Pith. sign in

REVIEW 4 major objections 6 minor 240 references

This survey argues that fusing asynchronous event-camera streams with conventional frame capture has become a reliable, broadly applicable recipe for restoring degraded videos and reconstructing 3D scenes.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review

2026-08-04 18:21 UTC pith:JCAQ7BNM

load-bearing objection A useful but flawed map of event-frame fusion literature; fix the dataset table before release. the 4 major comments →

arxiv 2509.09971 v1 pith:JCAQ7BNM submitted 2025-09-12 cs.CV

Event Camera Guided Visual Media Restoration & 3D Reconstruction: A Survey

Aupendu Kar , Vishnu Raj , Guan-Ming Su This is my paper
classification cs.CV
keywords event cameravideo restorationframe interpolationmotion deblurringHDR imaginglow-light enhancement3D reconstructionneural radiance fields
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The authors set out to establish that event cameras—bio-inspired sensors that asynchronously report per-pixel brightness changes—are no longer a niche curiosity but a practical complement to standard RGB capture. By systematically organizing recent deep-learning work, they argue this fusion delivers consistent gains across temporal tasks (video reconstruction, frame interpolation, motion deblurring) and spatial tasks (super-resolution, HDR, low-light enhancement, artifact removal), and extends into photorealistic 3D reconstruction with NeRF and 3D Gaussian Splatting. A careful reader would care because the survey consolidates a fragmented field into a map with two axes, plus a compiled dataset table that makes the landscape actionable for benchmarking and future research.

Core claim

On the paper's own terms, the central discovery is that event data supplies exactly the information frame-based cameras lack: microsecond-resolution motion cues and edge/brightness-change signals that survive extreme lighting and fast motion. When fused with RGB frames, this complementary signal lifts restoration quality in ways that pure frame-based or pure event-based methods cannot match. The survey traces this benefit through three domains—temporal enhancement, spatial enhancement, and 3D reconstruction—and shows a clear historical drift from hand-crafted event-integration models toward deep networks, transformer-based hybrids, diffusion models, and spiking architectures. It also documen

What carries the argument

The load-bearing object is the event stream, a sparse sequence of tuples (x, y, p, t) recording the polarity, position, and microsecond timestamp of each per-pixel brightness change. Its formal event-generation model, ΔL = ∇L·v + ∂L/∂t, ties events to scene motion and brightness dynamics, and it is this continuous motion/brightness signal that the surveyed methods fuse with RGB frames—via voxel grids, time surfaces, graphs, spikes, or learned representations—to fill temporal gaps, sharpen motion, and constrain 3D optimization.

Load-bearing premise

The map is only as trustworthy as the survey's reading of each cited paper; if any key method is misdescribed or misattributed, the proposed organization of the field loses its reliability.

What would settle it

Take one representative method from each surveyed category (e.g., a frame-interpolation model) and re-run its reported ablation on a fixed benchmark, comparing the full event+RGB input against the same model with the event stream removed; if the event-conditioned version does not show a consistent, meaningful gain over the frame-only baseline across tasks, the survey's central thesis that event fusion benefits restoration would be called into question.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X Bluesky LinkedIn Reddit HN

If this is right

  • Event-guided frame interpolation and deblurring can handle non-linear and fast motion that frame-only methods miss, including rolling-shutter and blind-exposure cases.
  • Spatial enhancement tasks—super-resolution, HDR, low-light, occlusion removal, and rain removal—are improved by event fusion, with real-world datasets increasingly used.
  • 3D reconstruction with NeRF and 3D Gaussian Splatting can be made robust to motion blur, low light, and inaccurate camera poses by adding event streams.
  • The field is converging on unified multi-task frameworks that jointly address deblurring, interpolation, rolling-shutter correction, and continuous-time reconstruction.
  • The compiled dataset list exposes gaps—color events, multi-view data, semantic annotations, and broad lighting conditions—that the authors argue are the next barriers.

Where Pith is reading between the lines

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

  • If the fusion thesis holds generally, a single cross-modal backbone might be trained across all listed tasks; the survey's taxonomy implicitly predicts that task-specific designs will eventually be subsumed.
  • The cited finding that low-resolution sensors can beat high-resolution ones in low light suggests that event-density, not resolution, is the limiting factor; a testable design principle is to invest in noise reduction and temporal aggregation rather than pixel count.
  • The survey's emphasis on sensor non-idealities (latency, threshold variation, noise) implies that simulation-to-real transfer will only close the gap once simulators model these physical effects—an implied research direction.
  • Event-guided 3D may relax the need for precise camera poses, which could make consumer-grade hand-held 3D capture practical; that is an extrapolation beyond the surveyed results.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper is a survey of event-camera-guided visual media restoration and 3D reconstruction. It covers event camera fundamentals, event representations, simulators, event denoising/super-resolution, temporal enhancement (event-to-video reconstruction, frame interpolation, deblurring), spatial enhancement (super-resolution, HDR, low-light enhancement, de-occlusion, deraining, focus control), and recent NeRF/3D Gaussian Splatting methods that use event streams. It also includes a table of publicly available datasets and a list of future research opportunities. The central claim is that fusing asynchronous event streams with conventional frame-based capture significantly benefits restoration and 3D reconstruction, especially in fast-motion, low-light, and high-dynamic-range scenarios.

Significance. If the literature and dataset map are accurate, this survey fills a useful niche between general event-vision surveys and task-specific papers. Its organization along temporal enhancement, spatial enhancement, and 3D reconstruction is clear, and it captures recent developments through 2025, including diffusion models, spiking networks, and event-driven 3DGS. The dataset table is a practical contribution for benchmarking. The paper does not derive new methods; its value rests on faithful summarization and correct citation mapping. The overall thesis is plausible and consistent with spot-checked literature, but the survey's reliability as a guide depends on fixing the bibliographic and dataset-table errors detailed below.

major comments (4)
  1. [Table 2, RLED row] The RLED dataset is attributed to [121] (Liu et al., 'Seeing Motion at Nighttime with an Event Camera'), but Section 4.3 correctly credits RELED to [95] (Kim et al., 'Towards Real-World Event-Guided Low-Light Video Enhancement and Deblurring'). This is an internal contradiction in the benchmark map. A reader using Table 2 will be directed to the wrong paper for the RELED dataset. The citation should be corrected and every Table 2 dataset-source pair checked against the original publication.
  2. [Reference list, [215]/[216] and [221]/[222]] References [215] and [216] are the same paper, 'CrossZoom: Simultaneous Motion Deblurring and Event Super-Resolving', listed with different publication details and cited as separate items in Sections 2.4.3 and 3.2.2. Similarly, [221] and [222] are both 'Neuromorphic Imaging With Joint Image Deblurring and Event Denoising' and are cited separately in Sections 2.4.3 and 3.2.2. These duplicates make the numbered citation system ambiguous and prevent readers from distinguishing distinct works. The bibliography and in-text citations must be de-duplicated.
  3. [Table 2, DSEC row] The DSEC row lists tasks including 'Super-resolution' and 'Occlusion removal', but DSEC [54] is a stereo event camera dataset for driving scenarios and does not, to my knowledge, provide benchmarks for those tasks. This kind of task mislabeling in the dataset table is misleading for readers choosing a benchmark. The table should either be revised to state the actual benchmark tasks supported by each dataset or qualified with a note about the original paper's purpose. Adding a dedicated reference column would also improve verifiability.
  4. [Section 2.4.3 and reference [163]] The text states that 'Shariff et al. [163] integrate binary spikes with Sigma Delta Neural Networks (SDNNs)' for event super-resolution, but reference [163] is listed as 'Event cameras in automotive sensing: A review' by Shariff et al. This appears to be a citation mismatch: a survey paper would not present an SDNN-based method as a contribution. The citation should be corrected to the actual method paper, or the sentence should be reworded. This is part of a broader pattern of bibliographic inaccuracies that need systematic auditing.
minor comments (6)
  1. [Abstract] The phrase 'fusing event-stream captured with traditional frame-based capture' is ungrammatical. Suggest 'fusing event-stream data with traditional frame-based capture'.
  2. [Section 3.2.1] The author name is misspelled as 'Kiliccet et al.'; the reference [89] is by Kilicc et al. Please correct the spelling.
  3. [Section 4.3, reference [212]] The SEE-Net citation is incomplete: it has no venue, year, or publication identifier. The entry 'LU Yunfan et al. ... In: ()' should be filled in with the full bibliographic details.
  4. [Section 4.5] The phrase 'rain steaks' should be 'rain streaks'. Also 'Sun et al. [170] propose approach include' is grammatically incomplete and should be revised.
  5. [Table 2] The table would be easier to verify if a dedicated reference/URL column were added for every dataset. Currently some rows embed a reference in the dataset name while others (e.g., Erf-X170FPS, HighREV) do not clearly indicate their source paper.
  6. [Figure 4 caption] The caption uses 'marked as Green' and 'red-marked areas'; capitalization should be consistent (e.g., 'green' and 'red') and the sentence could be rephrased for clarity.

Circularity Check

0 steps flagged

No significant circularity: the survey's central claim rests on external literature, not on the authors' own definitions, fitted parameters, or self-citations.

full rationale

This paper is a literature survey, not a derivation or prediction pipeline. It introduces no new model, fits no parameters, and proves no theorem; its central claim—that fusing event streams with frame-based capture benefits restoration and 3D reconstruction—is supported by summaries of externally published methods, benchmarks, and datasets. The only self-citations ([31], [32], [76]) appear as examples of event representation techniques in Section 2.2 and are not load-bearing: the survey's conclusions would be unchanged if these entries were removed. The dataset table (Table 2) contains apparent attribution and duplication errors (e.g., the RLED row cites [121] while the RELED dataset is introduced in [95]; [215] and [216] both describe CrossZoom), but these are factual correctness/reliability concerns, not circular reasoning. There is no equation or claimed result that reduces by construction to its own input, no fitted quantity is renamed as a prediction, and no uniqueness or foundational premise is imported solely from the authors' prior work. Accordingly, the paper is not circular in the sense defined here.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

No free parameters or invented entities; the paper is a literature review and introduces no new formalism, model, or fitted quantity.

axioms (2)
  • domain assumption The cited papers report results as summarized in this survey.
    The survey's conclusions about the state of the art rest on the fidelity of its summaries of the cited works; this is an unverifiable assumption in a review.
  • domain assumption Event cameras and frame cameras can be meaningfully fused for the tasks discussed.
    The survey takes the benefit of event-frame fusion as its organizing premise, citing prior work rather than proving it.

reviewed 2026-08-04 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Event Camera Guided Visual Media Restoration & 3D Reconstruction: A Survey." pith.science (2026). https://pith.science/paper/JCAQ7BNM

@misc{pith2026250909971,
  author       = {Pith},
  title        = {Pith review of: Event Camera Guided Visual Media Restoration & 3D Reconstruction: A Survey},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JCAQ7BNM}},
  note         = {Machine review of arXiv:2509.09971}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Event camera sensors are bio-inspired sensors which asynchronously capture per-pixel brightness changes and output a stream of events encoding the polarity, location and time of these changes. These systems are witnessing rapid advancements as an emerging field, driven by their low latency, reduced power consumption, and ultra-high capture rates. This survey explores the evolution of fusing event-stream captured with traditional frame-based capture, highlighting how this synergy significantly benefits various video restoration and 3D reconstruction tasks. The paper systematically reviews major deep learning contributions to image/video enhancement and restoration, focusing on two dimensions: temporal enhancement (such as frame interpolation and motion deblurring) and spatial enhancement (including super-resolution, low-light and HDR enhancement, and artifact reduction). This paper also explores how the 3D reconstruction domain evolves with the advancement of event driven fusion. Diverse topics are covered, with in-depth discussions on recent works for improving visual quality under challenging conditions. Additionally, the survey compiles a comprehensive list of openly available datasets, enabling reproducible research and benchmarking. By consolidating recent progress and insights, this survey aims to inspire further research into leveraging event camera systems, especially in combination with deep learning, for advanced visual media restoration and enhancement.

Figures

Figures reproduced from arXiv: 2509.09971 by Aupendu Kar, Guan-Ming Su, Vishnu Raj.

Figure 1
Figure 1. Figure 1: Organization of the study. tion blur artifacts in deblurring tasks [171]. These methods leverage event cameras’ µ-second-resolution temporal data to overcome traditional vision bottlenecks, making them highly beneficial for applications in autonomous navigation, computational photography, and real-time augmented real￾ity. Deep learning can unlock unprecedented visual fidelity in dynamic environments by ali… view at source ↗
Figure 2
Figure 2. Figure 2: An example of a captured intensity frame and its [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: The typical pipeline of how temporal (in Section 3) and spatial (in Section 4) enhancement methods work with event [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: The summary of recent works in photorealistic 3D reconstruction using NeRF and 3DGS with the help of Events. [PITH_FULL_IMAGE:figures/full_fig_p019_4.png] view at source ↗

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

240 extracted references · 32 linked inside Pith

  1. [1]

    Investigation of event-based sur- faces for high-speed detection, unsupervised feature extraction, and object recognition

    Saeed Afshar et al. “Investigation of event-based sur- faces for high-speed detection, unsupervised feature extraction, and object recognition”. In:Frontiers in neuroscience12 (2019), p. 1047

    2019

  2. [2]

    MAEVI: Motion Aware Event-Based Video Frame Interpolation

    Ahmet Akman, Onur Selim Kilicc, and Aydin Alatan. “MAEVI: Motion Aware Event-Based Video Frame Interpolation”. In:ArXivabs/2303.02025 (2023).URL: https : / / api . semanticscholar . org / CorpusID : 257353657

    Pith/arXiv arXiv 2023

  3. [3]

    ACE: An efficient asynchronous corner tracker for event cam- eras

    Ignacio Alzugaray and Margarita Chli. “ACE: An efficient asynchronous corner tracker for event cam- eras”. In:2018 International Conference on 3D Vision (3DV). IEEE. 2018, pp. 653–661

    2018

  4. [4]

    Accurate and efficient frame-based event representation for aer object recognition

    WeiJie Bai et al. “Accurate and efficient frame-based event representation for aer object recognition”. In: 2022 International Joint Conference on Neural Networks (IJCNN). IEEE. 2022, pp. 1–6

    2022

  5. [5]

    Time-ordered recent event (tore) volumes for event cameras

    R Wes Baldwin et al. “Time-ordered recent event (tore) volumes for event cameras”. In:IEEE Transac- tions on Pattern Analysis and Machine Intelligence45.2 (2022), pp. 2519–2532

    2022

  6. [6]

    Improving fast auto-focus with event polarity

    Yuhan Bao et al. “Improving fast auto-focus with event polarity”. In:Optics Express31.15 (2023), pp. 24025–24044

    2023

  7. [7]

    One-Step Event-Driven High- Speed Autofocus

    Yuhan Bao et al. “One-Step Event-Driven High- Speed Autofocus”. In:Proceedings of the Computer Vision and Pattern Recognition Conference (CVPR). June 2025, pp. 6222–6230

    2025

  8. [8]

    Temporal-mapping photography for event cameras

    Yuhan Bao et al. “Temporal-mapping photography for event cameras”. In:European Conference on Com- puter Vision. Springer. 2024, pp. 55–72

    2024

  9. [9]

    Simultaneous optical flow and inten- sity estimation from an event camera

    Patrick Bardow, Andrew J Davison, and Stefan Leutenegger. “Simultaneous optical flow and inten- sity estimation from an event camera”. In:Proceedings of the IEEE conference on computer vision and pattern recognition. 2016, pp. 884–892

    2016

  10. [10]

    E3d: event-based 3d shape reconstruction

    Alexis Baudron et al. “E3d: event-based 3d shape reconstruction”. In:arXiv preprint arXiv:2012.05214 (2020)

    Pith/arXiv arXiv 2012

  11. [11]

    Event-based visual flow

    Ryad Benosman et al. “Event-based visual flow”. In:IEEE transactions on neural networks and learning systems25.2 (2013), pp. 407–417

    2013

  12. [12]

    Evdnerf: Reconstructing event data with dynamic neural radiance fields

    Anish Bhattacharya et al. “Evdnerf: Reconstructing event data with dynamic neural radiance fields”. In:Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. 2024, pp. 5846–5855

    2024

  13. [13]

    Graph-based spatio-temporal feature learning for neuromorphic vision sensing

    Yin Bi et al. “Graph-based spatio-temporal feature learning for neuromorphic vision sensing”. In:IEEE Transactions on Image Processing29 (2020), pp. 9084– 9098

    2020

  14. [14]

    A 240×180 130 db 3µs latency global shutter spatiotemporal vision sensor

    Christian Brandli et al. “A 240×180 130 db 3µs latency global shutter spatiotemporal vision sensor”. In:IEEE Journal of Solid-State Circuits49.10 (2014), pp. 2333–2341

    2014

  15. [15]

    SPADE-E2VID: Spatially-Adaptive Denormalization for Event-Based Video Reconstruction

    Pablo Rodrigo Gantier Cadena et al. “SPADE-E2VID: Spatially-Adaptive Denormalization for Event-Based Video Reconstruction”. In:IEEE Transactions on Image Processing30 (2021), pp. 2488–2500.URL: https://api. semanticscholar.org/CorpusID:231759393

    2021

  16. [16]

    Mitigating motion blur in neural radiance fields with events and frames

    Marco Cannici and Davide Scaramuzza. “Mitigating motion blur in neural radiance fields with events and frames”. In:Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition. 2024, pp. 9286–9296

    2024

  17. [17]

    A differentiable recurrent sur- face for asynchronous event-based data

    Marco Cannici et al. “A differentiable recurrent sur- face for asynchronous event-based data”. In:Com- puter Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XX

    2020

  18. [18]

    2020, pp

    Springer. 2020, pp. 136–152

    2020

  19. [19]

    Event-driven video restoration with spiking-convolutional architecture

    Chengzhi Cao et al. “Event-driven video restoration with spiking-convolutional architecture”. In:IEEE Transactions on Neural Networks and Learning Systems (2023)

    2023

  20. [20]

    Noise2Image: Noise-Enabled Static Scene Recovery for Event Cameras

    Ruiming Cao et al. “Noise2Image: Noise-Enabled Static Scene Recovery for Event Cameras”. In: ArXivabs/2404.01298 (2024).URL: https : / / api . semanticscholar.org/CorpusID:268856956. 23

    Pith/arXiv arXiv 2024

  21. [21]

    Event-based 3D reconstruction from neuromorphic retinas

    Jo ˜ao Carneiro et al. “Event-based 3D reconstruction from neuromorphic retinas”. In:Neural Networks45 (2013), pp. 27–38

    2013

  22. [22]

    Recent event cam- era innovations: A survey

    Bharatesh Chakravarthi et al. “Recent event cam- era innovations: A survey”. In:arXiv preprint arXiv:2408.13627(2024)

    Pith/arXiv arXiv 2024

  23. [23]

    Revisiting Event-Based Video Frame Interpolation

    Jiaben Chen et al. “Revisiting Event-Based Video Frame Interpolation”. In:2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE. 2023, pp. 1292–1299

    2023

  24. [24]

    Repurposing pre-trained video diffusion models for event-based video interpola- tion

    Jingxi Chen et al. “Repurposing pre-trained video diffusion models for event-based video interpola- tion”. In:Proceedings of the Computer Vision and Pattern Recognition Conference. 2025, pp. 12456–12466

    2025

  25. [25]

    EvLight++: Low-Light Video Enhancement with an Event Camera: A Large-Scale Real-World Dataset, Novel Method, and More

    Kanghao Chen et al. “EvLight++: Low-Light Video Enhancement with an Event Camera: A Large-Scale Real-World Dataset, Novel Method, and More”. In: arXiv preprint arXiv:2408.16254(2024)

    Pith/arXiv arXiv 2024

  26. [26]

    A simple framework for contrastive learning of visual representations

    Ting Chen et al. “A simple framework for contrastive learning of visual representations”. In:International conference on machine learning. PmLR. 2020, pp. 1597– 1607

    2020

  27. [27]

    Learning continuous image representation with local implicit image function

    Yinbo Chen, Sifei Liu, and Xiaolong Wang. “Learning continuous image representation with local implicit image function”. In:Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition. 2021, pp. 8628–8638

    2021

  28. [28]

    A novel rain removal approach for outdoor dynamic vision sensor event videos

    Long Cheng et al. “A novel rain removal approach for outdoor dynamic vision sensor event videos”. In: Frontiers in Neurorobotics16 (2022), p. 928707

    2022

  29. [29]

    Recovering Continuous Scene Dynamics from A Single Blurry Image with Events

    Zhang Cheng et al. “Recovering Continuous Scene Dynamics from A Single Blurry Image with Events”. In:ArXivabs/2304.02695 (2023).URL: https://api. semanticscholar.org/CorpusID:257984979

    Pith/arXiv arXiv 2023

  30. [30]

    Non-Coaxial Event-guided Mo- tion Deblurring with Spatial Alignment

    Hoonhee Cho et al. “Non-Coaxial Event-guided Mo- tion Deblurring with Spatial Alignment”. In:2023 IEEE/CVF International Conference on Computer Vision (ICCV)(2023), pp. 12458–12469.URL: https : / / api . semanticscholar.org/CorpusID:267024980

    2023

  31. [31]

    TTA-EVF: Test-Time Adap- tation for Event-based Video Frame Interpolation via Reliable Pixel and Sample Estimation

    Hoonhee Cho et al. “TTA-EVF: Test-Time Adap- tation for Event-based Video Frame Interpolation via Reliable Pixel and Sample Estimation”. In:2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)(2024), pp. 25701–25711.URL: https : / / api . semanticscholar . org / CorpusID : 272722967

    2024

  32. [32]

    Modeling Event Camera Frame Sequence using Neural Field

    Anustup Choudhury and Guan-Ming Su. “Modeling Event Camera Frame Sequence using Neural Field”. In:IEEE International Conference on Multimedia Infor- mation Processing and Retrieval. 2025

    2025

  33. [33]

    Triplane Learning for Event Stream Repre- sentation

    Anustup Choudhury, Guan-Ming Su, and Jingxi Chen. “Triplane Learning for Event Stream Repre- sentation”. In:59th Asilomar Conference on Signals, Systems and Computers. 2025

    2025

  34. [34]

    Interacting maps for fast visual interpretation

    Matthew Cook et al. “Interacting maps for fast visual interpretation”. In:The 2011 International Joint Confer- ence on Neural Networks. IEEE. 2011, pp. 770–776

    2011

  35. [35]

    Color event enhanced single- exposure HDR imaging

    Mengyao Cui et al. “Color event enhanced single- exposure HDR imaging”. In:Proceedings of the AAAI Conference on Artificial Intelligence. Vol. 38. 2. 2024, pp. 1399–1407

    2024

  36. [36]

    Amae: Adaptive motion-agnostic encoder for event-based object classification

    Yongjian Deng, Youfu Li, and Hao Chen. “Amae: Adaptive motion-agnostic encoder for event-based object classification”. In:IEEE Robotics and Automa- tion Letters5.3 (2020), pp. 4596–4603

    2020

  37. [37]

    A voxel graph cnn for object classification with event cameras

    Yongjian Deng et al. “A voxel graph cnn for object classification with event cameras”. In:Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022, pp. 1172–1181

    2022

  38. [38]

    E-MLB: Multilevel benchmark for event-based camera denoising

    Saizhe Ding et al. “E-MLB: Multilevel benchmark for event-based camera denoising”. In:IEEE Transactions on Multimedia26 (2023), pp. 65–76

    2023

  39. [39]

    EventAid: Benchmarking event- aided image/video enhancement algorithms with real-captured hybrid dataset

    Peiqi Duan et al. “EventAid: Benchmarking event- aided image/video enhancement algorithms with real-captured hybrid dataset”. In:IEEE Transactions on Pattern Analysis and Machine Intelligence(2025)

    2025

  40. [40]

    EventZoom: Learning to denoise and super resolve neuromorphic events

    Peiqi Duan et al. “EventZoom: Learning to denoise and super resolve neuromorphic events”. In:Proceed- ings of the IEEE/CVF conference on computer vision and pattern recognition. 2021, pp. 12824–12833

    2021

  41. [41]

    Neurozoom: Denoising and super resolving neuromorphic events and spikes

    Peiqi Duan et al. “Neurozoom: Denoising and super resolving neuromorphic events and spikes”. In:IEEE Transactions on Pattern Analysis and Machine Intelli- gence45.12 (2023), pp. 15219–15232

    2023

  42. [42]

    LED: A large-scale real-world paired dataset for event camera denoising

    Yuxing Duan. “LED: A large-scale real-world paired dataset for event camera denoising”. In:Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2024, pp. 25637–25647

    2024

  43. [43]

    Hypere2VID: Improving event- based video reconstruction via hypernetworks

    Burak Ercan et al. “Hypere2VID: Improving event- based video reconstruction via hypernetworks”. In: IEEE Transactions on Image Processing(2024)

    2024

  44. [44]

    AE-NeRF: Augmenting Event- Based Neural Radiance Fields for Non-ideal Con- ditions and Larger Scene

    Chaoran Feng et al. “AE-NeRF: Augmenting Event- Based Neural Radiance Fields for Non-ideal Con- ditions and Larger Scene”. In:arXiv preprint arXiv:2501.02807(2025)

    Pith/arXiv arXiv 2025

  45. [45]

    Event-Driven Heterogeneous Network for Video Deraining

    Xueyang Fu et al. “Event-Driven Heterogeneous Network for Video Deraining”. In:International Jour- nal of Computer Vision132.12 (2024), pp. 5841–5861

    2024

  46. [46]

    Focus is all you need: Loss func- tions for event-based vision

    Guillermo Gallego, Mathias Gehrig, and Davide Scaramuzza. “Focus is all you need: Loss func- tions for event-based vision”. In:Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2019, pp. 12280–12289

    2019

  47. [47]

    Event-based vision: A sur- vey

    Guillermo Gallego et al. “Event-based vision: A sur- vey”. In:IEEE transactions on pattern analysis and machine intelligence44.1 (2020), pp. 154–180

    2020

  48. [48]

    VECtor: A Versatile Event-Centric Benchmark for Multi-Sensor SLAM

    Ling Gao et al. “VECtor: A Versatile Event-Centric Benchmark for Multi-Sensor SLAM”. In:IEEE Robotics and Automation Letters7 (2022), pp. 8217– 8224

    2022

  49. [49]

    Sim- cse: Simple contrastive learning of sentence embed- dings

    Tianyu Gao, Xingcheng Yao, and Danqi Chen. “Sim- cse: Simple contrastive learning of sentence embed- dings”. In:arXiv preprint arXiv:2104.08821(2021)

    Pith/arXiv arXiv 2021

  50. [50]

    SuperFast: 200× Video Frame Inter- polation via Event Camera

    Yue Gao et al. “SuperFast: 200× Video Frame Inter- polation via Event Camera”. In:IEEE Transactions on Pattern Analysis and Machine Intelligence45 (2022), pp. 7764–7780.URL: https : / / api . semanticscholar. org/CorpusID:253838481

    2022

  51. [51]

    EventMamba: Enhancing Spatio- Temporal Locality with State Space Models for Event-Based Video Reconstruction

    Chengjie Ge et al. “EventMamba: Enhancing Spatio- Temporal Locality with State Space Models for Event-Based Video Reconstruction”. In:Proceedings of 24 the AAAI Conference on Artificial Intelligence. Vol. 39. 3. 2025, pp. 3104–3112

    2025

  52. [52]

    Neuromorphic event signal- driven network for video de-raining

    Chengjie Ge et al. “Neuromorphic event signal- driven network for video de-raining”. In:Proceedings of the AAAI Conference on Artificial Intelligence. Vol. 38

  53. [53]

    Are high- resolution event cameras really needed?

    Daniel Gehrig and Davide Scaramuzza. “Are high- resolution event cameras really needed?” In:arXiv preprint arXiv:2203.14672(2022)

    Pith/arXiv arXiv 2022

  54. [54]

    End-to-end learning of repre- sentations for asynchronous event-based data

    Daniel Gehrig et al. “End-to-end learning of repre- sentations for asynchronous event-based data”. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. 2019, pp. 5633–5643

    2019

  55. [55]

    Dsec: A stereo event camera dataset for driving scenarios

    Mathias Gehrig et al. “Dsec: A stereo event camera dataset for driving scenarios”. In:IEEE Robotics and Automation Letters6.3 (2021), pp. 4947–4954

    2021

  56. [56]

    Mamba: Linear-time se- quence modeling with selective state spaces

    Albert Gu and Tri Dao. “Mamba: Linear-time se- quence modeling with selective state spaces”. In: arXiv preprint arXiv:2312.00752(2023)

    Pith/arXiv arXiv 2023

  57. [57]

    Ef- ficiently modeling long sequences with structured state spaces

    Albert Gu, Karan Goel, and Christopher R ´e. “Ef- ficiently modeling long sequences with structured state spaces”. In:arXiv preprint arXiv:2111.00396 (2021)

    Pith/arXiv arXiv 2021

  58. [58]

    The spatio-temporal poisson point process: A simple model for the alignment of event camera data

    Cheng Gu et al. “The spatio-temporal poisson point process: A simple model for the alignment of event camera data”. In:Proceedings of the IEEE/CVF Interna- tional Conference on Computer Vision. 2021, pp. 13495– 13504

    2021

  59. [59]

    How to learn a domain-adaptive event simulator?

    Daxin Gu et al. “How to learn a domain-adaptive event simulator?” In:Proceedings of the 29th ACM International Conference on Multimedia. 2021, pp. 1275– 1283

    2021

  60. [60]

    Tactilesgnet: A spiking graph neural network for event-based tactile object recog- nition

    Fuqiang Gu et al. “Tactilesgnet: A spiking graph neural network for event-based tactile object recog- nition”. In:2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE. 2020, pp. 9876–9882

    2020

  61. [61]

    Event-guided image super- resolution reconstruction

    Guangsha Guo et al. “Event-guided image super- resolution reconstruction”. In:Sensors23.4 (2023), p. 2155

    2023

  62. [62]

    REDIR: Refocus-Free Event-Based De- occlusion Image Reconstruction

    Qi Guo et al. “REDIR: Refocus-Free Event-Based De- occlusion Image Reconstruction”. In:European Con- ference on Computer Vision. Springer. 2024, pp. 419– 435

    2024

  63. [63]

    Event-assisted 12-stop HDR Imaging of Dynamic Scene

    Shi Guo et al. “Event-assisted 12-stop HDR Imaging of Dynamic Scene”. In:arXiv preprint arXiv:2412.14705(2024)

    Pith/arXiv arXiv 2024

  64. [64]

    Improved Event-Based Image De- Occlusion

    Yufei Guo et al. “Improved Event-Based Image De- Occlusion”. In:IEEE Signal Processing Letters(2024)

    2024

  65. [65]

    Event-3DGS: Event-based 3D Reconstruction Using 3D Gaussian Splatting

    Haiqian Han et al. “Event-3DGS: Event-based 3D Reconstruction Using 3D Gaussian Splatting”. In: Advances in Neural Information Processing Systems37 (2025), pp. 128139–128159

    2025

  66. [66]

    Physical-Based Event Camera Simulator

    Haiqian Han et al. “Physical-Based Event Camera Simulator”. In:European Conference on Computer Vi- sion. Springer. 2024, pp. 19–35

    2024

  67. [67]

    Evintsr-net: Event guided multiple la- tent frames reconstruction and super-resolution

    Jin Han et al. “Evintsr-net: Event guided multiple la- tent frames reconstruction and super-resolution”. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. 2021, pp. 4882–4891

    2021

  68. [68]

    Hybrid high dynamic range imaging fusing neuromorphic and conventional images

    Jin Han et al. “Hybrid high dynamic range imaging fusing neuromorphic and conventional images”. In: IEEE Transactions on pattern analysis and machine intel- ligence45.7 (2023), pp. 8553–8565

    2023

  69. [69]

    Neuromorphic camera guided high dynamic range imaging

    Jin Han et al. “Neuromorphic camera guided high dynamic range imaging”. In:Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2020, pp. 1730–1739

    2020

  70. [70]

    Momentum contrast for unsuper- vised visual representation learning

    Kaiming He et al. “Momentum contrast for unsuper- vised visual representation learning”. In:Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2020, pp. 9729–9738

    2020

  71. [71]

    Timereplayer: Unlocking the po- tential of event cameras for video interpolation

    Weihua He et al. “Timereplayer: Unlocking the po- tential of event cameras for video interpolation”. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022, pp. 17804–17813

    2022

  72. [72]

    Event-aided direct sparse odom- etry

    Javier Hidalgo-Carri ´o, Guillermo Gallego, and Da- vide Scaramuzza. “Event-aided direct sparse odom- etry”. In:Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022, pp. 5781–5790

    2022

  73. [73]

    De- noising Diffusion Probabilistic Models

    Jonathan Ho, Ajay Jain, and P . Abbeel. “De- noising Diffusion Probabilistic Models”. In:ArXiv abs/2006.11239 (2020).URL: https : / / api . semanticscholar.org/CorpusID:219955663

    Pith/arXiv arXiv 2006

  74. [74]

    Learning to exploit multiple vision modalities by using grafted networks

    Yuhuang Hu, Tobi Delbruck, and Shih-Chii Liu. “Learning to exploit multiple vision modalities by using grafted networks”. In:European Conference on Computer Vision. Springer. 2020, pp. 85–101

    2020

  75. [75]

    v2e: From video frames to realistic DVS events

    Yuhuang Hu, Shih-Chii Liu, and Tobi Delbruck. “v2e: From video frames to realistic DVS events”. In:Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2021, pp. 1312–1321

    2021

  76. [76]

    IncEventGS: Pose-Free Gaussian Splatting from a Single Event Camera

    Jian Huang et al. “IncEventGS: Pose-Free Gaussian Splatting from a Single Event Camera”. In:Proceed- ings of the Computer Vision and Pattern Recognition Conference (CVPR). June 2025, pp. 26933–26942

    2025

  77. [77]

    Neural Representations For Event Voxel Grid

    Tsung-Wei Huang, Anustup Choudhury, and Guan- Ming Su. “Neural Representations For Event Voxel Grid”. In:IEEE International Conference on Image Pro- cessing (ICIP) 2025 workshop on Generative AI for World Simulations and Communications. 2025

    2025

  78. [78]

    Eventpoint: Self-supervised interest point detection and description for event-based cam- era

    Ze Huang et al. “Eventpoint: Self-supervised interest point detection and description for event-based cam- era”. In:Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. 2023, pp. 5396– 5405

    2023

  79. [79]

    Bilateral event mining and com- plementary for event stream super-resolution

    Zhilin Huang et al. “Bilateral event mining and com- plementary for event stream super-resolution”. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2024, pp. 34–43

    2024

  80. [80]

    Ev-nerf: Event based neural radiance field

    Inwoo Hwang, Junho Kim, and Young Min Kim. “Ev-nerf: Event based neural radiance field”. In: Proceedings of the IEEE/CVF Winter Conference on Ap- plications of Computer Vision. 2023, pp. 837–847

    2023

Showing first 80 references.

This paper was first reviewed by deepseek-v4-flash on August 4, 2026.