Pith. sign in

REVIEW 5 major objections 6 minor 32 references

AW-GATCN: Adaptive Weighted Graph Attention Convolutional Network for Event Camera Data Joint Denoising and Object Recognition

T0 review · 5 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read Adaptive graph-based denoising prunes event-camera noise by maximizing degree-distribution variance and lifts recognition accuracy on four benchmarks.

desk verdict A plausible engineering paper whose variance-maximizing denoising rule is the load-bearing but least-supported piece; worth refereeing with requests for code and a better-specified threshold. read the letter →

arxiv 2505.11232 v2 pith:CDCOBN7M submitted 2025-05-16 cs.CV

classification cs.CV
keywords eventcamerasdatadenoisinggraphattentionnetworksobjectrecognitionadaptivethresholdingdegreedistributionvarianceneuromorphicvisionvoxelsegmentation
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

Event cameras emit sparse, asynchronous streams that mix object structure with large amounts of redundant and noisy events. The paper claims that this noise can be removed and the remaining events classified by building a graph per local voxel, weighting each edge by a combination of distance, velocity, angular difference, and polarity, and then pruning edges with an adaptive threshold chosen to maximize the variance of the normalized degree distribution. It reports accuracies of 83.77% on N-Caltech101, 76.79% on CIFAR10-DVS, 99.30% on MNIST-DVS, and 96.89% on N-CARS, beating previous graph-based approaches by up to 8.79% and improving recognition accuracy by up to 19.57% when the denoising step is switched on in ablations. A sympathetic reader would care because the method avoids fixed-radius neighbor rules and frame conversion, which are the usual weak points in event-based recognition.

What carries the argument

The load-bearing mechanism is the adaptive edge-pruning step: a multifactorial edge weight that combines Euclidean distance, velocity magnitude difference, angular difference, and polarity consistency, together with a per-voxel threshold $T$ set by maximizing the variance of the normalized degree distribution. The threshold is meant to keep the graph's structure most differentiated while discarding weakly connected noise. The same edge weights are then inverted inside a graph attention convolution, so the network's feature aggregation is steered by the same measure of event-point correlation that drives denoising.

What would settle it

Take a clean event stream with known labels, inject Poisson background events at several controlled rates, and check two things: whether the per-voxel threshold $T$ increases with the injection rate, and whether recognition accuracy with denoising stays above accuracy without denoising at every rate. If $T$ is flat or accuracy collapses at moderate injection rates, the variance-maximizing heuristic is not tracking noise.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central discovery is that the graph formulation itself can do the denoising, if the graph is built and pruned adaptively. Starting from an event stream $E=\{(x_k,y_k,t_k,p_k)\}$, the method segments events into density-balanced windows and then voxels, constructs a graph per voxel with edge weights $w_{ij}=\alpha D_{ij}+\beta\Delta v_{ij}+\gamma\theta_{ij}+\delta P_{ij}$, and removes edges whose weight exceeds a threshold $T$ chosen per voxel as $T=\arg\max_\delta \sigma^2_\delta$, where $\sigma^2_\delta$ is the variance of the normalized degree distribution at threshold $\delta$. The surviving graph feeds two graph-attention convolutional layers whose attention coefficients include $1/w_{ij}$, so strongly correlated pairs dominate feature aggregation. The paper reports that this joint denoising-and-recognition pipeline surpasses prior graph-based methods on all four tested benchmarks, with the largest gains on the noisiest datasets.

Load-bearing premise

The load-bearing premise is that the threshold $T$ maximizing the variance of the normalized degree distribution separates meaningful event structure from noise; if that heuristic does not track the actual noise level, denoising will discard signal or keep noise and the accuracy gains will not carry over to new datasets.

Editorial extensions

If this is right

  • If the reported results hold, event-based recognition can be improved without dense frame conversion, preserving the sparsity and temporal resolution of the raw stream.
  • The adaptive threshold makes denoising self-tuning per voxel, so the method should transfer across scene densities without re-tuning a global radius.
  • The ablation comparison shows the balanced multifactor weight (Comb 3) beats the distance-only configuration by 4.50 points on D-END and 6.26 points on N-CARS.
  • The denoising ablation attributes most of the gain to the noise filter: accuracy rises by 17.19% on D-END, 19.57% on N-CARS, and 12.2% on N-END when denoising is enabled.

Reading between the lines

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

  • The variance-maximizing threshold may serve as an unsupervised noise-level indicator: on a dataset with known injected noise, the chosen $T$ should rise with noise rate, which would let the same mechanism calibrate event-camera preprocessing without labels.
  • Although the paper evaluates offline recognition, the per-voxel threshold computation is local and cheap enough that a streaming or online variant is a natural next step for latency-sensitive applications such as drone navigation.
  • The fixed weight parameters $\alpha=0.7,\beta=0.1,\gamma=0.1,\delta=0.1$ are chosen by search on END and N-CARS; making them learnable end-to-end could transfer better to new event-camera domains, but that extension goes beyond what the paper tests.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 6 minor

Summary. This manuscript proposes AW-GATCN, a graph-based method for joint denoising and object recognition of event camera data. The pipeline consists of three main components: adaptive event-point segmentation based on normalized density and a square-root-law voxelization (Algorithm 1), an adaptive denoising procedure that builds weighted graphs using Euclidean distance, velocity magnitude difference, angular difference, and polarity consistency and then selects a threshold by maximizing the variance of the normalized degree distribution (Algorithm 2, Eq. (10)), and a graph attention convolutional network that uses the resulting edge weights in the attention computation (Eqs. (13)--(14)). The method is evaluated on N-Caltech101, CIFAR10-DVS, MNIST-DVS, and N-CARS, where it is reported to achieve 83.77%, 76.79%, 99.30%, and 96.89% accuracy, respectively, outperforming prior graph-based methods by up to 8.79%. An ablation study on D-END, N-END, and N-CARS compares four weight configurations and reports accuracy improvements of 12.2--19.57 percentage points from the proposed denoising step.

Significance. If the reported results hold, the accuracy gains on N-Caltech101 and CIFAR10-DVS over prior graph-based methods such as NVS and EvS are meaningful, and the idea of adapting the graph threshold from the degree distribution is a simple and potentially useful contribution. The claims are empirical and falsifiable: the four benchmark numbers can be checked by independent implementation. However, the paper provides no code and no variance estimates, and the central denoising claim is not directly evidenced. The paper's contribution would be strengthened substantially by a concrete specification of Algorithm 2, a direct noise-removal evaluation, and a clarification of how the hyperparameters were chosen relative to the reported test sets.

major comments (5)
  1. [Section III-B, Algorithm 2 and Eq. (10)] Algorithm 2 is under-specified because the loop 'for δ = 0 to t' has no discretization step. The edge weights in Eq. (6) are continuous, so there are infinitely many candidate values of δ between 0 and t, and the argmax in Eq. (10) is not uniquely defined by the text. The authors should specify whether δ is swept over the sorted unique edge weights, over a fixed grid, or over some other finite set. Without this, the denoising step and the reported results are not reproducible.
  2. [Section III-B, Eq. (6)] The edge weight in Eq. (6) sums four heterogeneous terms: Euclidean distance, velocity magnitude difference, angular difference, and a binary polarity term. These quantities have different units and ranges, yet no normalization or scaling is described. Consequently, the fixed coefficients α=0.7, β=0.1, γ=0.1, δ=0.1 are not meaningful unless each term is normalized to a common scale, and the threshold T in Algorithm 2 will depend on the arbitrary scales of the input coordinates and timestamps. The authors should specify how each term is normalized or otherwise made commensurable.
  3. [Section IV-C2 and Table III] Table III is presented as evidence of 'noise reduction performance' or 'denoising effectiveness,' but it only compares recognition accuracy with and without the denoising step. An accuracy improvement from removing edges is not a measurement of noise removal: the procedure could act as a generic graph sparsifier that helps the classifier even if it removes signal edges. To support the joint-denoising claim, the authors need a direct evaluation of noise-removal quality, for example by adding synthetic noise with known ground truth and reporting precision/recall of retained events, or by comparing against established event-denoising methods on noisy event streams.
  4. [Section IV-C1, Table II, and Table I] The weight parameters are tuned on N-CARS and the same N-CARS result is then reported as a main result. Table II shows that Comb 3 achieved 96.89% on N-CARS, and Table I reports 96.89% for AW-GATCN on N-CARS. If the parameter selection in Section IV-C1 used the N-CARS test data, then this is test-set tuning and the reported N-CARS accuracy is optimistically biased. The authors should clarify how N-CARS was split during parameter selection or restrict the hyperparameter choice to validation folds.
  5. [Section IV-A and Table I] The main comparison in Table I reports only single accuracy numbers for each method, despite stating that all models are trained with 5-fold cross-validation. Without standard deviations, confidence intervals, or the per-fold results, it is not possible to assess whether the reported gains, some of which are small (e.g., 99.3% vs. 99.4% for AsyncNet on MNIST-DVS), are statistically significant. The authors should report the mean and variance across folds for their own method, and ideally for the compared methods.
minor comments (6)
  1. [Algorithm 1, line 12] The computed Nvoxels is a real number (a square root), but it is used as a count of voxels. The manuscript should specify whether rounding, ceiling, or floor is applied before the clamping operation.
  2. [Section IV-A] The preprocessing constants Nmin, Cscale, Nmin_vox, and Nmax_vox appear in Algorithm 1 but their values are never reported. Without these values, the segmentation and the subsequent graph construction are not fully reproducible.
  3. [Section III-B, Eqs. (8)--(9)] Eq. (9) reuses the symbol Φδ for both the raw and the normalized degree distribution, which makes the normalization step confusing. A different symbol, for example Φ̂δ, should be used.
  4. [Section III-C, Eq. (13)] The attention coefficient uses 1/wij, but nothing in the text rules out wij = 0, for example when two events have identical coordinates and timestamps or when the polarity term is zero. The authors should state how division by zero is avoided.
  5. [Abstract and Section IV-C2] The abstract claims 'improving noise reduction performance by up to 19.57%,' but Table III reports improvement in recognition accuracy, not in noise reduction. The wording should be changed to 'improving recognition accuracy' or a direct noise-removal metric should be reported.
  6. [Section IV-A, dataset description] The description of CIFAR10-DVS states that it contains 'a sixth of the original CIFAR10 dataset, totaling 60,000 samples (6,000 per class),' which is internally inconsistent: one sixth of 60,000 is 10,000. The authors should correct the dataset statistics or clarify the version used.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported accuracies are empirical benchmark results, and the adaptive denoising heuristic is fully specified in the paper.

full rationale

The paper's central claim is an empirical accuracy comparison across four event-camera datasets. The method consists of (1) adaptive event segmentation based on density heuristics, (2) a multifactor edge-weighting rule with parameters tuned via cross-validation, (3) an adaptive threshold chosen by maximizing variance of the normalized degree distribution (Algorithm 2, Eq. 10), and (4) a graph attention network. None of these steps defines its output in terms of the recognition accuracy it is later compared against. The variance-maximizing threshold is a heuristic that is not derived from the classification objective, and the paper does not claim to 'predict' accuracy from the threshold by construction. The only self-citation is Ref. [26] (AGSF), which the paper cites as inspiration for the adaptive graph formulation; the actual algorithm is fully described in the text (Algorithm 2, Eq. 6-12) and its performance is measured experimentally. The weight parameters (α=0.7, β=0.1, γ=0.1, δ=0.1) are selected via a cross-validated ablation study (Section IV-C1), and the final numbers are reported on the same datasets; this is standard hyperparameter selection, not a fitted input masquerading as a prediction. The 'noise reduction performance' claim (Table III) uses recognition-accuracy improvement as a proxy, which is an evaluation-validity concern but not a circular-reasoning step: no equation reduces to another by construction, and the ablation genuinely compares with and without the denoising stage. Overall, the derivation chain is self-contained against the benchmarks, and no quoted passage exhibits a definitional equivalence or a load-bearing self-citation that forces the reported result.

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

The method introduces no new physical entities. It relies on several tuned constants and two heuristic rules: the square root law for voxelization and the variance-maximizing threshold for denoising. Independent evidence for these heuristics beyond the reported accuracy numbers is absent.

free parameters (4)
  • alpha, beta, gamma, delta = 0.7, 0.1, 0.1, 0.1
    Edge weight coefficients in Equation 6, selected by ablation on D-END, N-CARS, and N-END (Section IV-C1, Table II) rather than derived.
  • Nmin = not specified
    Minimum points per window in Algorithm 1, equation 2; value not given in the paper.
  • Cscale = not specified
    Scaling factor for window size in Equation 2; value not given.
  • Nmin_vox, Nmax_vox = not specified
    Clamps for voxel count in Algorithm 1; values not given.
assumptions (4)
  • ad hoc to paper Square root law for voxel count (Equation 4)
    The number of voxels is set to sqrt(X*Y*T) with no derivation; this is a heuristic choice that affects segmentation granularity.
  • ad hoc to paper Maximizing degree-distribution variance removes noise (Equation 10)
    The core denoising assumption that the threshold maximizing variance of normalized degree distribution separates meaningful structure from noise is asserted without proof or direct noise-label validation.
  • domain assumption Velocity vectors from neighboring events are meaningful (Equation 7)
    The method assumes local event neighbors define a reliable motion vector, but in noisy or asynchronous data this may not hold.
  • standard math MST sets a valid upper threshold for connectivity
    Using the minimum spanning tree to guarantee full connectivity is a standard algorithmic assumption.

how reviews work

0 comments
Cite this review

Pith. "Pith review of AW-GATCN: Adaptive Weighted Graph Attention Convolutional Network for Event Camera Data Joint Denoising and Object Recognition." pith.science (2026). https://pith.science/paper/CDCOBN7M

@misc{pith2026250511232,
  author       = {Pith},
  title        = {Pith review of: AW-GATCN: Adaptive Weighted Graph Attention Convolutional Network for Event Camera Data Joint Denoising and Object Recognition},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CDCOBN7M}},
  note         = {Machine review of arXiv:2505.11232}
}
read the original abstract

Event cameras, which capture brightness changes with high temporal resolution, inherently generate a significant amount of redundant and noisy data beyond essential object structures. The primary challenge in event-based object recognition lies in effectively removing this noise without losing critical spatial-temporal information. To address this, we propose an Adaptive Graph-based Noisy Data Removal framework for Event-based Object Recognition. Specifically, our approach integrates adaptive event segmentation based on normalized density analysis, a multifactorial edge-weighting mechanism, and adaptive graph-based denoising strategies. These innovations significantly enhance the integration of spatiotemporal information, effectively filtering noise while preserving critical structural features for robust recognition. Experimental evaluations on four challenging datasets demonstrate that our method achieves superior recognition accuracies of 83.77%, 76.79%, 99.30%, and 96.89%, surpassing existing graph-based methods by up to 8.79%, and improving noise reduction performance by up to 19.57%, with an additional accuracy gain of 6.26% compared to traditional Euclidean-based techniques.

Figures

Figures reproduced from arXiv: 2505.11232 by the authors.

Figure 1
Figure 1. The adaptive edge weight wij , designed to capture event point relevance, facilitates noise filtering and attention-based feature aggregation, enhancing robust recognition by emphasizing the most informative node connections. assessment aids in comprehending the temporal dynamics of the event stream, enhancing our understanding of its temporal characteristics. With both spatial and temporal ranges defined, the norma… view at source ↗
Figure 2
Figure 2. The circular region represents the optimal threshol [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Compared to Original Data, Comb 3 effectively reduce [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 9 canonical work pages

  1. [1]

    A 240×180 10mW 12us latency sparse-out put vision sensor for mobile applications,

    R. Berner et al., “A 240×180 10mW 12us latency sparse-out put vision sensor for mobile applications,” in 2013 Symposium on VLSI Circuits , IEEE, pp. C186–C187, 2013

  2. [2]

    EfficientNet: Rethinking Model Scal- ing for Convolutional Neural Networks,

    M. Tan and Q. V . Le, “EfficientNet: Rethinking Model Scal- ing for Convolutional Neural Networks,” arXiv, Cornell Uni ver- sity, 2019. [Online]. Available: https://arxiv.org/abs/ 1905.11946. doi: 10.48550/arxiv.1905.11946

  3. [3]

    End-to-End Learning of Representatio ns for Asynchronous Event-Based Data,

    D. Gehrig et al., “End-to-End Learning of Representatio ns for Asynchronous Event-Based Data,” arXiv, Cornell Universit y,

  4. [4]

    Events-to-Video: Bringing Modern Computer Vision to Event Cameras

    H. Rebecq et al., “Events-to-Video: Bringing Modern Com puter Vision to Event Cameras,” arXiv, Cornell University, 2019. [Onlin e]. Available: https://arxiv.org/abs/1904.08298. doi: 10.48550/arxiv.1904.08298

  5. [5]

    A Differentiable Recurrent Surface f or Asynchronous Event-Based Data,

    M. Cannici et al., “A Differentiable Recurrent Surface f or Asynchronous Event-Based Data,” arXiv, Cornell University, 2020. [Onli ne]. Available: https://arxiv.org/abs/2001.03455. doi: 10.48550/arxiv.2001.03455

  6. [6]

    HA TS: Histograms of Averaged Time Surf aces for Robust Event-based Object Classification,

    A. Sironi et al., “HA TS: Histograms of Averaged Time Surf aces for Robust Event-based Object Classification,” arXiv, Corn ell Univer- sity, 2018. [Online]. Available: https://arxiv.org/abs/ 1803.07913. doi: 10.48550/arxiv.1803.07913

  7. [7]

    HOTS: A Hierarchy of Event-Based Time -Surfaces for Pattern Recognition,

    X. Lagorce et al., “HOTS: A Hierarchy of Event-Based Time -Surfaces for Pattern Recognition,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 39, no. 7, pp. 1346–1359, 2017. doi: 10.1109/TPAMI.2016.2574707

  8. [8]

    HFirst: A Temporal Approach to Object Recogni- tion,

    G. Orchard et al., “HFirst: A Temporal Approach to Object Recogni- tion,” 2015. [Online]. Available: https://arxiv.org/abs /1508.01176. doi: 10.48550/arxiv.1508.01176

Show all 32 references
  1. [9]

    Effective AER Object Classification Using Segmented Probability-Maximization Learning in Spiking N eural Networks,

    Q. Liu et al., “Effective AER Object Classification Using Segmented Probability-Maximization Learning in Spiking N eural Networks,” arXiv, Cornell University, 2020. [Online]. Ava ilable: https://arxiv.org/abs/2002.06199. doi: 10.48550/arxiv.2002.06199

  2. [10]

    Learning Visual Motion Segmentat ion Using Event Surfaces,

    A. Mitrokhin et al., “Learning Visual Motion Segmentat ion Using Event Surfaces,” in 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , IEEE, pp. 14402–14411, 2020. doi: 10.1109/CVPR42600.2020.01442

  3. [11]

    Graph-Based Object Classification for Neu romor- phic Vision Sensing,

    Y . Bi et al., “Graph-Based Object Classification for Neu romor- phic Vision Sensing,” in 2019 IEEE/CVF International Confer- ence on Computer Vision (ICCV) , IEEE, pp. 491–501, 2019. doi: 10.1109/ICCV .2019.00058

  4. [12]

    Space-Time Event C louds for Gesture Recognition: From RGB Cameras to Event Cameras,

    Q. Wang, Y . Zhang, J. Y uan, and Y . Lu, “Space-Time Event C louds for Gesture Recognition: From RGB Cameras to Event Cameras, ” in 2019 IEEE Winter Conference on Applications of Computer Vision (WACV) , Waikoloa, HI, USA, pp. 1826–1835, 2019. doi: 10.1109/W ACV .2019.00199. ...

  5. [13]

    Interacting Maps for Fast Visual Interp retation,

    M. Cook et al., “Interacting Maps for Fast Visual Interp retation,” in The 2011 International Joint Conference on Neural Networks , IEEE, pp. 770–776, 2011. doi: 10.1109/IJCNN.2011.6033299

  6. [14]

    Inductive Rep - resentation Learning on Large Graphs,

    W. L. Hamilton, R. Ying, and J. Leskovec, “Inductive Rep - resentation Learning on Large Graphs,” arXiv, Cornell Univ er- sity, 2017. [Online]. Available: https://arxiv.org/abs/ 1706.02216. doi: 10.48550/arxiv.1706.02216

  7. [15]

    Graph Attention Networks,

    P . V eliˇ ckovi´ c et al., “Graph Attention Networks,” arXiv, 2017. [Online]. Available: https://arxiv.org/abs/1710.10903

  8. [16]

    Event-Stream Representation for Human Gaits Identi- fication Using Deep Neural Networks,

    Y . Wang et al., “Event-Stream Representation for Human Gaits Identi- fication Using Deep Neural Networks,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 44, no. 7, pp. 3436–3449, 2022. doi: 10.1109/TPAMI.2021.3054886

  9. [17]

    A Spiking Neural Network Architectu re for Visual Motion Estimation,

    G. Orchard et al., “A Spiking Neural Network Architectu re for Visual Motion Estimation,” in 2013 IEEE Biomedical Circuits and Systems Conference (BioCAS) , IEEE, pp. 298–301, 2013. doi: 10.1109/Bio- CAS.2013.6679698

  10. [18]

    Unsupervised Learning of Dense Optical Fl ow, Depth and Egomotion from Sparse Event Data,

    C. Y e et al., “Unsupervised Learning of Dense Optical Fl ow, Depth and Egomotion from Sparse Event Data,” arXiv, Cornell Unive r- sity, 2018. [Online]. Available: https://arxiv.org/abs/ 1809.08625. doi: 10.48550/arxiv.1809.08625

  11. [19]

    Event-based High Dynamic Range Im- age and V ery High Frame Rate Video Generation using Condi- tional Generative Adversarial Networks,

    S. M. M. I et al., “Event-based High Dynamic Range Im- age and V ery High Frame Rate Video Generation using Condi- tional Generative Adversarial Networks,” 2018. [Online]. Available: https://arxiv.org/abs/1811.08230. doi: 10.48550/arxiv.1811.08230

  12. [20]

    Event-based Asynchronous Spar se Convolu- tional Networks,

    N. Messikommer et al., “Event-based Asynchronous Spar se Convolu- tional Networks,” arXiv, Cornell University, 2020. [Onlin e]. Available: https://arxiv.org/abs/2003.09148. doi: 10.48550/arxiv.2003.09148

  13. [21]

    3D Semant ic Seg- mentation with Submanifold Sparse Convolutional Networks ,

    B. Graham, M. Engelcke, and L. van der Maaten, “3D Semant ic Seg- mentation with Submanifold Sparse Convolutional Networks ,” in 2018 IEEE/CVF Conference on Computer Vision and Pattern Recogni tion, IEEE, pp. 9224–9232, 2018. doi: 10.1109/CVPR.2018.00961

  14. [22]

    Generative PointNet: Deep Energy-Based Learning on Unordered Point Sets for 3D Generation, Reconstruction a nd Classification,

    J. Xie et al., “Generative PointNet: Deep Energy-Based Learning on Unordered Point Sets for 3D Generation, Reconstruction a nd Classification,” arXiv, Cornell University, 2020. [Online ]. Available: https://arxiv.org/abs/2004.01301. doi: 10.48550/arxiv.2004.01301

  15. [23]

    PointNet++: Deep Hierarchical Feature Learn- ing on Point Sets in a Metric Space,

    C. R. Qi et al., “PointNet++: Deep Hierarchical Feature Learn- ing on Point Sets in a Metric Space,” arXiv, Cornell Univer- sity, 2017. [Online]. Available: https://arxiv.org/abs/ 1706.02413. doi: 10.48550/arxiv.1706.02413

  16. [24]

    EventNet: Asynchro nous Recursive Event Processing,

    Y . Sekikawa, K. Hara, and H. Saito, “EventNet: Asynchro nous Recursive Event Processing,” in 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , IEEE, pp. 3882–3891, 2019. doi: 10.1109/CVPR.2019.00401

  17. [25]

    Event-Based Vision: A Survey,

    G. Gallego et al., “Event-Based Vision: A Survey,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 44, no. 1, pp. 154– 180, 2022. doi: 10.1109/TPAMI.2020.3008413

  18. [26]

    AGSF: Adaptive Graph Fo rmulation and Hand-Crafted Graph Spectral Features for Shape Represe ntation,

    B. Alwaely and C. Abhayaratne, “AGSF: Adaptive Graph Fo rmulation and Hand-Crafted Graph Spectral Features for Shape Represe ntation,” IEEE Access , vol. 8, pp. 182260–182272, 2020. doi: 10.1109/AC- CESS.2020.3028696

  19. [27]

    Converting Static Image Datasets to Spik- ing Neuromorphic Datasets Using Saccades,

    G. Orchard et al., “Converting Static Image Datasets to Spik- ing Neuromorphic Datasets Using Saccades,” arXiv, Cornell Univer- sity, 2015. [Online]. Available: https://arxiv.org/abs/ 1507.07629. doi: 10.48550/arxiv.1507.07629

  20. [28]

    CIFAR10-DVS: An Event-Stream Dataset for Object Classification,

    H. Li et al., “CIFAR10-DVS: An Event-Stream Dataset for Object Classification,” Frontiers in Neuroscience , vol. 11, MA Y , pp. 309–309,

  21. [29]

    Pok er-DVS and MNIST-DVS. Their History, How They Were Made, and Other De- tails,

    T. Serrano-Gotarredona and B. Linares-Barranco, “Pok er-DVS and MNIST-DVS. Their History, How They Were Made, and Other De- tails,” Frontiers in Neuroscience , vol. 9, pp. 481–481, 2015. doi: 10.3389/fnins.2015.00481

  22. [30]

    E-MLB: Multilevel Benchmark for Event- Based Camera Denoising,

    S. Ding et al., “E-MLB: Multilevel Benchmark for Event- Based Camera Denoising,” IEEE Transactions on Multimedia , vol. 26, pp. 1–12, 2024. doi: 10.1109/TMM.2023.3260638

  23. [2017]

    doi: 10.3389/fnins.2017.00309

  24. [2019]

    Available: https://arxiv.org/abs/1904

    [Online]. Available: https://arxiv.org/abs/1904. 08245. doi: 10.48550/arxiv.1904.08245

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.