Pith. sign in

REVIEW 5 major objections 5 minor 113 references

EvReflection: Event-Driven Micro-Dynamics for Reflection Removal

T0 review · 5 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Event cameras break the reflection-removal ambiguity by exploiting depth-dependent micro-motion

desk verdict First event-camera reflection removal paper with a promising new dataset and large claimed gains, but the theoretical uniqueness proof is internally inconsistent and the evaluation has fixable protocol gaps. read the letter →

arxiv 2608.06184 v1 pith:NRZECM6S submitted 2026-08-06 cs.CV

Jiaxiao Wang , Dachun Kai , Huyue Zhu , Quanquan Hu , Zhenyang Xu , Xiaoyan Sun This is my paper
classification cs.CV
keywords eventcamerareflectionremovalmotionparallaxmicro-dynamicslayerseparationevent-drivennetworkEVR2datasetparallax-awaresimulationcross-modalattention
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper claims that reflection removal, which is ill-posed for a single image because infinitely many transmission/reflection pairs sum to the same observation, becomes well-posed when the camera records event signals during subtle handheld motion. Because the glass reflection and the background behind it sit at different depths, a tiny camera shake moves the two layers differently, and event cameras capture that differential motion at microsecond resolution. The paper builds a network, EvReflection, that turns event streams into layer-specific motion priors and uses those priors to suppress reflection artifacts in an RGB image, reporting state-of-the-art results on both the synthetic SIR2 benchmark (29.09 dB) and a new real-world EVR2 dataset (27.25 dB). If these results hold, event cameras supply a physically grounded cue that lets reflection removal work under casual handheld motion instead of requiring large camera baselines or strong static-image priors.

What carries the argument

The load-bearing object is the linear system A x = b built from event observations, where the rows of A record the scalar velocity magnitudes of the transmission and reflection layers at successive instants inside a micro-second window. The claim that this system has a unique solution rests on the Hessian H = A^T A being positive definite, which the paper equates with the two layers' motion magnitude vectors S_T and S_R being linearly independent, i.e., with real motion parallax v_T(t) != v_R(t). On the network side, the carrying mechanism is the Micro-Dynamics Decoupler, a dual-branch recurrent module that iteratively separates the mixed event stream into distinct transmission and reflection dynamics priors, and the Parallax-Attention Rectifier, a cross-modal attention block that uses those priors as query keys to decide which spatial regions of the RGB feature map belong to transmission and which to reflection.

What would settle it

The cleanest check is to re-derive the velocity sequences under the paper's own pure-translation model: if the two depth layers have speeds that differ only by a constant factor, the matrix A has proportional columns and the Hessian is singular, which would contradict the claimed positive-definiteness. An experimental falsifier is to record EVR2-style scenes with a purely translational slide (no rotation or speed variation) and measure whether EvReflection's margin over RGB baselines shrinks or vanishes, since the theory predicts the motion cue should then be uninformative.

Watch

Extended reading notes

Core claim

The central claim is that event signals of micro-dynamics make reflection removal well-posed. The paper models the observed image as a sum of transmission and reflection layers, relates the event stream to the temporal derivative of log intensity, and derives an Event-Gradient Constraint: the measured signal equals the negative gradient of each layer projected along its own motion. Aggregating several event observations within a short window yields a linear system Ax=b whose Hessian A^T A, the paper argues, is positive definite when the two layers move at different speeds. The paper states that distinct micro-dynamics guarantee a unique solution, resolving the ambiguity that plagues single-image methods. It then instantiates this theory as EvReflection, whose Micro-Dynamics Decoupler (MDD) separates the event stream into transmission and reflection motion priors and whose Parallax-Attention Rectifier (PAR) uses those priors to spatially modulate RGB features and remove artifacts. The empirical claim is a gain of 1.63 dB over the best RGB competitor on SIR2 (29.09 versus 27.46 dB) and 1.28 dB on the real-world EVR2 dataset (27.25 versus 25.97 dB).

Load-bearing premise

The whole argument depends on the claim that subtle camera shake makes the reflection and transmission layers move at different speeds over time, so the measured event signals can pick out each layer's motion; if the two layers' motions are proportional or too weak to detect, the uniqueness argument collapses and the event cue degenerates.

Editorial extensions

If this is right

  • Handheld devices with event-RGB hybrid sensors could remove reflections from casual photographs without requiring the user to move the camera deliberately or capture multiple views with a large baseline.
  • The EVR2 dataset, the first real-world event-based reflection removal benchmark, gives the field a common test bed stratified by glass thickness, and the parallax-aware simulator offers a way to generate paired training data for other event-guided layer-separation tasks.
  • Reflection removal performance degrades gracefully as glass thickens: the method still leads competitors on 8 mm glass, where double-reflection ghosting is strongest.
  • If the theoretical argument is correct, similar event-gradient constraints could be applied to other additive superposition problems, such as removing rain, snow, or glare, where the degradations lie at different depths than the scene.

Reading between the lines

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

  • The paper's own motion model for a translating camera makes the linear-independence argument questionable: when flow magnitude scales as 1/Z under pure translation, the two layers' speed vectors differ only by a constant ratio, so they are proportional and the Hessian is singular. The unique-solution claim therefore needs the camera to rotate or the trajectory to vary in a way that produces genuin
  • A directly testable extension would be a controlled experiment in pure lateral translation on the EVR2 rig: if the theory's necessary condition fails under pure translation, performance should drop measurably compared with trajectories containing rotation, even though depth differences are present.
  • The method seems likely to transfer to event-guided de-raining, de-hazing, or glare suppression, where the offending layer occupies a distinct depth plane, but those tasks would need their own simulators and benchmarks because reflection has the special property that the virtual layer lies behind the glass.
  • The ablation numbers suggest that motion priors matter most on thick glass, so future work could specialize the attention mechanism to the larger displacement regimes rather than treating all thicknesses uniformly.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 5 minor

Summary. The paper introduces EvReflection, an event-driven reflection removal method that exploits micro-dynamics captured by event cameras. It proposes a Micro-Dynamics Decoupler (MDD) and a Parallax-Attention Rectifier (PAR), together with a parallax-aware simulation pipeline and a new real-world benchmark dataset (EVR2). The paper reports state-of-the-art results on SIR2 (29.09 dB PSNR) and EVR2 (27.25 dB PSNR), and it includes a theoretical claim in Section 3.1 and Appendix B that event-based motion parallax makes the layer-gradient estimation problem well-posed.

Significance. The core idea is genuinely novel and potentially influential: event cameras provide a new cue for reflection removal under subtle handheld motion, and the release of code and a real dataset would support follow-up work. The empirical gains over RGB-only baselines are large, and the ablations in Section 4.4 and Appendix C consistently show that the event modality and the proposed modules help. The explicit limitation section (static scenes, low light, computational cost) is a useful honesty check. However, the advertised theoretical guarantee is not established under the paper's own motion model, and several evaluation choices prevent the empirical claims from being fully trusted as reported.

major comments (5)
  1. [Section 3.1 and Appendix B.2, Eqs. (31)-(33)] The proof that H is positive definite asserts that ZT != ZR implies the velocity-magnitude vectors ST and SR are linearly independent. Under the model actually written in Eq. (31), where optical flow for a translating camera is proportional to 1/Z, the two magnitude vectors are exactly proportional: S_T = (Z_R / Z_T) S_R. Therefore the columns of A in Eq. (2) are scalar multiples, rank(A) = 1, H is singular, and the least-squares problem has infinitely many minimizers. Pointwise inequality v_T(t) != v_R(t) does not imply linear independence of the sampled vectors, so the claimed uniqueness guarantee does not follow. The authors need to either state an explicit additional assumption (for example, a depth-independent rotational flow component, or a time-varying depth ratio) and prove independence under that assumption, or temper the well-posedness claim in Section 3.1.
  2. [Section 4.1 vs Section 4.3] The simulation pipeline in Section 4.1 states that images from both PASCAL VOC and SIR2 are used as source layers, while Section 4.3 says the model is trained exclusively on 7,643 pairs from PASCAL VOC. If any SIR2 images are used to generate synthetic training pairs, then evaluating on the SIR2 benchmark (Table 1) is contaminated by training data leakage. Please clarify the exact composition of the training set and confirm that no SIR2 test image, or close variant, appears during training.
  3. [Section 4.1, SIR2 evaluation protocol] The paper does not specify how event streams are generated for the static SIR2 benchmark images. Because the method is trained on events synthesized with the same parallax-aware pipeline, the SIR2 numbers could reflect a simulator-specific parallax cue rather than a robust property of real event streams. Please provide the full protocol: virtual depths for transmission and reflection, the camera trajectory C(t), glass thickness or offset parameters, event thresholds, and the exact set of source images used to create each test event stream. Ideally, also evaluate on events captured by a real event camera for at least a subset of SIR2-style scenes.
  4. [Section 4.1, EVR2 test-set construction] The EVR2 test set is selected by a 'performance-guided stratified sampling' strategy that keeps only 11 of 140 scenes, chosen by mapping the error distribution of a baseline model. This procedure is vulnerable to selection bias unless the rule is fixed before seeing the proposed model's outputs and is applied only to a development partition. Please report the full 140-scene aggregate results or a random, pre-registered split, and specify the selection criterion in enough detail that the benchmark is reproducible.
  5. [Table 2, baseline retraining consistency] The real-world comparison treats baselines inconsistently: methods marked with an asterisk are evaluated with official pre-trained models, while unmarked methods are retrained on EVR2. Differences in training data, resolution, and augmentation make the reported margin over the asterisked baselines difficult to interpret. Please retrain all baselines under the same protocol, or clearly justify why official weights are the appropriate comparison, and report which training split and preprocessing each baseline used.
minor comments (5)
  1. [Section 4.1 and Table 2] The text alternates between 'EVR2' and 'EVR 2' (with a space), and Table 2 headers such as 'EVR2-T3' and 'EVR 2-T5' are inconsistent; please normalize the notation throughout.
  2. [Section 3.1, Eq. (1)] The notation 'I · E' in Eq. (1) is ambiguous because I is a scalar intensity and E is defined as a continuous event signal; please clarify whether this denotes pointwise multiplication and define the domain of each quantity explicitly.
  3. [Section 4.2, loss function] The paper says the total loss 'strictly adheres' to RDNet's objective but does not give the weights of the MSE, gradient, and perceptual terms; please provide the exact loss formula and weights for reproducibility.
  4. [Appendix F, Table F.1] The MDD iteration count is listed as M=12 in Table F.1 but the main text in Section 3.3 only says 'N iterations' and never defines N; please reconcile the notation and state the value used in the experiments.
  5. [Conclusion, Limitation paragraph] The limitation paragraph correctly notes that the method struggles in strictly static scenes, but the Introduction and Abstract should be rephrased so that the need for micro-dynamics is stated up front rather than only in the final limitation section.

Circularity Check

1 steps flagged · score 6.0 of 10

The well-posedness theorem's key step asserts ST≠kSR from depth difference, but under the paper's own model distinct depths force ST∝SR, so the uniqueness guarantee is assumed by construction.

  1. self definitional [Section 3.1 'Theoretical Solvability'; Appendix B.2, Eqs. (31)-(33)]
    "Given that the transmission scene and the virtual reflection image usually reside at distinct depths (ZT≠ZR) and the camera trajectory includes non-trivial dynamics, the induced velocity fields differ in magnitude scaling over time. Consequently, the aggregated velocity vectors satisfy: ST≠k·SR,∀k∈R."

    Under the linear system in Eq. (22), the paper's own model restricts both layers to a shared direction d: uT=vTd, uR=vRd. If the camera translates with velocity V(t), then vT(t)=V(t)·d/ZT and vR(t)=V(t)·d/ZR, exactly as in Eq. (31)'s inverse-depth law. Therefore ST=(ZR/ZT)SR as vectors, so the columns of A are scalar multiples, rank(A)=1, det(H)=0, and the least-squares minimizer is not unique. The assertion 'ST≠k·SR' is thus not a consequence of ZT≠ZR; it is exactly the linear-independence condition needed to conclude positive definiteness. The theorem's conclusion is inserted as a premise.

full rationale

The empirical core of the paper is not circular: EvReflection is trained supervised on synthetic VOC pairs and tested on external SIR2 and the authors' own EVR2 benchmark; no constants are fitted to the reported benchmark numbers, and the same-group self-citations (Kai et al., Xiao et al.) are contextual rather than load-bearing. The circular step is localized to the well-posedness theorem advertised as the foundation ('Guided by this proof, we design EvReflection'). In Appendix B.2, the proof that distinct depths yield linearly independent velocity vectors is asserted under a model that actually makes the vectors proportional, so the advertised 'theoretically confirms ... guarantee a unique solution' reduces to the very independence it must prove. Because this theorem is central to the paper's strongest claim, the overall circularity score is 6 rather than 0-2; the supervised empirical comparison itself remains self-contained and externally benchmarked.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The central approach rests on a chain of modeling choices: linear layer superposition, per-layer brightness constancy, collinear motion, a simulation pipeline that manufactures parallax, and a proof step that asserts linear independence of the two motion histories. The last of these is not derived and is false under the paper's own depth model for translation. The network itself is a learned system with standard supervised losses, so the circularity burden is modest, but the 'theoretical guarantee' is built on an unproven assertion.

free parameters (4)
  • virtual_depth_ratio = dT ≠ dR, values not disclosed
    The parallax-aware simulator assigns distinct depths to transmission and reflection layers; this hand-chosen setup is what makes micro-dynamics visible in the synthetic events and is a core design choice, not derived from data.
  • temporal_bins_B = 20
    Event voxel grid uses B=20 bins (Section D), up from the usual 5, to preserve micro-dynamics. No ablation justifies this specific value.
  • glass_thicknesses = 3 mm, 5 mm, 8 mm
    Real benchmark is stratified by these chosen thicknesses; they define the task difficulty but are not derived from any theory.
  • camera_trajectory_C(t) = random smooth trajectory, not specified
    The simulator's random camera trajectory controls the parallax magnitude in all synthetic training and test data; its distribution is not reported, making it an uncontrolled experimental variable.
assumptions (6)
  • domain assumption Observed image is a linear superposition of transmission and reflection layers: I = T + R.
    Invoked in Section 3.1 and Appendix A; standard reflection model but assumes no nonlinear interaction, whereas the simulator later adds a nonlinear term Φ(T,R), an inconsistency.
  • domain assumption Brightness constancy and the Optical Flow Constraint Equation hold independently for each layer.
    Used to derive Eq. (1) and Appendix B.1; standard but fails at occlusions, specularities, and intensity threshold crossings.
  • ad hoc to paper The two layers' motion vectors share a common local direction d, with magnitudes vT, vR.
    Assumed in Section 3.1 before Eq. (2); restricts to collinear motion, which is not guaranteed by the 3D camera model in the simulator.
  • ad hoc to paper Distinct depths (ZT ≠ ZR) imply ST and SR are linearly independent.
    Appendix B.2, Eq. (31-33). This is false under pure translation with the paper's own u∝1/Z model, where the magnitudes differ by a constant ratio, making the columns linearly dependent.
  • domain assumption ESIM, fed with interpolated video, produces event streams faithful to real sensor dynamics.
    Used in Section 4.1 for all synthetic training and SIR2 evaluation; assumes the simulator's threshold model and interpolation are realistic, which is plausible but unvalidated against real sensor noise.
  • ad hoc to paper Static benchmark images (SIR2) can be evaluated by generating synthetic events for them.
    Implicit in Section 4.3; the paper never states how events are obtained for SIR2, yet the method requires them. The validity of the resulting evaluation is unestablished.

how reviews work

0 comments
Cite this review

Pith. "Pith review of EvReflection: Event-Driven Micro-Dynamics for Reflection Removal." pith.science (2026). https://pith.science/paper/NRZECM6S

@misc{pith2026260806184,
  author       = {Pith},
  title        = {Pith review of: EvReflection: Event-Driven Micro-Dynamics for Reflection Removal},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NRZECM6S}},
  note         = {Machine review of arXiv:2608.06184}
}
abstract

Despite remarkable progress in reflection removal, current methods primarily exploit static image priors from a single frame and still suffer from severe residual artifacts due to the inherent ambiguity between the reflection and transmission layers. In this paper, we propose leveraging event signals to break this ambiguity. By employing event cameras to capture micro-dynamics, we reveal the differential motion between these two layers. We thereby present a novel event-driven reflection removal network, EvReflection, that utilizes these dynamic cues for layer separation. Specifically, we design a Micro-Dynamics Decoupler to disentangle layer-specific motions from event streams as priors, which then guide a Parallax-Attention Rectifier to cleanly remove artifacts from the RGB image. Furthermore, to address data scarcity, we develop a parallax-aware simulation pipeline and construct the EVR$^2$ benchmark dataset, the first real-world dataset for this task. Extensive experiments demonstrate that EvReflection achieves state-of-the-art performance on both synthetic and real-world benchmarks, surpassing the best competing method by more than 1.6 dB and 1.2 dB in PSNR, respectively. The code, dataset, and pre-trained models are available at https://github.com/JiaxiaoWang/EvReflection.

Figures

Figures reproduced from arXiv: 2608.06184 by the authors.

Figure 1
Figure 1. Visual comparison in a highly reflective scene. Existing methods like DAI (Hu et al., 2026) and RDNet (Zhao et al., 2025) still leave visible artifacts due to the inherent ambiguity between layers. By leveraging event signals (d) to capture micro-dynamics, our method successfully removes reflections and recovers a clean image with a significant PSNR gain (+9.70 dB over DAI). Despite its importance, reflection remova… view at source ↗
Figure 2
Figure 2. Illustration of event-driven micro-dynamics via motion parallax. (a) Since the transmission and reflection layers reside at different depths, even subtle camera motion induces distinct pixel displacements between the two layers. (b) Event voxel slices reveal this phenomenon: the spatial offset ∆d between transmission (green) and reflection (red) layers differs across temporal bins. Dynamics Decoupler (MDD) to disent… view at source ↗
Figure 3
Figure 3. Overview of the EvReflection framework. Event streams are voxelized into a spatiotemporal representation. The Micro￾Dynamics Decoupler disentangles mixed event dynamics into layer-specific transmission and reflection priors. These priors are fed into the Parallax-Attention Rectifier, which leverages cross-modal attention to spatially modulate RGB features and remove reflection artifacts. 2026a;b). Fusing event signa… view at source ↗
Figures from the paper (3 more)
Figure 5
Figure 5. Figure 5: Real-world data acquisition setup for the EVR2 dataset. An Event-RGB hybrid camera is mounted on a motorized linear slide, which ensures smooth and repeatable horizontal translation to induce consistent motion parallax across all recorded scenes. cally accurate micro-d…
Figure 6
Figure 6. Figure 6: Qualitative comparison on the SIR2 benchmark (Wan et al., 2017). Competing methods leave obvious residual artifacts in the zoomed regions (red boxes), while our method cleanly removes reflections and restores sharp details consistent with the ground truth. (a) Input im…
Figure 7
Figure 7. Figure 7: Qualitative comparison on the real-world EVR2 benchmark. Competing methods fail to remove severe ghosting artifacts caused by glass thickness (red boxes), whereas our method cleanly eliminates these reflections and restores clear background details. et al., 2020). Desp…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

113 extracted references · 67 canonical work pages

  1. [1]

    Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

    Reversible decoupling network for single image reflection removal , author=. Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

  2. [2]

    Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

    Removing reflections from raw photos , author=. Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

  3. [3]

    IEEE Transactions on Circuits and Systems for Video Technology , year=

    Real-world image reflection removal: An ultra-high-definition dataset and an efficient baseline , author=. IEEE Transactions on Circuits and Systems for Video Technology , year=

  4. [4]

    IEEE International Conference on Robotics and Automation , year=

    Specular reflection removal on high-speed camera for robot vision , author=. IEEE International Conference on Robotics and Automation , year=

  5. [5]

    Proceedings of the IEEE International Conference on Computer Vision , pages=

    Video reflection removal through spatio-temporal optimization , author=. Proceedings of the IEEE International Conference on Computer Vision , pages=

  6. [6]

    IEEE Transactions on Multimedia , year=

    Rethinking depth guided reflection removal , author=. IEEE Transactions on Multimedia , year=

  7. [7]

    IEEE signal processing letters , year=

    Spatio-temporal Multi-image Reflection Removal , author=. IEEE signal processing letters , year=

  8. [8]

    Proceedings of the AAAI Conference on Artificial Intelligence , pages=

    Dereflection any image with diffusion priors and diversified data , author=. Proceedings of the AAAI Conference on Artificial Intelligence , pages=

Show all 113 references
  1. [9]

    IEEE Access , volume=

    Single-image reflection removal using deep learning: a systematic review , author=. IEEE Access , volume=. 2022 , publisher=

    2022

  2. [10]

    Recovering intrinsic scene characteristics , author=. Comput. vis. syst , volume=

  3. [11]

    Artificial intelligence , year=

    Determining optical flow , author=. Artificial intelligence , year=

  4. [12]

    IEEE transactions on neural networks and learning systems , year=

    Event-based visual flow , author=. IEEE transactions on neural networks and learning systems , year=

  5. [13]

    Proceedings IEEE Conference on Computer Vision and Pattern Recognition , year=

    Layer extraction from multiple images containing reflections and transparency , author=. Proceedings IEEE Conference on Computer Vision and Pattern Recognition , year=

  6. [14]

    IEEE transactions on pattern analysis and machine intelligence , year=

    Blind separation of superimposed moving images using image statistics , author=. IEEE transactions on pattern analysis and machine intelligence , year=

  7. [15]

    European Conference on Computer Vision , pages=

    Reflection backdoor: A natural backdoor attack on deep neural networks , author=. European Conference on Computer Vision , pages=. 2020 , organization=

    2020

  8. [16]

    Asian Conference on Machine Learning , pages=

    Robust Scene Text Detection via Learnable Scene Transformations , author=. Asian Conference on Machine Learning , pages=

  9. [17]

    International conference on machine learning , pages=

    Test-time training with self-supervision for generalization under distribution shifts , author=. International conference on machine learning , pages=. 2020 , organization=

    2020

  10. [18]

    IEEE Transactions on Multimedia , year=

    Ustc-td: A test dataset and benchmark for image and video coding in 2020s , author=. IEEE Transactions on Multimedia , year=

  11. [19]

    arXiv preprint arXiv:2107.00649 , year=

    On the practicality of deterministic epistemic uncertainty , author=. arXiv preprint arXiv:2107.00649 , year=

    arXiv

  12. [20]

    arXiv preprint arXiv:2212.06079 , year=

    Robust perception through equivariance , author=. arXiv preprint arXiv:2212.06079 , year=

    arXiv

  13. [21]

    European conference on computer vision , pages=

    Segmenting transparent objects in the wild , author=. European conference on computer vision , pages=. 2020 , organization=

    2020

  14. [22]

    ECCV 2024 Workshop on Wild 3D: 3D Modeling, Reconstruction, and Generation in the Wild , year=

    D3 RoMa: Disparity Diffusion-based Depth Sensing for Material-Agnostic Robotic Manipulation , author=. ECCV 2024 Workshop on Wild 3D: 3D Modeling, Reconstruction, and Generation in the Wild , year=

    2024

  15. [23]

    Conference on Robot Learning , pages=

    Simnet: Enabling robust unknown object manipulation from pure synthetic data via stereo , author=. Conference on Robot Learning , pages=. 2022 , organization=

    2022

  16. [24]

    arXiv preprint arXiv:2110.00087 , year=

    Seeing glass: joint point cloud and depth completion for transparent objects , author=. arXiv preprint arXiv:2110.00087 , year=

    arXiv

  17. [25]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Single image reflection separation via component synergy , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  18. [26]

    Applied Intelligence , volume=

    Two-stage single image reflection removal with reflection-aware guidance , author=. Applied Intelligence , volume=. 2023 , publisher=

    2023

  19. [27]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Language-guided image reflection separation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  20. [28]

    European Conference on Computer Vision , pages=

    L-differ: Single image reflection removal with language-based diffusion model , author=. European Conference on Computer Vision , pages=

  21. [29]

    Advances in Neural Information Processing Systems , volume=

    Trash or treasure? an interactive dual-stream strategy for single image reflection separation , author=. Advances in Neural Information Processing Systems , volume=

  22. [30]

    Advances in Neural Information Processing Systems , volume=

    Single image reflection separation via dual-stream interactive transformers , author=. Advances in Neural Information Processing Systems , volume=

  23. [31]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Revisiting single image reflection removal in the wild , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  24. [32]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Learning to see through obstructions , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  25. [33]

    IEEE transactions on pattern analysis and machine intelligence , volume=

    Learning to see through obstructions with layered decomposition , author=. IEEE transactions on pattern analysis and machine intelligence , volume=

  26. [34]

    Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision , pages=

    Burst reflection removal using reflection motion aggregation cues , author=. Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision , pages=

  27. [35]

    IEEE journal of solid-state circuits , volume=

    A 128 128 120 dB 15 s latency asynchronous temporal contrast vision sensor , author=. IEEE journal of solid-state circuits , volume=. 2008 , publisher=

    2008

  28. [36]

    Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

    Event-based Video Super-Resolution via State Space Models , author=. Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

  29. [37]

    IEEE/CAA Journal of Automatica Sinica , year=

    PromptFusion: Harmonized semantic prompt learning for infrared and visible image fusion , author=. IEEE/CAA Journal of Automatica Sinica , year=

  30. [38]

    Proceedings of the 41st International Conference on Machine Learning , pages =

    Learning Scale-Aware Spatio-temporal Implicit Representation for Event-based Motion Deblurring , author =. Proceedings of the 41st International Conference on Machine Learning , pages =

  31. [39]

    Proceedings of the 42nd International Conference on Machine Learning , pages =

    Hybrid Spiking Vision Transformer for Object Detection with Event Cameras , author =. Proceedings of the 42nd International Conference on Machine Learning , pages =

  32. [40]

    IEEE Transactions on Multimedia , year=

    Learning Dual Modality Interactions for Event-based Motion Deblurring , author=. IEEE Transactions on Multimedia , year=

  33. [41]

    European Conference on Computer Vision , pages=

    Contourlet residual for prompt learning enhanced infrared image super-resolution , author=. European Conference on Computer Vision , pages=

  34. [42]

    Forty-second International Conference on Machine Learning , year=

    EvFocus: Learning to Reconstruct Sharp Images from Out-of-Focus Event Streams , author=. Forty-second International Conference on Machine Learning , year=

  35. [43]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=

    User assisted separation of reflections from a single image using a sparsity prior , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=. 2007 , publisher=

    2007

  36. [44]

    Advances in Neural Information Processing Systems , volume=

    Learning to perceive transparency from the statistics of natural scenes , author=. Advances in Neural Information Processing Systems , volume=

  37. [45]

    Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2004

    Separating reflections from a single image using local features , author=. Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2004. CVPR 2004. , volume=. 2004 , organization=

    2004

  38. [46]

    2009 Workshop on Applications of Computer Vision (WACV) , pages=

    Interference reflection separation from a single image , author=. 2009 Workshop on Applications of Computer Vision (WACV) , pages=. 2009 , organization=

    2009

  39. [47]

    Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

    Single image layer separation using relative smoothness , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

  40. [48]

    Proceedings of the IEEE international conference on computer vision , pages=

    Benchmarking single-image reflection removal algorithms , author=. Proceedings of the IEEE international conference on computer vision , pages=

  41. [49]

    2013 IEEE International Symposium on Circuits and Systems (ISCAS) , pages=

    Separation of weak reflection from a single superimposed image using gradient profile sharpness , author=. 2013 IEEE International Symposium on Circuits and Systems (ISCAS) , pages=. 2013 , organization=

    2013

  42. [50]

    Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

    Reflection removal using ghosting cues , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

  43. [51]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Location-aware single image reflection removal , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  44. [52]

    Proceedings of the IEEE International Conference on Computer Vision , pages=

    A generic deep architecture for single image reflection removal and image smoothing , author=. Proceedings of the IEEE International Conference on Computer Vision , pages=

  45. [53]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Robust single image reflection removal against adversarial attacks , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  46. [54]

    arXiv preprint arXiv:2402.02374 , year=

    Promptrr: Diffusion models as prompt generators for single image reflection removal , author=. arXiv preprint arXiv:2402.02374 , year=

    arXiv

  47. [55]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Single image reflection removal exploiting misaligned training data and network enhancements , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  48. [56]

    Proceedings of the european conference on computer vision (ECCV) , pages=

    Seeing deeply and bidirectionally: A deep learning approach for single image reflection removal , author=. Proceedings of the european conference on computer vision (ECCV) , pages=

  49. [57]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Single image reflection removal through cascaded refinement , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  50. [58]

    International Journal of Computer Vision , volume=

    Separation of reflection components using color and polarization , author=. International Journal of Computer Vision , volume=. 1997 , publisher=

    1997

  51. [59]

    IEEE transactions on pattern analysis and machine intelligence , volume=

    A physically-based approach to reflection separation: from physical modeling to constrained optimization , author=. IEEE transactions on pattern analysis and machine intelligence , volume=. 2013 , publisher=

    2013

  52. [60]

    IEEE Transactions on Image Processing , volume=

    High-quality reflection separation using polarized images , author=. IEEE Transactions on Image Processing , volume=. 2011 , publisher=

    2011

  53. [61]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Polarized reflection removal with perfect alignment in the wild , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  54. [62]

    European Conference on Computer Vision , pages=

    Reflection separation via multi-bounce polarization state tracing , author=. European Conference on Computer Vision , pages=. 2020 , organization=

    2020

  55. [63]

    Advances in neural information processing systems , volume=

    Reflection separation using a pair of unpolarized and polarized images , author=. Advances in neural information processing systems , volume=

  56. [64]

    Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

    PolarFree: Polarization-based Reflection-Free Imaging , author=. Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

  57. [65]

    ACM SIGGRAPH 2005 Papers , pages=

    Removing photography artifacts using gradient projection and flash-exposure sampling , author=. ACM SIGGRAPH 2005 Papers , pages=

    2005

  58. [66]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Robust reflection removal with reflection-free flash-only cues , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  59. [67]

    International Journal of Computer Vision , volume=

    Light flickering guided reflection removal , author=. International Journal of Computer Vision , volume=. 2024 , publisher=

    2024

  60. [68]

    Proceedings of the IEEE international conference on computer vision , pages=

    Exploiting reflection change for automatic reflection removal , author=. Proceedings of the IEEE international conference on computer vision , pages=

  61. [69]

    Conference on Robot Learning , pages=

    Esim: an open event camera simulator , author=. Conference on Robot Learning , pages=

  62. [70]

    Tenth IEEE International Conference on Computer Vision (ICCV'05) Volume 1 , volume=

    Separating transparent layers of repetitive dynamic behaviors , author=. Tenth IEEE International Conference on Computer Vision (ICCV'05) Volume 1 , volume=. 2005 , organization=

    2005

  63. [71]

    ACM Transactions on Graphics (TOG) , volume=

    A computational approach for obstruction-free photography , author=. ACM Transactions on Graphics (TOG) , volume=. 2015 , publisher=

    2015

  64. [72]

    Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages=

    Robust optical flow estimation of double-layer images under transparency or reflection , author=. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages=

  65. [73]

    IEEE transactions on pattern analysis and machine intelligence , volume=

    Event-based vision: A survey , author=. IEEE transactions on pattern analysis and machine intelligence , volume=

  66. [74]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=

    EvTexture++: Event-Driven Texture Enhancement for Video Super-Resolution , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=

  67. [75]

    Kai, Dachun and Lu, Jiayao and Zhang, Yueyi and Sun, Xiaoyan , booktitle =

  68. [76]

    Proceedings of the AAAI Conference on Artificial Intelligence , pages=

    Event-Enhanced Blurry Video Super-Resolution , author=. Proceedings of the AAAI Conference on Artificial Intelligence , pages=

  69. [77]

    IEEE International Conference on Image Processing , pages=

    Video Super-Resolution Via Event-Driven Temporal Alignment , author=. IEEE International Conference on Image Processing , pages=

  70. [78]

    Proceedings of the AAAI Conference on Artificial Intelligence , pages=

    Seeing the Unseen: Zooming in the Dark with Event Cameras , author=. Proceedings of the AAAI Conference on Artificial Intelligence , pages=

  71. [79]

    Forty-first International Conference on Machine Learning , year=

    Evggs: A collaborative learning framework for event-based generalizable gaussian splatting , author=. Forty-first International Conference on Machine Learning , year=

  72. [80]

    International Conference on Learning Representations , year=

    Adam: A Method for Stochastic Optimization , author=. International Conference on Learning Representations , year=

  73. [81]

    Loshchilov, Ilya and Hutter, Frank , booktitle=

  74. [82]

    Proceedings of the 31st ACM International Conference on Multimedia , pages=

    Personalized single image reflection removal network through adaptive cascade refinement , author=. Proceedings of the 31st ACM International Conference on Multimedia , pages=

  75. [83]

    International Journal of Computer Vision , volume=

    The pascal visual object classes (voc) challenge , author=. International Journal of Computer Vision , volume=

  76. [84]

    Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

    Single image reflection separation with perceptual losses , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

  77. [85]

    IEEE Transactions on Image Processing , volume=

    Learning dense and continuous optical flow from an event camera , author=. IEEE Transactions on Image Processing , volume=. 2022 , publisher=

    2022

  78. [86]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Towards robust event-guided low-light image enhancement: a large-scale real-world event-image dataset and novel approach , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  79. [87]

    Proceedings of the AAAI Conference on Artificial Intelligence , pages=

    Deblursr: Event-based motion deblurring under the spiking representation , author=. Proceedings of the AAAI Conference on Artificial Intelligence , pages=

  80. [88]

    European conference on computer vision , pages=

    Event-based fusion for motion deblurring with cross-modal attention , author=. European conference on computer vision , pages=

  81. [89]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=

    Learning to see through with events , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=. 2022 , publisher=

    2022

  82. [90]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Event-based video reconstruction via potential-assisted spiking neural network , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  83. [91]

    arXiv preprint arXiv:2512.09016 , year=

    Learning to Remove Lens Flare in Event Camera , author=. arXiv preprint arXiv:2512.09016 , year=

  84. [92]

    Proceedings of the European conference on computer vision (ECCV) workshops , year=

    Esrgan: Enhanced super-resolution generative adversarial networks , author=. Proceedings of the European conference on computer vision (ECCV) workshops , year=

  85. [93]

    Advances in Neural Information Processing Systems , pages=

    Focal modulation networks , author=. Advances in Neural Information Processing Systems , pages=

  86. [94]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Unsupervised event-based learning of optical flow, depth, and egomotion , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  87. [95]

    2021 IEEE international conference on computational photography (ICCP) , pages=

    Eventgan: Leveraging large scale image datasets for event cameras , author=. 2021 IEEE international conference on computational photography (ICCP) , pages=. 2021 , organization=

    2021

  88. [96]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Evunroll: Neuromorphic events based rolling shutter image correction , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  89. [97]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Learning event guided high dynamic range video reconstruction , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  90. [98]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Eventps: Real-time photometric stereo using an event camera , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  91. [99]

    IEEE Transactions on pattern analysis and machine intelligence , volume=

    Hybrid high dynamic range imaging fusing neuromorphic and conventional images , author=. IEEE Transactions on pattern analysis and machine intelligence , volume=

  92. [100]

    European conference on computer vision , pages=

    Raft: Recurrent all-pairs field transforms for optical flow , author=. European conference on computer vision , pages=. 2020 , organization=

    2020

  93. [101]

    European conference on computer vision , pages=

    Simple baselines for image restoration , author=. European conference on computer vision , pages=. 2022 , organization=

    2022

  94. [102]

    European conference on computer vision , pages=

    Perceptual losses for real-time style transfer and super-resolution , author=. European conference on computer vision , pages=

  95. [103]

    Knowledge-Based Systems , volume=

    Polarization information restoration for visual reflection removal via cross dual-stream network , author=. Knowledge-Based Systems , volume=. 2026 , publisher=

    2026

  96. [104]

    European Conference on Computer Vision , pages=

    Event-adapted video super-resolution , author=. European Conference on Computer Vision , pages=

  97. [105]

    Proceedings of the AAAI Conference on Artificial Intelligence , pages=

    Evstvsr: Event guided space-time video super-resolution , author=. Proceedings of the AAAI Conference on Artificial Intelligence , pages=

  98. [106]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , year =

    Yan, Haojie and Chen, Zehao and Liu, Yan and Gu, Shi and Lin, Peng and Ma, De and Tang, Huajin and Zheng, Qian and Pan, Gang , title =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , year =

  99. [107]

    IEEE Transactions on Image Processing , volume=

    Learning frame-event fusion for motion deblurring , author=. IEEE Transactions on Image Processing , volume=

  100. [108]

    Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision , pages=

    Learned dual-view reflection removal , author=. Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision , pages=

  101. [109]

    Advances in Neural Information Processing Systems , year=

    Efficient Rectified Flow for Image Fusion , author=. Advances in Neural Information Processing Systems , year=

  102. [110]

    Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

    Difiisr: A diffusion model with gradient guidance for infrared image super-resolution , author=. Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

  103. [111]

    Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

    DCEvo: Discriminative Cross-Dimensional Evolutionary Learning for Infrared and Visible Image Fusion , author=. Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

  104. [112]

    arXiv preprint arXiv:1503.02531 , year=

    Distilling the knowledge in a neural network , author=. arXiv preprint arXiv:1503.02531 , year=

    arXiv

  105. [113]

    arXiv preprint arXiv:1704.04861 , year=

    Mobilenets: Efficient convolutional neural networks for mobile vision applications , author=. arXiv preprint arXiv:1704.04861 , year=

    arXiv

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.