Pith. sign in

REVIEW 4 major objections 4 minor 48 references

CREST: An Efficient Conjointly-trained Spike-driven Framework for Event-based Object Detection Exploiting Spatiotemporal Dynamics

T0 review · 4 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A conjointly-trained spike-driven framework called CREST claims state-of-the-art event-based object detection accuracy with up to 100x lower energy than prior spiking detectors.

desk verdict Real contributions in SNN training and event representation, but the 100x energy claim is a cost-model artifact that needs hardware validation. read the letter →

arxiv 2412.12525 v3 pith:Q3EORLPA submitted 2024-12-17 cs.CV cs.AI

Ruixin Mao , Aoyu Shen , Lin Tang , Jun Zhou This is my paper · ORCID
classification cs.CVcs.AI
keywords event-basedvisionspikingneuralnetworksobjectdetectionfew-spikesneuronsurrogategradientlearningenergyefficiencytemporalcodingneuromorphichardware
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

CREST claims that a fully spike-driven neural network can surpass prior spiking detectors on event-based object recognition and detection while reducing estimated inference energy by up to two orders of magnitude. The framework combines a conjoint learning rule, which trains through a surrogate network with discrete-level activations instead of unrolling gradients over spike times, with a multi-scale spatiotemporal event integrator and a spatiotemporal IoU loss. On the NCARs, Gen1, and PKU-Vidar-DVS benchmarks, the paper reports accuracy above the SNN baselines it compares against, with the largest gains coming from the new training rule and the event representation. If the energy model holds, CREST would make high-accuracy event-based detection practical on low-power neuromorphic hardware.

What carries the argument

The load-bearing mechanism is the conjoint learning rule. Its surrogate DL-Net defines activations by clipping the membrane potential to a bounded range and rounding to the nearest discrete level (Eqs. 9-10), then replaces the temporal accumulation over K timesteps with a single convolution (Eq. 11); backward passes approximate the round derivative as identity (Eq. 12), so error propagates only spatially. This removes the iterative spatiotemporal gradient of Eq. 4, cutting the time loop out of training and easing gradient vanishing. MESTOR supplies multi-scale spatiotemporal encoding: an FSN convolution layer accumulates fixed 3x3-filter responses over time bins and retains pixels whose spike count reaches K-1, yielding a spatiotemporal-continuous channel, combined with a spatial accumulation channel and a short-bin temporal channel. ST-IoU computes the absolute difference of spike density between predicted and ground-truth boxes and adds it to CIoU, so the loss is sensitive to the fact that objects produce denser spikes than background. FSN, the few-spikes neuron from prior work, provides the efficient temporal coding that makes layer outputs spike trains of K bits.

What would settle it

Measure the actual per-inference energy of CREST on an FSN-based neuromorphic processor (or a cycle-accurate simulator that includes membrane decay, memory access, and control logic) and compare it with a LIF-based SNN detector on the same task; if the measured ratio is far below 100x, the central energy claim is refuted. The accuracy claims can be checked independently by retraining the three components on the three public datasets.

Watch

Extended reading notes

Core claim

The core discovery is that the expensive spatiotemporal backpropagation of SNNs can be replaced by a spatial surrogate gradient computed in a discrete-level network (DL-Net), without sacrificing accuracy on event-based detection. During forward propagation, the DL-Net clips and rounds membrane potentials to mimic FSN spike trains; during backward propagation, the round function is treated as an identity, so gradients flow through a single spatial pass rather than across K timesteps. This conjoint learning rule is complemented by MESTOR, which converts raw event streams into three channels—spatiotemporal-continuous events, spatial accumulation, and temporal short-bin maps—and by ST-IoU, which adds a spike-density IoU term to the standard CIoU coordinate loss. The paper reports that CREST outperforms the compared SNN detectors on all three datasets: 0.952 accuracy on NCARs, 0.360 mAP50:95 and 0.632 mAP50 on Gen1, and 0.432 mAP50 on PKU-Vidar-DVS.

Load-bearing premise

The energy-efficiency claim rests on the assumption that FSN-based hardware incurs negligible membrane-decay cost, so only spike-integration accumulations consume energy; if real chips pay significant decay, memory-access, or control overhead, the reported up-to-100x savings would shrink.

Editorial extensions

If this is right

  • The conjoint learning rule reduces per-epoch training time by about 3.4x relative to FSN-BP on the studied settings, so directly-trained SNNs become feasible for larger detection models.
  • MESTOR's spatiotemporal-continuous filtering lowers firing rates while raising accuracy, meaning the same framework scales to longer event streams without proportional energy growth.
  • ST-IoU improves mAP over coordinate-only CIoU, showing that spike-density agreement is a reusable loss ingredient for event-based detectors.
  • Because the DL-Net forward pass and the FSN forward pass share weights, a model trained on GPU/TPU can be deployed on FSN hardware without re-training.

Reading between the lines

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

  • The paper's energy numbers are analytical rather than measured; if real FSN chips incur non-negligible decay or memory-access overhead, the efficiency lead over LIF-based SNNs could shrink, although the accuracy results would stand.
  • The same conjoint-learning trick of treating a rounded activation as an identity in the backward pass could transfer to other temporal coding schemes beyond FSN, such as rank or latency coding, where unrolling over time is equally expensive.
  • MESTOR's filtering of time-continuous events is a general denoising step for event streams and might be reused for tracking, segmentation, or other event-based tasks that suffer from background and noise events.
  • Training energy on GPU/TPU is not included in the efficiency comparison; a full life-cycle accounting could alter the headline, but inference-side savings remain the relevant metric for edge deployment.
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

4 major / 4 minor

Summary. CREST proposes a conjointly-trained spike-driven framework for event-based object detection and recognition. The method combines a surrogate DL-Net with discrete-level activations to approximate FSN-based spatiotemporal backpropagation, a multi-scale spatiotemporal event representation (MESTOR), and a spatiotemporal IoU loss (ST-IoU) that mixes a spiking-density term with CIoU. The paper reports state-of-the-art accuracy on NCARs, Gen1, and PKU-Vidar-DVS compared with prior SNN detectors, and claims up to 100x energy efficiency over SOTA SNNs on the basis of an analytical energy model.

Significance. If the accuracy results hold, CREST makes a useful contribution to event-based vision: it demonstrates that a surrogate-training scheme built around few-spikes neurons can train deep SNN detectors to competitive accuracy on three established benchmarks, and the ablations isolate the effects of the learning rule, the MESTOR representation, and the ST-IoU loss. The release of code and the use of multiple backbones and datasets are strengths. The central caveat is that the headline energy-efficiency claim is not a measurement: it follows from an analytical model that assigns full MAC costs to LIF decay and zero decay cost to FSN neurons, so the 100x number is contingent on hardware assumptions that are not validated in this manuscript. The accuracy results themselves are credible and do not depend on that assumption.

major comments (4)
  1. [Experimental Settings, Eqs. (24)-(26), and Table 2] The 'up to 100x energy efficiency' claim is computed, not measured. Eq. (25) charges LIF neurons a 4.6 pJ MAC for every non-spiking neuron-time step, while Eq. (26) assigns FSN neurons zero decay cost, citing the authors' STELLAR hardware paper (Mao et al. 2024). This asymmetry drives most of the reported gap: on Gen1, CREST (DenseNet121-16) is listed at 3.48 mJ versus EMS-34 at 650.13 mJ, and the (1−fr) decay term in Eq. (25) contributes roughly 624 mJ of that total. If a real FSN implementation incurs non-negligible static, memory, or control energy, or if a LIF implementation realizes decay with a cheaper operation than a full 4.6 pJ MAC, the advertised multiple shrinks substantially. Please either report measured power on FSN hardware, or provide a sensitivity analysis that treats the LIF decay cost and FSN overhead as parameters; without that, the energy claim should be reworded as conditional on the STELLAR hardware model.
  2. [Eq. (23) and Figure 5(b)] The final values of the ST-IoU weights a and b are never reported. The text and figure show experiments with (a,b) = (0.5,0.5), (0.3,0.7), and (0.1,0.9), but it is not stated which pair is used in Tables 1-3. Since ST-IoU is one of the two novel components claimed to improve detection accuracy, the missing values prevent reproduction and a fair comparison. Please report the exact weights used for each reported result.
  3. [Eq. (15) and Eq. (17)] The MESTOR spatiotemporal-continuity stage uses 'fixed-value filters of size 3×3' in Eq. (15), but the filter values are never specified. These values directly determine which pixels are retained through the threshold condition in Eq. (17), and they are essential for reproducing MESTOR. Please provide the exact filter kernel (or state that it is random and fixed) in the main text or supplementary material.
  4. [Tables 1-6] All experimental results are single-run, with no error bars or multiple-seed reporting. Several accuracy gaps over prior work are modest (e.g., 0.952 vs 0.937 on NCARs in Table 5, and 0.360 vs 0.321 mAP50:95 on Gen1 in Table 2). Without variance estimates, the claim of 'superior recognition & detection performance' is not statistically supported. Please report mean ± std over at least three seeds for the main comparisons and for the key ablations.
minor comments (4)
  1. [Eq. (22)] Spiking-IoU is defined as an absolute difference of spike densities, not as an intersection-over-union; the name is misleading and should be clarified in the text.
  2. [Table 5] The header 'SFOD w/ FSN-BP(ours)' is ambiguous: it could mean the original SFOD method or the authors' reimplementation with FSN-BP. Please label the column more explicitly.
  3. [Conclusion and Supplementary D] There are several typos ('effcient' in the Conclusion, 'choosed' and 'Futhermore' in the Supplementary Material) that should be corrected.
  4. [Figure 5(a)] The figure reports energy and mAP for different time windows, but it does not indicate which K corresponds to the final default setting (K=5) of the experiments; adding a marker for the reported configuration would improve clarity.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation: accuracy is externally benchmarked; the energy-efficiency estimate rests on a self-cited FSN hardware assumption that is a modeling caveat, not a logical reduction.

full rationale

I find no circular step in the derivation chain. The recognition and detection results in Tables 1-3, 5-6 are obtained on external benchmarks with publicly available datasets, and the ablations compare defined variants (FSN-BP vs CLR, MESTOR components, IoU losses) rather than re-reporting fitted quantities. The conjoint learning rule is an approximation of the FSN forward model whose accuracy and 3.4x training speedup are measured, not derived from the claim. MESTOR and ST-IoU are new modules validated by ablations. The only self-referential element is the energy model: Eq. 26 credits the authors' STELLAR hardware paper (Mao et al. 2024) for the claim that FSN-based SNNs have negligible membrane-decay energy, while Eq. 25 charges LIF SNNs a full 4.6 pJ MAC for every non-spiking step. This asymmetry largely produces the reported up-to-100x ratio, and no measured power data for CREST are given in this manuscript. That is an evidentiary limitation requiring a hardware-aware sensitivity check, but it is not circular: the assumption is consistent with the FSN dynamics in Eqs. 1-3, is attributed to an external peer-reviewed hardware study, and is not fitted from the benchmark results. The central accuracy claim is self-contained against external baselines.

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

The central empirical claims rest on standard benchmark evaluations and ablations. The energy claim rests on an analytical model with assumptions inherited from the authors' STELLAR paper. The training method relies on a surrogate gradient approximation that is standard in practice but not formally justified.

free parameters (4)
  • Time window K = 5 (default; swept across 4,5,6,7)
    Chosen by hand to trade accuracy against firing rate and energy; used in FSN-SNN and MESTOR threshold (Eq. 17).
  • Neuron threshold alpha = 3 for middle layers, 1 for input encoding layer
    Controls the ReLU emulation range in FSN/DL-Net; set as a hyperparameter.
  • ST-IoU weights a and b = Not explicitly stated in final model; grid searched over (0.5,0.5), (0.3,0.7), (0.1,0.9) in Fig. 5b
    Weights balancing Spiking-IoU and CIoU in Eq. 23; selected to maximize mAP on validation.
  • FSN ReLU emulation parameters = U_th(t)=d(t)=alpha * 2^-t
    A chosen setting to make FSN emulate ReLU (Eq. 8 context), adopted from prior work (Stockl and Maass 2021; Mao et al. 2024).
assumptions (4)
  • domain assumption The derivative of the round function in DL-Net is approximated as identity (Eq. 12), so gradients flow through the surrogate as if the discretization were a smooth linear map.
    This straight-through-estimator assumption is what makes CLR a valid surrogate gradient; it is not proven, only empirically supported by ablation.
  • domain assumption Events from the same object are spatiotemporally continuous whereas background and noise are not, so a pixel that fires at least K-1 times over K bins is treated as object-related (Eqs. 14-17).
    This heuristic underpins MESTOR's spatiotemporal filtering and is not derived from data statistics.
  • domain assumption FSN-based SNN inference energy is only K * OP_AC * fr * E_AC (Eq. 26), with negligible decay and no MAC cost during inference.
    The energy model assumes the STELLAR FSN hardware's 'negligible decay' claim; this is not measured in the paper.
  • domain assumption The FSN neuron with U_th(t)=d(t)=alpha*2^-t emulates a bounded ReLU, allowing DL-Net's clip+round to mimic the spiking behavior.
    The correspondence is approximate, with rounding error decreasing as K increases; the paper relies on this approximation for weight sharing.

how reviews work

0 comments
Cite this review

Pith. "Pith review of CREST: An Efficient Conjointly-trained Spike-driven Framework for Event-based Object Detection Exploiting Spatiotemporal Dynamics." pith.science (2026). https://pith.science/paper/Q3EORLPA

@misc{pith2026241212525,
  author       = {Pith},
  title        = {Pith review of: CREST: An Efficient Conjointly-trained Spike-driven Framework for Event-based Object Detection Exploiting Spatiotemporal Dynamics},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Q3EORLPA}},
  note         = {Machine review of arXiv:2412.12525}
}
read the original abstract

Event-based cameras feature high temporal resolution, wide dynamic range, and low power consumption, which is ideal for high-speed and low-light object detection. Spiking neural networks (SNNs) are promising for event-based object recognition and detection due to their spiking nature but lack efficient training methods, leading to gradient vanishing and high computational complexity, especially in deep SNNs. Additionally, existing SNN frameworks often fail to effectively handle multi-scale spatiotemporal features, leading to increased data redundancy and reduced accuracy. To address these issues, we propose CREST, a novel conjointly-trained spike-driven framework to exploit spatiotemporal dynamics in event-based object detection. We introduce the conjoint learning rule to accelerate SNN learning and alleviate gradient vanishing. It also supports dual operation modes for efficient and flexible implementation on different hardware types. Additionally, CREST features a fully spike-driven framework with a multi-scale spatiotemporal event integrator (MESTOR) and a spatiotemporal-IoU (ST-IoU) loss. Our approach achieves superior object recognition & detection performance and up to 100X energy efficiency compared with state-of-the-art SNN algorithms on three datasets, providing an efficient solution for event-based object detection algorithms suitable for SNN hardware implementation.

Figures

Figures reproduced from arXiv: 2412.12525 by the authors.

Figure 1
Figure 1. Comparison of (a) FSN-conversion, (b) FSN-BP, and (c) FSN-conjointly learning. The last layer uses the accumulated [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. The overview architecture of CREST. under the distribution of 𝑈 𝑙−1 𝑚 (Eq. 8), with 𝐾 𝑙−1 pre-set by the requirements of accuracy and firing rate. min 𝜀(𝑈 𝑙−1 𝑡 ℎ (𝑡), 𝑑𝑙−1 (𝑡); 𝐾 𝑙−1 ) = E𝑢𝑚 (𝑡) |𝑈 𝑙−1 𝑚 − 𝑈 𝑙−1 𝑝 | (8) A possible setting is 𝑈 𝑙−1 𝑡 ℎ (𝑡)= 𝑑 𝑙−1 (𝑡)= 𝛼 𝑙−12 −𝑡 . For case ❶, when 𝑈 𝑙−1 𝑚 ≤ 0, FSN 𝑚 generates no spikes (𝑈 𝑙−1 𝑝 = 0). For case ❷, when 𝑈 𝑙−1 𝑚 is an integer multiple of 𝛼 𝑙−12 −𝐾𝑙−1 , 𝑈… view at source ↗
Figure 3
Figure 3. Gen1/PKU-Vidar-DVS encoded with MESTOR. Method Acc SFOD w/ FSN-BP(ours) w/ CLR(ours) DenseNet121-16 0.937 0.948 0.952 DenseNet121-32 0.923 0.948 0.953 DenseNet169-16 0.921 0.946 0.950 DenseNet169-32 0.894 0.944 0.953 [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (1 more)
Figure 5
Figure 5. Figure 5: (a) Energy and mAP change with time-window [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 26 canonical work pages

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Abbott, L. F. 1999. Lapicque’s introduction of the integrate-and-fire model neuron (1907). Brain research bulletin, 50(5-6): 303--304

    1907

  4. [4]

    Bochkovskiy, A.; Wang, C.-Y.; and Liao, H.-Y. M. 2020. Yolov4: Optimal speed and accuracy of object detection. arXiv preprint arXiv:2004.10934

    arXiv 2020

  5. [5]

    B.; Kreuzberg, L.; and Behnke, S

    Bodden, L.; Schwaiger, F.; Ha, D. B.; Kreuzberg, L.; and Behnke, S. 2024. Spiking CenterNet: A Distillation-boosted Spiking Neural Network for Object Detection. In International Joint Conference on Neural Networks (IJCNN)

    work page 2024

  6. [6]

    Cannici, M.; Ciccone, M.; Romanoni, A.; and Matteucci, M. 2019. Asynchronous convolutional networks for object detection in neuromorphic cameras. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops

    work page 2019

  7. [7]

    Chen, N. F. 2018. Pseudo-labels for supervised learning on dynamic vision sensor data, applied to object detection under ego-motion. In Proceedings of the IEEE conference on computer vision and pattern recognition workshops, 644--653

    work page 2018

  8. [8]

    Cordone, L.; Miramond, B.; and Thierion, P. 2022. Object detection with spiking neural networks on automotive event data. In 2022 International Joint Conference on Neural Networks (IJCNN), 1--8. IEEE

    2022

Show all 48 references
  1. [9]

    Dampfhoffer, M.; Mesquida, T.; Valentian, A.; and Anghel, L. 2022. Are SNNs really more energy-efficient than ANNs An in-depth hardware-aware study. IEEE Transactions on Emerging Topics in Computational Intelligence, 7(3): 731--741

    2022

  2. [10]

    de Tournemire, P.; Nitti, D.; Perot, E.; Migliore, D.; and Sironi, A. 2020. A Large Scale Event-based Detection Dataset for Automotive. arXiv:2001.08499

    2020 arXiv

  3. [11]

    Fan, Y.; Zhang, W.; Liu, C.; Li, M.; and Lu, W. 2024. SFOD: Spiking Fusion Object Detector. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 17191--17200

    2024

  4. [12]

    Fang, W.; Yu, Z.; Chen, Y.; Masquelier, T.; Huang, T.; and Tian, Y. 2021. Incorporating learnable membrane time constant to enhance learning of spiking neural networks. In Proceedings of the IEEE/CVF international conference on computer vision, 2661--2671

    2021

  5. [13]

    J.; Conradt, J.; Daniilidis, K.; et al

    Gallego, G.; Delbr \"u ck, T.; Orchard, G.; Bartolozzi, C.; Taba, B.; Censi, A.; Leutenegger, S.; Davison, A. J.; Conradt, J.; Daniilidis, K.; et al. 2020. Event-based vision: A survey. IEEE transactions on pattern analysis and machine intelligence, 44(1): 154--180

    2020

  6. [14]

    Gerstner, W.; and Kistler, W. M. 2002. Spiking neuron models: Single neurons, populations, plasticity. Cambridge university press

    2002

  7. [15]

    Glorot, X.; Bordes, A.; and Bengio, Y. 2011. Deep sparse rectifier neural networks. In Proceedings of the fourteenth international conference on artificial intelligence and statistics, 315--323. JMLR Workshop and Conference Proceedings

    2011

  8. [16]

    Hasssan, A.; Meng, J.; and sun Seo, J. 2023. LT - SNN : Self-Adaptive Spiking Neural Network for Event-based Classification and Object Detection

    2023

  9. [17]

    Horowitz, M. 2014. 1.1 computing's energy problem (and what we can do about it). In 2014 IEEE international solid-state circuits conference digest of technical papers (ISSCC), 10--14. IEEE

    2014

  10. [18]

    Hu, Y.; Delbruck, T.; and Liu, S.-C. 2020. Learning to exploit multiple vision modalities by using grafted networks. In European Conference on Computer Vision, 85--101. Springer

    2020

  11. [19]

    Huang, G.; Liu, Z.; Van Der Maaten, L.; and Weinberger, K. Q. 2017. Densely connected convolutional networks. In Proceedings of the IEEE conference on computer vision and pattern recognition, 4700--4708

    2017

  12. [20]

    Iacono, M.; Weber, S.; Glover, A.; and Bartolozzi, C. 2018. Towards event-driven object detection with off-the-shelf deep learning. In 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 1--9. IEEE

    2018

  13. [21]

    P.; Young, C.; Patil, N.; Patterson, D.; Agrawal, G.; Bajwa, R.; Bates, S.; Bhatia, S.; Boden, N.; Borchers, A.; et al

    Jouppi, N. P.; Young, C.; Patil, N.; Patterson, D.; Agrawal, G.; Bajwa, R.; Bates, S.; Bhatia, S.; Boden, N.; Borchers, A.; et al. 2017. In-datacenter performance analysis of a tensor processing unit. In Proceedings of the 44th annual international symposium on computer archit...

    2017

  14. [22]

    Kim, S.; Park, S.; Na, B.; and Yoon, S. 2020. Spiking-yolo: spiking neural network for energy-efficient object detection. In Proceedings of the AAAI conference on artificial intelligence, volume 34, 11270--11277

    2020

  15. [23]

    Kugele, A.; Pfeil, T.; Pfeiffer, M.; and Chicca, E. 2021. Hybrid SNN-ANN: Energy-efficient classification and object detection for event-based vision. In DAGM German Conference on Pattern Recognition, 297--312. Springer

    2021

  16. [24]

    Li, J.; Wang, X.; Zhu, L.; Li, J.; Huang, T.; and Tian, Y. 2022 a . Retinomorphic object detection in asynchronous visual streams. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 36, 1332--1340

    2022

  17. [25]

    Li, Y.; He, X.; Dong, Y.; Kong, Q.; and Zeng, Y. 2022 b . Spike calibration: Fast and accurate conversion of spiking neural network for object detection and segmentation. arXiv preprint arXiv:2207.02702

    2022 arXiv

  18. [26]

    Li, Y.; Zhou, H.; Yang, B.; Zhang, Y.; Cui, Z.; Bao, H.; and Zhang, G. 2021. Graph-based asynchronous event processing for rapid object recognition. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 934--943

    2021

  19. [27]

    Lichtsteiner, P.; Posch, C.; and Delbruck, T. 2008. A 128 128 120 dB 15 s latency asynchronous temporal contrast vision sensor. IEEE journal of solid-state circuits, 43(2): 566--576

    2008

  20. [28]

    Ma, N.; Zhang, X.; Zheng, H.-T.; and Sun, J. 2018. Shufflenet v2: Practical guidelines for efficient cnn architecture design. In Proceedings of the European conference on computer vision (ECCV), 116--131

    2018

  21. [29]

    L.; Hannun, A

    Maas, A. L.; Hannun, A. Y.; Ng, A. Y.; et al. 2013. Rectifier nonlinearities improve neural network acoustic models. In Proc. icml, volume 30, 3. Atlanta, GA

    2013

  22. [30]

    Maass, W. 1997. Networks of spiking neurons: the third generation of neural network models. Neural networks, 10(9): 1659--1671

    1997

  23. [31]

    Mao, R.; Tang, L.; Yuan, X.; Liu, Y.; and Zhou, J. 2024. Stellar: Energy-Efficient and Low-Latency SNN Algorithm and Hardware Co-Design with Spatiotemporal Computation. In 2024 IEEE International Symposium on High-Performance Computer Architecture (HPCA), 172--185. IEEE

    2024

  24. [32]

    Messikommer, N.; Gehrig, D.; Loquercio, A.; and Scaramuzza, D. 2020. Event-based asynchronous sparse convolutional networks. In Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part VIII 16, 415--431. Springer

    2020

  25. [33]

    M.; and Roy, K

    Nagaraj, M.; Liyanagedera, C. M.; and Roy, K. 2023. Dotie-detecting objects through temporal isolation of events using a spiking architecture. In 2023 IEEE International Conference on Robotics and Automation (ICRA), 4858--4864. IEEE

    2023

  26. [34]

    O.; Mostafa, H.; and Zenke, F

    Neftci, E. O.; Mostafa, H.; and Zenke, F. 2019. Surrogate gradient learning in spiking neural networks: Bringing the power of gradient-based optimization to spiking neural networks. IEEE Signal Processing Magazine, 36(6): 51--63

    2019

  27. [35]

    Nickolls, J.; Buck, I.; Garland, M.; and Skadron, K. 2008. Scalable parallel programming with cuda: Is cuda the parallel programming model that application developers have been waiting for? Queue, 6(2): 40--53

    2008

  28. [36]

    Peng, Y.; Zhang, Y.; Xiao, P.; Sun, X.; and Wu, F. 2023. Better and faster: Adaptive event conversion for event-based object detection. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, 2056--2064

    2023

  29. [37]

    Rezatofighi, H.; Tsoi, N.; Gwak, J.; Sadeghian, A.; Reid, I.; and Savarese, S. 2019. Generalized intersection over union: A metric and a loss for bounding box regression. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 658--666

    2019

  30. [38]

    Roy, K.; Jaiswal, A.; and Panda, P. 2019. Towards spike-based machine intelligence with neuromorphic computing. Nature, 575(7784): 607--617

    2019

  31. [39]

    Sironi, A.; Brambilla, M.; Bourdis, N.; Lagorce, X.; and Benosman, R. 2018. HATS: Histograms of averaged time surfaces for robust event-based object classification. In Proceedings of the IEEE conference on computer vision and pattern recognition, 1731--1740

    2018

  32. [40]

    St \"o ckl, C.; and Maass, W. 2021. Optimized spiking neurons can classify images with high accuracy through temporal coding with two spikes. Nature Machine Intelligence, 3(3): 230--238

    2021

  33. [41]

    Su, Q.; Chou, Y.; Hu, Y.; Li, J.; Mei, S.; Zhang, Z.; and Li, G. 2023. Deep directly-trained spiking neural networks for object detection. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 6555--6565

    2023

  34. [42]

    Wu, Y.; Deng, L.; Li, G.; Zhu, J.; and Shi, L. 2018. Spatio-temporal backpropagation for training high-performance spiking neural networks. Frontiers in neuroscience, 12: 331

    2018

  35. [43]

    Yu, J.; Jiang, Y.; Wang, Z.; Cao, Z.; and Huang, T. 2016. Unitbox: An advanced object detection network. In Proceedings of the 24th ACM international conference on Multimedia, 516--520

    2016

  36. [44]

    Yuan, M.; Zhang, C.; Wang, Z.; Liu, H.; Pan, G.; and Tang, H. 2024. Trainable Spiking-YOLO for low-latency and high-performance object detection. Neural Networks, 172: 106092

    2024

  37. [45]

    Zhang, H.; Li, Y.; Leng, L.; Che, K.; Liu, Q.; Guo, Q.; Liao, J.; and Cheng, R. 2024. Automotive Object Detection via Learning Sparse Events by Spiking Neurons. IEEE Transactions on Cognitive and Developmental Systems

    2024

  38. [46]

    Zheng, Z.; Wang, P.; Ren, D.; Liu, W.; Ye, R.; Hu, Q.; and Zuo, W. 2021. Enhancing geometric factors in model learning and inference for object detection and instance segmentation. IEEE transactions on cybernetics, 52(8): 8574--8586

    2021

  39. [47]

    Zubi \'c , N.; Gehrig, D.; Gehrig, M.; and Scaramuzza, D. 2023. From chaos comes order: Ordering event representations for object recognition and detection. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 12846--12856

    2023

  40. [48]

    Zubic, N.; Gehrig, M.; and Scaramuzza, D. 2024. State space models for event cameras. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 5819--5828

    2024

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.