pith. machine review for the scientific record. sign in

arxiv: 2604.10273 · v1 · submitted 2026-04-11 · 💻 cs.CV

Recognition: unknown

Dual-Exposure Imaging with Events

Mingyuan Lin , Hongyi Liu , Chu He , Wen Yang , Gui-Song Xia , Lei Yu
Authors on Pith no claims yet

Pith reviewed 2026-05-10 16:04 UTC · model grok-4.3

classification 💻 cs.CV
keywords dual-exposure imagingevent camerasmotion deblurringlow-light enhancementfeature alignmentimage reconstructionevent-based vision
0
0 comments X

The pith

Event data from cameras aligns short- and long-exposure images to remove motion artifacts in low-light reconstruction.

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

The paper claims that dual-exposure imaging, which pairs short and long exposures to balance detail and brightness in dark scenes, can be made reliable by adding event camera data. Existing methods produce blur and mismatches because scene motion displaces features differently across exposure times. Events supply microsecond-scale motion signals that let a network align the two frames and guide both deblurring and brightness recovery. The authors split the problem into two parallel paths, one for each exposure, and introduce a fusion module that uses events to match and combine features before final synthesis. Experiments on real paired data show cleaner results than prior image-only approaches.

Core claim

The E-DEI network reconstructs high-quality low-light images from dual-exposure pairs plus events by decomposing the task into event-guided motion deblurring and enhancement, using a dual-path architecture whose Dual-path Feature Alignment and Fusion module aligns and merges features across exposures with the help of high-temporal-resolution event information.

What carries the argument

The Dual-path Feature Alignment and Fusion (DFAF) module, which takes event streams as auxiliary input to align and fuse features extracted from the short-exposure and long-exposure images.

If this is right

  • Dual-exposure pairs become usable in dynamic low-light environments where motion previously made them unreliable.
  • Motion deblurring and low-light enhancement can be solved jointly rather than sequentially when event data is available.
  • A new real-world dataset of paired low-light and normal-light images with events supports training and benchmarking of similar methods.
  • The dual-path design with event-assisted fusion generalizes across multiple test sets, indicating the alignment step is the main source of improvement.

Where Pith is reading between the lines

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

  • The same event-assisted alignment principle could be applied to other multi-frame fusion tasks such as HDR merging or burst denoising.
  • In robotics or surveillance, adding event sensors to existing dual-exposure cameras might yield reliable vision without extra hardware beyond the event camera itself.
  • If event density is low in very dark regions, performance may degrade unless the network learns to fall back to image-only cues.

Load-bearing premise

Event streams supply sufficiently complete and accurate motion information between and within the two exposure frames to correct spatial displacements and exposure mismatches without creating new artifacts or needing scene-specific tuning.

What would settle it

Capture dual-exposure pairs plus events in a scene with rapid object motion, then check whether the output still contains visible ghosting or residual blur when compared pixel-by-pixel against a static reference image taken at the same average light level.

Figures

Figures reproduced from arXiv: 2604.10273 by Chu He, Gui-Song Xia, Hongyi Liu, Lei Yu, Mingyuan Lin, Wen Yang.

Figure 1
Figure 1. Figure 1: (a) Illustration of the proposed Event-based Dual-Exposure Imag [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Comparison between (a) existing Frame-based Dual-Exposure Imaging [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: (a) The overall architecture of the proposed network of Event-based Dual-Exposure Imaging (E-DEI), which is a dual-path feature propagation and supervision framework. (b) The proposed Dual-path Feature Alignment and Fusion (DFAF) module, where features of the Deblurring path are spatially aligned to the short-exposure target time with features of the Enhancement path as the temporal reference, and then we … view at source ↗
Figure 4
Figure 4. Figure 4: Details of the proposed Cross-Guided Fusion (CGF) module, which is designed to fuse the results and features of the Deblurring and Enhancement paths for the final result. Deblurring path receives the long-exposure image Il and events E[ts,te] as inputs and outputs the deblurred image Lˆ l , while the Enhancement path is fed with the short-exposure image Is and surrounding events E[ts−∆t,ts+∆t] to produce t… view at source ↗
Figure 5
Figure 5. Figure 5: (a) The illustration of a beam splitter-based hybrid camera system for building the PIED dataset. (b) PIED contains the normal- and low-light image [PITH_FULL_IMAGE:figures/full_fig_p006_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Statistical properties of the E-DEI dataset. Distributions of (a) Motion magnitude, (b) Illumination, (c) Texture richness, and (d) Event firing rate. higher frame rate (72 FPS) than the abovementioned datasets, as shown in Table I. It can be used to simultaneously eval￾uate multiple frame-/event-based low-light imaging methods, e.g., long-exposure motion deblurring, short-exposure image enhancement, and d… view at source ↗
Figure 7
Figure 7. Figure 7: Qualitative comparisons with state-of-the-art methods on the REDS-LL (upper low) and PIED (lower row) datasets. For a better view, details are [PITH_FULL_IMAGE:figures/full_fig_p008_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Qualitative ablations of the backbone and modules of our network and their absolute differences to the ground-truth clear image. Enc. and Dec. indicate the architecture of the encoder and decoder. performance, which demonstrates the complementary benefits of event data for both contributions. 2) Effectiveness of Parallel Interactive Structure: As desribed in Section IV-A, an intuitive approach to solve the… view at source ↗
Figure 9
Figure 9. Figure 9: Visualization of intermediate features from the DFAF module of our E-DEI. (a) and (d) the input dual-exposure image pairs. (b-c) Fl and F ′ l are the features before and after the DA block in the Delurring path. (e-f) Fs and F ′ s are the features before and after the CAF block in the Enhancement path. path parallel architecture, which is composed of two parallel encoder-decoder structures, i.e., the Deblu… view at source ↗
Figure 11
Figure 11. Figure 11: (a) Sensitivity analysis of the E-DEI network to the temporal misalignment between cross-modal image pairs and event streams. (b) Gen￾eralization of the E-DEI network to the unseen exposure-time ratios. TABLE VI QUANTITATIVE COMPARISONS OF THE NETWORK PARAMETER AND INFERENCE EFFICIENCY ON THE REDS-LL DATASETS. THE ITEMS OF RUNTIME AND FLOPS ARE EVALUATED WITH THE INPUT IMAGES AND EVENTS AT THE RESOLUTION … view at source ↗
Figure 12
Figure 12. Figure 12: Failure cases under non-ideal scenes, e.g., extremely dim environment (upper), blocked and discarded events (middle), and coupled degradations (lower). better than the state-of-the-art F-DEI method, i.e., D2HNet [7], with much smaller model size (9.46 M vs. 79.58 M) and less inference time (33.0 ms vs. 83.7 ms). Our base model (w/o CGF) requires only 8.58M parameters and 1344.76 GFLOPs, operating at 28.6 … view at source ↗
read the original abstract

By combining complementary benefits of short- and long-exposure images, Dual-Exposure Imaging (DEI) enhances image quality in low-light scenarios. However, existing DEI approaches inevitably suffer from producing artifacts due to spatial displacement from scene motion and image feature discrepancies from different exposure times. To tackle this problem, we propose a novel Event-based DEI (E-DEI) algorithm, which reconstructs high-quality images from dual-exposure image pairs and events, leveraging high temporal resolution of event cameras to provide accurate inter-/intra-frame dynamic information. Specifically, we decompose this complex task into an integration of two sub-tasks, i.e., event-based motion deblurring and low-light image enhancement tasks, which guides us to design E-DEI network as a dual-path parallel feature propagation architecture. We propose a Dual-path Feature Alignment and Fusion (DFAF) module to effectively align and fuse features extracted from dual-exposure images with assistance of events. Furthermore, we build a real-world Dataset containing Paired low-/normal-light Images and Events (PIED). Experiments on multiple datasets show the superiority of our method. The code and dataset are available at github.

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

2 major / 2 minor

Summary. The paper proposes an Event-based Dual-Exposure Imaging (E-DEI) algorithm that reconstructs high-quality low-light images from paired short- and long-exposure images plus event streams. It decomposes the problem into event-based motion deblurring and low-light enhancement, implemented via a dual-path parallel feature propagation network with a Dual-path Feature Alignment and Fusion (DFAF) module that uses events for inter- and intra-frame alignment. The authors also introduce the PIED real-world dataset of paired low-/normal-light images and events, and claim superior performance on multiple datasets.

Significance. If the experimental claims hold, the work would be significant for low-light imaging applications by showing how event cameras' high temporal resolution can mitigate motion artifacts and exposure discrepancies without scene-specific tuning. The release of the PIED dataset would also be a concrete contribution for future research in event-based vision.

major comments (2)
  1. [Experiments] Experiments section: the abstract and results claim superiority on multiple datasets and introduce the PIED dataset, yet the provided description contains no quantitative metrics (PSNR, SSIM, etc.), ablation studies on the DFAF module or event contribution, or error analysis; this leaves the central claim of effective alignment without new artifacts unverified.
  2. [Methods (DFAF)] Methods, DFAF module description: the design assumes events supply sufficiently complete and accurate inter-/intra-frame dynamic information for feature alignment, but no experiments or analysis test robustness when event density is low (as occurs in low-light, low-contrast, or slowly varying regions) or when events are corrupted by typical sensor noise; this directly bears on whether the dual-path decomposition avoids introducing new artifacts.
minor comments (2)
  1. [Introduction] The notation for 'inter-/intra-frame' dynamic information is used repeatedly but never formally defined or illustrated with an example of what intra-frame cues events are expected to provide.
  2. [Abstract] The paper states that code and dataset are available at github but provides no specific link or repository identifier in the text.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback. We will revise the manuscript to provide more detailed experimental validation and robustness analysis as requested.

read point-by-point responses

  1. Referee: [Experiments] Experiments section: the abstract and results claim superiority on multiple datasets and introduce the PIED dataset, yet the provided description contains no quantitative metrics (PSNR, SSIM, etc.), ablation studies on the DFAF module or event contribution, or error analysis; this leaves the central claim of effective alignment without new artifacts unverified.

    Authors: We acknowledge that the experiments section in the submitted version may not have presented the quantitative results with sufficient detail. The paper does report comparisons on multiple datasets using standard metrics such as PSNR and SSIM, and includes the PIED dataset. However, to fully address this concern, we will expand the section in the revised manuscript to include explicit quantitative tables, comprehensive ablation studies on the DFAF module and the contribution of event data, and an error analysis demonstrating that the alignment does not introduce new artifacts. revision: yes

  2. Referee: [Methods (DFAF)] Methods, DFAF module description: the design assumes events supply sufficiently complete and accurate inter-/intra-frame dynamic information for feature alignment, but no experiments or analysis test robustness when event density is low (as occurs in low-light, low-contrast, or slowly varying regions) or when events are corrupted by typical sensor noise; this directly bears on whether the dual-path decomposition avoids introducing new artifacts.

    Authors: We agree with the referee that validating the assumption regarding event data completeness is important. Our current experiments on the real-world PIED dataset include challenging low-light conditions where event density can vary. Nevertheless, we will add specific robustness tests in the revised version, including controlled experiments with reduced event density and added noise, to analyze the performance of the DFAF module and confirm that the dual-path approach mitigates artifacts effectively even under these conditions. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical network design with external validation

full rationale

The manuscript proposes an empirical dual-path neural network (E-DEI with DFAF module) for event-assisted dual-exposure imaging and introduces a new paired dataset (PIED). No equations, parameter-fitting steps, or first-principles derivations appear; the architecture is presented as a design choice justified by task decomposition and empirical superiority on multiple datasets. No self-citations are invoked as load-bearing uniqueness theorems or ansatzes, and no predictions reduce to fitted inputs by construction. The central claims rest on external experimental benchmarks rather than any closed self-referential loop.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review; no equations, parameters, or theoretical constructs are supplied.

pith-pipeline@v0.9.0 · 5501 in / 1139 out tokens · 87342 ms · 2026-05-10T16:04:50.001898+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

71 extracted references · 2 canonical work pages

  1. [1]

    Low-light image and video enhancement using deep learning: A survey,

    C. Li, C. Guo, L. Han, J. Jiang, M.-M. Cheng, J. Gu, and C. C. Loy, “Low-light image and video enhancement using deep learning: A survey,”IEEE TPAMI, vol. 44, no. 12, pp. 9396–9416, 2021. 1

    2021

  2. [2]

    Seeing dynamic scene in the dark: A high-quality video dataset with mechatronic alignment,

    R. Wang, X. Xu, C.-W. Fu, J. Lu, B. Yu, and J. Jia, “Seeing dynamic scene in the dark: A high-quality video dataset with mechatronic alignment,” inICCV, 2021, pp. 9700–9709. 1

    2021

  3. [3]

    Exposure bracketing is all you need for a high-quality image,

    Z. Zhang, S. Zhang, R. Wu, Z. Yan, and W. Zuo, “Exposure bracketing is all you need for a high-quality image,” inICLR, 2025. 1

    2025

  4. [4]

    Image deblurring with blurred/noisy image pairs,

    L. Yuan, J. Sun, L. Quan, and H.-Y . Shum, “Image deblurring with blurred/noisy image pairs,” inACM SIGGRAPH, 2007, pp. 1–10. 1, 2

    2007

  5. [5]

    Lsd2– joint denoising and deblurring of short and long exposure images with cnns,

    J. Mustaniemi, J. Kannala, J. Matas, S. S ¨arkk¨a, and J. Heikkil ¨a, “Lsd2– joint denoising and deblurring of short and long exposure images with cnns,” inBMVC, 2020, pp. 1–13. 1, 2, 7, 8, 9, 12

    2020

  6. [6]

    Low-light image restoration with short-and long-exposure raw pairs,

    M. Chang, H. Feng, Z. Xu, and Q. Li, “Low-light image restoration with short-and long-exposure raw pairs,”IEEE TMM, vol. 24, pp. 702–714,

  7. [7]

    1, 2, 3, 7, 8, 9, 12

  8. [8]

    D2hnet: Joint denoising and deblurring with hierarchical network for robust night image restoration,

    Y . Zhao, Y . Xu, Q. Yan, D. Yang, X. Wang, and L.-M. Po, “D2hnet: Joint denoising and deblurring with hierarchical network for robust night image restoration,” inECCV, 2022, pp. 91–110. 1, 2, 3, 7, 8, 9, 12

    2022

  9. [9]

    Proactive camera attribute control using bayesian optimization for illumination-resilient visual navigation,

    J. Kim, Y . Cho, and A. Kim, “Proactive camera attribute control using bayesian optimization for illumination-resilient visual navigation,”IEEE TRO, vol. 36, no. 4, pp. 1256–1271, 2020. 1

    2020

  10. [10]

    Learned camera gain and exposure control for improved visual feature detection and matching,

    J. Tomasi, B. Wagstaff, S. L. Waslander, and J. Kelly, “Learned camera gain and exposure control for improved visual feature detection and matching,”IEEE RAL, vol. 6, no. 2, pp. 2028–2035, 2021. 1

    2028

  11. [11]

    Exploring and evaluating image restoration potential in dynamic scenes,

    C. Zhang, S. Su, Y . Zhu, Q. Yan, J. Sun, and Y . Zhang, “Exploring and evaluating image restoration potential in dynamic scenes,” inCVPR, 2022, pp. 2067–2076. 1

    2022

  12. [12]

    Camera attributes control for visual odometry with motion blur awareness,

    B. Han, Y . Lin, Y . Dong, H. Wang, T. Zhang, and C. Liang, “Camera attributes control for visual odometry with motion blur awareness,” IEEE/ASME Transactions on Mechatronics, 2023. 1

    2023

  13. [13]

    Drl-isp: Multi-objective camera isp with deep reinforcement learning,

    U. Shin, K. Lee, and I. S. Kweon, “Drl-isp: Multi-objective camera isp with deep reinforcement learning,” inIROS, 2022, pp. 7044–7051. 1

    2022

  14. [14]

    Spatial temporal video enhancement using alternating exposures,

    W. Shen, M. Cheng, G. Lu, G. Zhai, L. Chen, M. S. Asif, and Z. Gao, “Spatial temporal video enhancement using alternating exposures,”IEEE TCSVT, vol. 32, no. 8, pp. 4912–4926, 2021. 1, 2

    2021

  15. [15]

    Ntire 2024 challenge on bracketing image restoration and enhancement: Datasets methods and results,

    Z. Zhang, S. Zhang, R. Wu, W. Zuo, R. Timofte, X. Xing, H. Park, S. Song, C. Kim, X. Konget al., “Ntire 2024 challenge on bracketing image restoration and enhancement: Datasets methods and results,” in CVPRW, 2024, pp. 6153–6166. 1

    2024

  16. [16]

    A 128×128 120 dB 15µs latency asynchronous temporal contrast vision sensor,

    P. Lichtsteiner, C. Posch, and T. Delbruck, “A 128×128 120 dB 15µs latency asynchronous temporal contrast vision sensor,”IEEE J. Solid- State Circuits, vol. 43, no. 2, pp. 566–576, 2008. 1, 3

    2008

  17. [17]

    Event- based vision: A survey,

    G. Gallego, T. Delbr ¨uck, G. Orchard, C. Bartolozzi, B. Taba, A. Censi, S. Leutenegger, A. J. Davison, J. Conradt, K. Daniilidiset al., “Event- based vision: A survey,”IEEE TPAMI, vol. 44, no. 1, pp. 154–180, 2020. 1, 2, 3

    2020

  18. [18]

    Asynchronous blob tracker for event cameras,

    Z. Wang, T. Molloy, P. van Goor, and R. Mahony, “Asynchronous blob tracker for event cameras,”IEEE TRO, vol. 40, pp. 4750–4767, 2024. 1, 3

    2024

  19. [19]

    Coherent event guided low-light video enhancement,

    J. Liang, Y . Yang, B. Li, P. Duan, Y . Xu, and B. Shi, “Coherent event guided low-light video enhancement,” inICCV, 2023, pp. 10 615– 10 625. 1, 8, 9, 12

    2023

  20. [20]

    Seeing motion at nighttime with an event camera,

    H. Liu, S. Peng, L. Zhu, Y . Chang, H. Zhou, and L. Yan, “Seeing motion at nighttime with an event camera,” inCVPR, 2024, pp. 25 648–25 658. 2, 6

    2024

  21. [21]

    From sim-to-real: Toward general event-based low-light frame interpolation with per-scene optimization,

    Z. Zhang, Y . Ma, Y . Chen, F. Zhang, J. Gu, T. Xue, and S. Guo, “From sim-to-real: Toward general event-based low-light frame interpolation with per-scene optimization,” inACM SIGGRAPH Asia, 2024, pp. 1–

    2024

  22. [22]

    Towards robust event- guided low-light image enhancement: a large-scale real-world event- image dataset and novel approach,

    G. Liang, K. Chen, H. Li, Y . Lu, and L. Wang, “Towards robust event- guided low-light image enhancement: a large-scale real-world event- image dataset and novel approach,” inCVPR, 2024, pp. 23–33. 2, 4, 6, 7, 8, 9, 12

    2024

  23. [23]

    Deep retinex decomposition for low-light enhancement,

    C. Wei, W. Wang, W. Yang, and J. Liu, “Deep retinex decomposition for low-light enhancement,” inBMVC, 2018, pp. 1–12. 2

    2018

  24. [24]

    Uretinex- net: Retinex-based deep unfolding network for low-light image enhance- ment,

    W. Wu, J. Weng, P. Zhang, X. Wang, W. Yang, and J. Jiang, “Uretinex- net: Retinex-based deep unfolding network for low-light image enhance- ment,” inCVPR, 2022, pp. 5901–5910. 2, 8, 9, 12

    2022

  25. [25]

    Learning a simple low-light image enhancer from paired low-light instances,

    Z. Fu, Y . Yang, X. Tu, Y . Huang, X. Ding, and K.-K. Ma, “Learning a simple low-light image enhancer from paired low-light instances,” in CVPR, 2023, pp. 22 252–22 261. 2

    2023

  26. [26]

    Snr-aware low-light image enhancement,

    X. Xu, R. Wang, C.-W. Fu, and J. Jia, “Snr-aware low-light image enhancement,” inCVPR, 2022, pp. 17 714–17 724. 2

    2022

  27. [27]

    Low-light image enhancement via structure modeling and guidance,

    X. Xu, R. Wang, and J. Lu, “Low-light image enhancement via structure modeling and guidance,” inCVPR, 2023, pp. 9893–9903. 2, 4

    2023

  28. [28]

    Exposure trajectory recovery from motion blur,

    Y . Zhang, C. Wang, S. J. Maybank, and D. Tao, “Exposure trajectory recovery from motion blur,”IEEE TPAMI, vol. 44, no. 11, pp. 7490– 7504, 2021. 2, 3

    2021

  29. [29]

    Self-supervised non-uniform kernel estimation with flow-based motion prior for blind image deblurring,

    Z. Fang, F. Wu, W. Dong, X. Li, J. Wu, and G. Shi, “Self-supervised non-uniform kernel estimation with flow-based motion prior for blind image deblurring,” inCVPR, 2023, pp. 18 105–18 114. 2

    2023

  30. [30]

    Rethinking coarse-to-fine approach in single image deblurring,

    S.-J. Cho, S.-W. Ji, J.-P. Hong, S.-W. Jung, and S.-J. Ko, “Rethinking coarse-to-fine approach in single image deblurring,” inICCV, 2021, pp. 4641–4650. 2

    2021

  31. [31]

    Multi-scale frequency separation network for image deblurring,

    Y . Zhang, Q. Li, M. Qi, D. Liu, J. Kong, and J. Wang, “Multi-scale frequency separation network for image deblurring,”IEEE TCSVT, vol. 33, no. 10, pp. 5525–5537, 2023. 2

    2023

  32. [32]

    Multi-stage progressive image restoration,

    S. W. Zamir, A. Arora, S. Khan, M. Hayat, F. S. Khan, M.-H. Yang, and L. Shao, “Multi-stage progressive image restoration,” inCVPR, 2021, pp. 14 821–14 831. 2, 6

    2021

  33. [33]

    Hinet: Half instance normalization network for image restoration,

    L. Chen, X. Lu, J. Zhang, X. Chu, and C. Chen, “Hinet: Half instance normalization network for image restoration,” inCVPRW, 2021, pp. 182–192. 2

    2021

  34. [34]

    Deblurgan: Blind motion deblurring using conditional adversarial net- works,

    O. Kupyn, V . Budzan, M. Mykhailych, D. Mishkin, and J. Matas, “Deblurgan: Blind motion deblurring using conditional adversarial net- works,” inCVPR, 2018, pp. 8183–8192. 2

    2018

  35. [35]

    Multiscale structure guided diffusion for image deblurring,

    M. Ren, M. Delbracio, H. Talebi, G. Gerig, and P. Milanfar, “Multiscale structure guided diffusion for image deblurring,” inICCV, 2023, pp. 10 721–10 733. 2

    2023

  36. [36]

    Blur removal via blurred-noisy image pair,

    C. Gu, X. Lu, Y . He, and C. Zhang, “Blur removal via blurred-noisy image pair,”IEEE TIP, vol. 30, pp. 345–359, 2020. 2, 3

    2020

  37. [37]

    Self-supervised image restoration with blurry and noisy pairs,

    Z. Zhang, R. Xu, M. Liu, Z. Yan, and W. Zuo, “Self-supervised image restoration with blurry and noisy pairs,”NeurIPS, vol. 35, pp. 29 179– 29 191, 2022. 2

    2022

  38. [38]

    Dual-camera joint deblurring-denoising,

    S. Shekarforoush, A. Walia, M. A. Brubaker, K. G. Derpanis, and A. Levinshtein, “Dual-camera joint deblurring-denoising,”arXiv preprint arXiv:2309.08826, 2023. 2

    work page arXiv 2023

  39. [39]

    Deep hybrid camera deblurring for smartphone cameras,

    J. Rim, J. Lee, H. Yang, and S. Cho, “Deep hybrid camera deblurring for smartphone cameras,” inACM SIGGRAPH, 2024, pp. 1–11. 2

    2024

  40. [40]

    Deformable convnets v2: More deformable, better results,

    X. Zhu, H. Hu, S. Lin, and J. Dai, “Deformable convnets v2: More deformable, better results,” inCVPR, 2019, pp. 9308–9316. 2, 5

    2019

  41. [41]

    Event-based fusion for motion deblurring with cross- modal attention,

    L. Sun, C. Sakaridis, J. Liang, Q. Jiang, K. Yang, P. Sun, Y . Ye, K. Wang, and L. V . Gool, “Event-based fusion for motion deblurring with cross- modal attention,” inECCV, 2022, pp. 412–428. 3, 8, 9, 12

    2022

  42. [42]

    Unifying motion deblurring and frame interpola- tion with events,

    X. Zhang and L. Yu, “Unifying motion deblurring and frame interpola- tion with events,” inCVPR, 2022, pp. 17 765–17 774. 3, 4, 5

    2022

  43. [43]

    Generalizing event- based motion deblurring in real-world scenarios,

    X. Zhang, L. Yu, W. Yang, J. Liu, and G.-S. Xia, “Generalizing event- based motion deblurring in real-world scenarios,” inCVPR, 2023, pp. 10 734–10 744. 3

    2023

  44. [44]

    Motion deblurring via spatial- temporal collaboration of frames and events,

    W. Yang, J. Wu, J. Ma, L. Li, and G. Shi, “Motion deblurring via spatial- temporal collaboration of frames and events,” inAAAI, vol. 38, no. 7, 2024, pp. 6531–6539. 3, 4, 5, 8, 9, 12

    2024

  45. [45]

    Asymmetric event- guided video super-resolution,

    Z. Xiao, D. Kai, Y . Zhang, X. Sun, and Z. Xiong, “Asymmetric event- guided video super-resolution,” inACM MM, 2024, pp. 2409–2418. 3

    2024

  46. [46]

    Event- adapted video super-resolution,

    Z. Xiao, D. Kai, Y . Zhang, Z.-J. Zha, X. Sun, and Z. Xiong, “Event- adapted video super-resolution,” inECCV. Springer, 2024, pp. 217–235. 3, 5

    2024

  47. [47]

    Event-based video super-resolution via state space models,

    Z. Xiao and X. Wang, “Event-based video super-resolution via state space models,” inCVPR. IEEE, 2025, pp. 12 564–12 574. 3, 5

    2025

  48. [48]

    Time lens: Event-based video frame interpolation,

    S. Tulyakov, D. Gehrig, S. Georgoulis, J. Erbach, M. Gehrig, Y . Li, and D. Scaramuzza, “Time lens: Event-based video frame interpolation,” in CVPR, 2021, pp. 16 155–16 164. 3, 5 SUBMISSION TO IEEE TMM 14

    2021

  49. [49]

    Timereplayer: Unlocking the potential of event cameras for video interpolation,

    W. He, K. You, Z. Qiao, X. Jia, Z. Zhang, W. Wang, H. Lu, Y . Wang, and J. Liao, “Timereplayer: Unlocking the potential of event cameras for video interpolation,” inCVPR, 2022, pp. 17 804–17 813. 3

    2022

  50. [50]

    Video frame interpolation with stereo event and intensity cameras,

    C. Ding, M. Lin, H. Zhang, J. Liu, and L. Yu, “Video frame interpolation with stereo event and intensity cameras,”IEEE TMM, vol. 26, pp. 9187– 9202, 2024. 3

    2024

  51. [51]

    Evunroll: Neuromorphic events based rolling shutter image correction,

    X. Zhou, P. Duan, Y . Ma, and B. Shi, “Evunroll: Neuromorphic events based rolling shutter image correction,” inCVPR, 2022, pp. 17 775– 17 784. 3

    2022

  52. [52]

    Evshutter: Transforming events for unconstrained rolling shutter correction,

    J. Erbach, S. Tulyakov, P. Vitoria, A. Bochicchio, and Y . Li, “Evshutter: Transforming events for unconstrained rolling shutter correction,” in CVPR, 2023, pp. 13 904–13 913. 3

    2023

  53. [53]

    Self-supervised shutter unrolling with events,

    M. Lin, Y . Wang, X. Zhang, B. Shi, W. Yang, C. He, G.-s. Xia, and L. Yu, “Self-supervised shutter unrolling with events,”IJCV, pp. 1–19,

  54. [54]

    Hdr reconstruction from bracketed exposures and events,

    R. Shaw, S. Catley-Chandar, A. Leonardis, and E. P ´erez-Pellitero, “Hdr reconstruction from bracketed exposures and events,” inBMVC, 2022, pp. 1–13. 3, 5

    2022

  55. [55]

    Multi-bracket high dynamic range imaging with event cameras,

    N. Messikommer, S. Georgoulis, D. Gehrig, S. Tulyakov, J. Erbach, A. Bochicchio, Y . Li, and D. Scaramuzza, “Multi-bracket high dynamic range imaging with event cameras,” inCVPRW, 2022, pp. 547–557. 3, 5

    2022

  56. [56]

    High-speed hdr video recon- struction from hybrid intensity frames and events,

    R. Samra, K. Mitra, and P. Shedligeri, “High-speed hdr video recon- struction from hybrid intensity frames and events,” inCVMI. Springer, 2023, pp. 179–190. 3, 5

    2023

  57. [57]

    Event- assisted 12-stop hdr imaging of dynamic scene,

    S. Guo, Z. Chen, Z. Zhang, Y . Chen, G. Xu, and T. Xue, “Event- assisted 12-stop hdr imaging of dynamic scene,”arXiv preprint arXiv:2412.14705, 2024. 3, 5

    work page arXiv 2024

  58. [58]

    Reblur2deblur: Deblurring videos via self-supervised learning,

    H. Chen, J. Gu, O. Gallo, M.-Y . Liu, A. Veeraraghavan, and J. Kautz, “Reblur2deblur: Deblurring videos via self-supervised learning,” in ICCP. IEEE, 2018, pp. 1–9. 3

    2018

  59. [59]

    Mssnet: Multi-scale-stage network for single image deblurring,

    K. Kim, S. Lee, and S. Cho, “Mssnet: Multi-scale-stage network for single image deblurring,” inECCV, 2022, pp. 524–539. 3, 8, 9, 12

    2022

  60. [60]

    Radiometric ccd camera calibration and noise estimation,

    G. E. Healey and R. Kondepudy, “Radiometric ccd camera calibration and noise estimation,”IEEE TPAMI, vol. 16, no. 3, pp. 267–276, 1994. 3

    1994

  61. [61]

    Image denoising in the deep learning era,

    S. Izadi, D. Sutton, and G. Hamarneh, “Image denoising in the deep learning era,”Artificial Intelligence Review, vol. 56, no. 7, pp. 5929– 5974, 2023. 3

    2023

  62. [62]

    Attention guided low-light image enhancement with a large scale low-light simulation dataset,

    F. Lv, Y . Li, and F. Lu, “Attention guided low-light image enhancement with a large scale low-light simulation dataset,”IJCV, vol. 129, no. 7, pp. 2175–2193, 2021. 3

    2021

  63. [63]

    Learning temporal consistency for low light video enhancement from single images,

    F. Zhang, Y . Li, S. You, and Y . Fu, “Learning temporal consistency for low light video enhancement from single images,” inCVPR, 2021, pp. 4967–4976. 3

    2021

  64. [64]

    Eemefn: Low-light image enhancement via edge-enhanced multi-exposure fusion network,

    M. Zhu, P. Pan, W. Chen, and Y . Yang, “Eemefn: Low-light image enhancement via edge-enhanced multi-exposure fusion network,” in AAAI, vol. 34, no. 07, 2020, pp. 13 106–13 113. 4

    2020

  65. [65]

    U-net: Convolutional networks for biomedical image segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-net: Convolutional networks for biomedical image segmentation,” inMICCAI, 2015, pp. 234–241. 4

    2015

  66. [66]

    Restormer: Efficient transformer for high-resolution image restoration,

    S. W. Zamir, A. Arora, S. Khan, M. Hayat, F. S. Khan, and M.-H. Yang, “Restormer: Efficient transformer for high-resolution image restoration,” inCVPR, 2022, pp. 5728–5739. 5

    2022

  67. [67]

    Two-frame motion estimation based on polynomial expansion,

    G. Farneb ¨ack, “Two-frame motion estimation based on polynomial expansion,” inScandinavian conference on Image analysis. Springer, 2003, pp. 363–370. 7

    2003

  68. [68]

    Ntire 2019 challenge on video deblurring and super-resolution: Dataset and study,

    S. Nah, S. Baik, S. Hong, G. Moon, S. Son, R. Timofte, and K. Mu Lee, “Ntire 2019 challenge on video deblurring and super-resolution: Dataset and study,” inCVPRW, 2019, pp. 1974––1984. 7

    2019

  69. [69]

    Real-time intermediate flow estimation for video frame interpolation,

    Z. Huang, T. Zhang, W. Heng, B. Shi, and S. Zhou, “Real-time intermediate flow estimation for video frame interpolation,” inECCV, 2022, pp. 624–642. 7

    2022

  70. [70]

    v2e: From video frames to realistic dvs events,

    Y . Hu, S.-C. Liu, and T. Delbruck, “v2e: From video frames to realistic dvs events,” inCVPR, 2021, pp. 1312–1321. 7, 12

    2021

  71. [71]

    Glow in the dark: Low-light image enhancement with external memory,

    D. Ye, Z. Ni, W. Yang, H. Wang, S. Wang, and S. Kwong, “Glow in the dark: Low-light image enhancement with external memory,”IEEE TMM, vol. 26, pp. 2148–2163, 2023. 8, 9, 12

    2023