use. A reader would care because underexposed photos appear often in everyday capture and improved correction could support downstream vision tasks without extra tuning.

Core claim

The central claim is that integrating knowledge from physical principles and implicit distributions from data into hybrid priors, then navigating a deep propagation procedure with them, allows simultaneous adjustment of reflectance and illumination to produce high-quality underexposed image corrections with appropriate luminance and abundant details.

What carries the argument

Hybrid priors navigated deep propagation procedure inside an energy-inspired model, which combines physical principles and data distributions to guide reflectance and illumination updates.

If this is right

  • The corrections achieve better subjective and objective quality than prior methods.
  • The results support improved performance on face detection as a downstream task.
  • The same framework applies to single-image haze removal with better outcomes than alternatives.
  • Both physical principles and data distributions are required for the propagation to succeed.

Where Pith is reading between the lines

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

  • The same hybrid-prior navigation could be tested on related low-light tasks such as denoising or contrast enhancement.
  • If the propagation step generalizes without per-image tuning, it could slot into camera pipelines for real-time use.
  • Extending the model to sequences might reveal whether temporal consistency follows from the same priors.

Load-bearing premise

A propagation procedure guided by hybrid priors can adjust reflectance and illumination together across many images without creating artifacts or needing extra tuning.

What would settle it

A collection of underexposed test images on which the method produces visible artifacts or lower scores on standard objective metrics than current leading approaches.

Figures

Figures reproduced from arXiv: 1907.07408 by Long Ma, Risheng Liu, Xin Fan, Yuxi Zhang, Zhongxuan Luo.

Figure 1
Figure 1. Figure 1: Underexposed image correction results comparison on an example image. It can be seen that there exists severe overexposure and some details [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The illustration of our propagations with hybrid priors navigation. The four dashed rectangles of left column is our core principles of designing [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Visual comparisons of our method with different prior strategies. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: Comparing the performance of two kinds of end-to-end learning [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: The visual performances and quantitative results of our method w.r.t the settings of algorithmic parameters. The NIQE (lower is better) curves of [PITH_FULL_IMAGE:figures/full_fig_p006_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Quantitative performance (i.e., NIQE, lower is better) on three different benchmark databases. [PITH_FULL_IMAGE:figures/full_fig_p006_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Comparisons on an example in Non-uniform dataset. The NIQE scores are reported below each image. [PITH_FULL_IMAGE:figures/full_fig_p007_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Comparisons on an example in LIME dataset. The NIQE scores are reported in the brackets. [PITH_FULL_IMAGE:figures/full_fig_p008_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Visual comparisons of underexposed image correction on real-world scenario. [PITH_FULL_IMAGE:figures/full_fig_p008_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Some sample images selected from DARK FACE dataset. [PITH_FULL_IMAGE:figures/full_fig_p009_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Comparison of face detection based on YOLOv3 [34] among LIME [3] (representative Retinex-based method), HDRNet [6] (end-to-end network), [PITH_FULL_IMAGE:figures/full_fig_p009_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Visual comparisons of single image haze removal on real-world scenario. [PITH_FULL_IMAGE:figures/full_fig_p010_13.png] view at source ↗
read the original abstract

Enhancing visual qualities for underexposed images is an extensively concerned task that plays important roles in various areas of multimedia and computer vision. Most existing methods often fail to generate high-quality results with appropriate luminance and abundant details. To address these issues, we in this work develop a novel framework, integrating both knowledge from physical principles and implicit distributions from data to solve the underexposed image correction task. More concretely, we propose a new perspective to formulate this task as an energy-inspired model with advanced hybrid priors. A propagation procedure navigated by the hybrid priors is well designed for simultaneously propagating the reflectance and illumination toward desired results. We conduct extensive experiments to verify the necessity of integrating both underlying principles (i.e., with knowledge) and distributions (i.e., from data) as navigated deep propagation. Plenty of experimental results of underexposed image correction demonstrate that our proposed method performs favorably against the state-of-the-art methods on both subjective and objective assessments. Additionally, we execute the task of face detection to further verify the naturalness and practical value of underexposed image correction. What's more, we employ our method to single image haze removal whose experimental results further demonstrate its superiorities.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

1 major / 0 minor

Summary. The paper claims to formulate underexposed image correction as an energy-inspired model navigated by hybrid priors that combine physical principles with data distributions. A propagation procedure is designed to simultaneously adjust reflectance and illumination components. Extensive experiments are said to verify the necessity of both knowledge-driven and data-driven components, with the method outperforming state-of-the-art approaches on subjective and objective assessments; additional validation is provided via face detection and single-image haze removal tasks.

Significance. If the hybrid-priors propagation indeed yields artifact-free results with appropriate luminance and detail recovery while outperforming existing methods, the work could advance low-light enhancement by demonstrating a principled way to fuse model-based and learning-based priors. The downstream-task experiments add evidence of practical utility.

major comments (1)
  1. Abstract: the central claim that hybrid priors are necessary rests on experimental comparisons whose details (equations, ablation results, error analysis) are absent from the provided text, so it is impossible to verify whether reported gains reduce to the hybrid formulation or to other implementation choices.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive comment. We address the major point below and clarify that the supporting details appear in the full manuscript.

read point-by-point responses

  1. Referee: [—] Abstract: the central claim that hybrid priors are necessary rests on experimental comparisons whose details (equations, ablation results, error analysis) are absent from the provided text, so it is impossible to verify whether reported gains reduce to the hybrid formulation or to other implementation choices.

    Authors: The abstract is a concise summary; the full manuscript supplies the requested details. Section 3 formulates the energy-inspired model with the hybrid priors (physical reflectance/illumination constraints combined with learned distributions) via explicit equations. Section 4 derives the propagation procedure that simultaneously updates reflectance and illumination. Section 5.3 contains the ablation studies: we report quantitative results (PSNR, SSIM, NIQE) for the full hybrid model versus ablated versions that remove either the knowledge-driven or data-driven prior, together with visual error maps and failure-case analysis. These controlled comparisons isolate the contribution of the hybrid formulation and show that performance degrades measurably when either component is omitted. The downstream face-detection and dehazing experiments further corroborate that the gains are not artifacts of other implementation choices. revision: no

Circularity Check

0 steps flagged

No significant circularity

full rationale

The paper formulates underexposed correction as an energy-inspired model with hybrid priors (physical + data) and designs a navigated propagation for reflectance/illumination. No quoted step reduces a claimed prediction or uniqueness result to a fitted parameter or self-citation by construction. Experiments are presented as external verification of necessity rather than tautological confirmation. The derivation chain is self-contained against the stated benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review supplies no explicit free parameters, axioms, or invented entities; all such elements would need to be extracted from the full manuscript.

pith-pipeline@v0.9.0 · 5747 in / 982 out tokens · 14772 ms · 2026-05-24T20:38:08.903056+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

43 extracted references · 43 canonical work pages · 1 internal anchor

  1. [1]

    Brightness preserving histogram equalization with maximum entropy: a variational perspective,

    C. Wang and Z. Ye, “Brightness preserving histogram equalization with maximum entropy: a variational perspective,” IEEE Transactions on Consumer Electronics, vol. 51, no. 4, pp. 1326–1334, 2005

    work page 2005

  2. [2]

    Brightness preserving dynamic fuzzy histogram equalization,

    D. Sheet, H. Garud, A. Suveer, M. Mahadevappa, and J. Chatterjee, “Brightness preserving dynamic fuzzy histogram equalization,” IEEE Transactions on Consumer Electronics , vol. 56, no. 4, 2010

    work page 2010

  3. [3]

    Lime: Low-light image enhancement via illumination map estimation,

    X. Guo, Y . Li, and H. Ling, “Lime: Low-light image enhancement via illumination map estimation,” IEEE Transactions on Image Processing , vol. 26, no. 2, pp. 982–993, 2017

    work page 2017

  4. [4]

    High-quality exposure correction of underexposed photos,

    Q. Zhang, W.-S. Zheng, G. Yuan, C. Xiao, and L. Zhu, “High-quality exposure correction of underexposed photos,” inACM Multimedia, 2018

    work page 2018

  5. [5]

    Structure-revealing low- light image enhancement via robust retinex model,

    M. Li, J. Liu, W. Yang, X. Sun, and Z. Guo, “Structure-revealing low- light image enhancement via robust retinex model,” IEEE Transactions on Image Processing , vol. 27, no. 6, pp. 2828–2841, 2018

    work page 2018

  6. [6]

    Deep bilateral learning for real-time image enhancement,

    M. Gharbi, J. Chen, J. T. Barron, S. W. Hasinoff, and F. Durand, “Deep bilateral learning for real-time image enhancement,” ACM Transactions on Graphics, vol. 36, no. 4, p. 118, 2017

    work page 2017

  7. [7]

    Deep retinex decomposition for low-light enhancement,

    C. Wei, W. Wang, W. Yang, and J. Liu, “Deep retinex decomposition for low-light enhancement,” in British Machine Vision Conference , 2018

    work page 2018

  8. [8]

    A simple and effective histogram equalization approach to image enhancement,

    H. Cheng and X. Shi, “A simple and effective histogram equalization approach to image enhancement,” Digital Signal Processing , vol. 14, no. 2, pp. 158–170, 2004

    work page 2004

  9. [9]

    Multi- histransactions on graphicsram equalization methods for contrast en- hancement and brightness preserving,

    M. J. Power, C. Whitlock, P. Bartlein, and L. R. Stevens, “Multi- histransactions on graphicsram equalization methods for contrast en- hancement and brightness preserving,” IEEE Transactions on Consumer Electronics, vol. 53, no. 3, pp. 1186–1194, 2011

    work page 2011

  10. [10]

    Contrast enhancement using a weighted histransactions on graphicsram equalization,

    S. H. Yun, H. K. Jin, and S. Kim, “Contrast enhancement using a weighted histransactions on graphicsram equalization,” in IEEE Inter- national Conference on Consumer Electronics , vol. 10, no. 11, 2011, pp. 203–204

    work page 2011

  11. [11]

    McCann, Retinex Theory

    J. McCann, Retinex Theory. Springer New York, 2016. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2015 11

    work page 2016

  12. [12]

    A joint intrinsic- extrinsic prior model for retinex,

    B. Cai, X. Xu, K. Guo, K. Jia, B. Hu, and D. Tao, “A joint intrinsic- extrinsic prior model for retinex,” in Proceeding of the IEEE Conference on Computer Vision and Pattern Recognition , 2017

    work page 2017

  13. [13]

    Lightness and retinex theory,

    E. H. Land and J. J. McCann, “Lightness and retinex theory,” Journal of the Optical Society of America , vol. 61, no. 1, pp. 1–11, 1971

    work page 1971

  14. [14]

    A variational framework for retinex,

    R. Kimmel, M. Elad, D. Shaked, R. Keshet, and I. Sobel, “A variational framework for retinex,” International Journal of Computer Vision , vol. 52, no. 1, pp. 7–23, 2003

    work page 2003

  15. [15]

    A total variation model for retinex,

    M. K. Ng and W. Wang, “A total variation model for retinex,” SIAM Journal on Imaging Sciences , vol. 4, no. 1, pp. 345–365, 2011

    work page 2011

  16. [16]

    A prob- abilistic method for image enhancement with simultaneous illumination and reflectance estimation,

    X. Fu, Y . Liao, D. Zeng, Y . Huang, X.-P. Zhang, and X. Ding, “A prob- abilistic method for image enhancement with simultaneous illumination and reflectance estimation,” IEEE Transactions on Image Processing , vol. 24, no. 12, pp. 4965–4977, 2015

    work page 2015

  17. [17]

    A weighted vari- ational model for simultaneous reflectance and illumination estimation,

    X. Fu, D. Zeng, Y . Huang, X.-P. Zhang, and X. Ding, “A weighted vari- ational model for simultaneous reflectance and illumination estimation,” in Proceeding of the IEEE Conference on Computer Vision and Pattern Recognition, 2016

    work page 2016

  18. [18]

    Deep image prior,

    D. Ulyanov, A. Vedaldi, and V . S. Lempitsky, “Deep image prior,” in Proceeding of the IEEE Conference on Computer Vision and Pattern Recognition, 2018

    work page 2018

  19. [19]

    Accurate image super-resolution using very deep convolutional networks,

    J. Kim, J. K. Lee, and K. M. Lee, “Accurate image super-resolution using very deep convolutional networks,” in Proceeding of the IEEE Conference on Computer Vision and Pattern Recognition , 2016, pp. 1646–1654

    work page 2016

  20. [20]

    Residual dense network for image super-resolution,

    Y . Zhang, Y . Tian, Y . Kong, B. Zhong, and Y . Fu, “Residual dense network for image super-resolution,” in Proceedings of the IEEE Con- ference on Computer Vision and Pattern Recognition , 2018, pp. 2472– 2481

    work page 2018

  21. [21]

    Proximal alternating direction network: A globally converged deep unrolling framework,

    R. Liu, X. Fan, S. Cheng, X. Wang, and Z. Luo, “Proximal alternating direction network: A globally converged deep unrolling framework,” in Association for the Advancement of Artificial Intelligence , 2018

    work page 2018

  22. [22]

    Learning collaborative generation correction modules for blind image deblurring and beyond,

    R. Liu, Y . He, S. Cheng, X. Fan, and Z. Luo, “Learning collaborative generation correction modules for blind image deblurring and beyond,” in ACM Multimedia, 2018

    work page 2018

  23. [23]

    Aod-net: All-in-one dehazing network,

    B. Li, X. Peng, Z. Wang, J. Xu, and D. Feng, “Aod-net: All-in-one dehazing network,” in Proceedings of the IEEE International Conference on Computer Vision , 2017

    work page 2017

  24. [24]

    Learning converged prop- agations with deep prior ensemble for image enhancement,

    R. Liu, L. Ma, Y . Wang, and L. Zhang, “Learning converged prop- agations with deep prior ensemble for image enhancement,” IEEE Transactions on Image Processing, vol. 28, no. 3, pp. 1528–1543, 2018

    work page 2018

  25. [25]

    Deep proximal unrolling: Algorithmic framework, convergence analysis and applications,

    R. Liu, S. Cheng, L. Ma, X. Fan, and Z. Luo, “Deep proximal unrolling: Algorithmic framework, convergence analysis and applications,” IEEE Transactions on Image Processing , 2019

    work page 2019

  26. [26]

    On the convergence of learning-based iterative methods for nonconvex inverse problems,

    R. Liu, S. Cheng, Y . He, X. Fan, Z. Lin, and Z. Luo, “On the convergence of learning-based iterative methods for nonconvex inverse problems,” IEEE transactions on pattern analysis and machine intelligence , 2019

    work page 2019

  27. [27]

    Llnet: A deep autoencoder approach to natural low-light image enhancement,

    K. G. Lore, A. Akintayo, and S. Sarkar, “Llnet: A deep autoencoder approach to natural low-light image enhancement,” Pattern Recognition, vol. 61, pp. 650–662, 2017

    work page 2017

  28. [28]

    Learning a deep single image contrast enhancer from multi-exposure images,

    J. Cai, S. Gu, and L. Zhang, “Learning a deep single image contrast enhancer from multi-exposure images,” IEEE Transactions on Image Processing, vol. 27, no. 4, pp. 2049–2062, 2018

    work page 2049

  29. [29]

    Kindling the darkness: A practical low-light image enhancer,

    Y . Zhang, J. Zhang, and X. Guo, “Kindling the darkness: A practical low-light image enhancer,” in ACM Multimedia, 2019

    work page 2019

  30. [30]

    Fields of experts,

    S. Roth and M. J. Black, “Fields of experts,” International Journal of Computer Vision, vol. 82, no. 2, pp. 205–229, 2009

    work page 2009

  31. [31]

    Poynton, Digital video and HD: Algorithms and Interfaces , 2012

    C. Poynton, Digital video and HD: Algorithms and Interfaces , 2012

    work page 2012

  32. [32]

    Imagenet classification with deep convolutional neural networks,

    A. Krizhevsky, I. Sutskever, and G. E. Hinton, “Imagenet classification with deep convolutional neural networks,” in Neural Information Pro- cessing Systems, 2012

    work page 2012

  33. [33]

    Making a completely blind image quality analyzer,

    A. Mittal, R. Soundararajan, and A. C. Bovik, “Making a completely blind image quality analyzer,” IEEE Signal Processing Letters , vol. 20, no. 3, pp. 209–212, 2013

    work page 2013

  34. [34]

    YOLOv3: An Incremental Improvement

    J. Redmon and A. Farhadi, “Yolov3: An incremental improvement,” arXiv preprint arXiv:1804.02767 , 2018

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  35. [35]

    Retinex processing for automatic image enhancement,

    Z.-u. Rahman, D. J. Jobson, and G. A. Woodell, “Retinex processing for automatic image enhancement,” Journal of Electronic Imaging , vol. 13, no. 1, pp. 100–111, 2004

    work page 2004

  36. [36]

    Globally optimized linear windowed tone mapping,

    Q. Shan, J. Jia, and M. S. Brown, “Globally optimized linear windowed tone mapping,” IEEE Transactions on Visualization and Computer Graphics, vol. 16, no. 4, pp. 663–675, 2010

    work page 2010

  37. [37]

    Naturalness preserved enhancement algorithm for non-uniform illumination images,

    S. Wang, J. Zheng, H.-M. Hu, and B. Li, “Naturalness preserved enhancement algorithm for non-uniform illumination images,” IEEE Transactions on Image Processing, vol. 22, no. 9, pp. 3538–3548, 2013

    work page 2013

  38. [38]

    Burst photography for high dynamic range and low-light imaging on mobile cameras,

    S. W. Hasinoff, D. Sharlet, R. Geiss, A. Adams, J. T. Barron, F. Kainz, J. Chen, and M. Levoy, “Burst photography for high dynamic range and low-light imaging on mobile cameras,” ACM Transactions on Graphics, vol. 35, no. 6, 2016

    work page 2016

  39. [39]

    Wider face: A face detection benchmark,

    S. Yang, P. Luo, C. C. Loy, and X. Tang, “Wider face: A face detection benchmark,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016

    work page 2016

  40. [40]

    On the duality between retinex and image dehazing,

    A. Galdran, A. Alvarez-Gila, A. Bria, J. Vazquez-Corral, and M. Bertalmıo, “On the duality between retinex and image dehazing,” in Proceeding of the IEEE Conference on Computer Vision and Pattern Recognition, 2018

    work page 2018

  41. [41]

    Single image haze removal using dark channel prior,

    K. He, J. Sun, and X. Tang, “Single image haze removal using dark channel prior,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 33, no. 12, pp. 2341–2353, 2011

    work page 2011

  42. [42]

    Non-local image dehazing,

    D. Berman, S. Avidan et al., “Non-local image dehazing,” in Proceeding of the IEEE Conference on Computer Vision and Pattern Recognition , 2016, pp. 1674–1682. Risheng Liu (M’12-) received the BSc and PhD degrees both in mathematics from the Dalian Univer- sity of Technology in 2007 and 2012, respectively. He was a visiting scholar in the Robotic Institut...

    work page 2016

  43. [43]

    His research interests include computational geometry and computer vision

    work page