Pith. sign in

REVIEW 5 major objections 5 minor 1 cited by

When Every Millisecond Counts: Real-Time Anomaly Detection via the Multimodal Asynchronous Hybrid Network

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

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

Pith's one-line read A hybrid network fusing event-camera streams with RGB frames detects driving anomalies at millisecond latency while holding accuracy at or near the top of published baselines.

desk verdict A plausible multimodal fusion system with a useful new latency metric, but the paper's central 'millisecond-level' real-time claim is contradicted by its own mResponse numbers, which are over a second. read the letter →

arxiv 2506.17457 v1 pith:LUSYNCTP submitted 2025-06-20 cs.CV

Dong Xiao , Guangyao Chen , Peixi Peng , Yangru Huang , Yifan Zhao , Yongxing Dai , Yonghong Tian This is my paper · ORCID
classification cs.CV
keywords real-timeanomalydetectionautonomousdrivingeventcamerasasynchronousgraphneuralnetworkmultimodalfusionresponsetimev2esyntheticeventstrafficaccident
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

This paper tries to establish that traffic anomaly detection for autonomous driving can be accurate and genuinely fast at the same time if event-camera streams are added to ordinary RGB video. The proposed multimodal asynchronous hybrid network routes the event stream through an asynchronous graph neural network and the RGB frames through a ResNet, fuses the two, and runs the fused object-level features through recurrent units; on the ROL and DoTA benchmarks it reports detection at roughly 579 frames per second with the lowest response delays among compared methods while matching or beating them on most accuracy metrics. The reason this matters is that an autonomous vehicle needs a warning in the milliseconds before a pedestrian darts out or a vehicle cuts in, and the event stream supplies motion cues between video frames that a frame-based detector cannot see. The paper also introduces a new latency metric, mResponse, that averages detection delay across many operating thresholds. Because no real event-camera dataset of driving anomalies exists, all event data are synthesized from ordinary video with the v2e converter, an assumption the authors test only partially.

What carries the argument

The load-bearing object is the asynchronous event graph. Every brightness-change event becomes a node carrying normalized pixel coordinates and a scaled timestamp; edges link nodes within a spatial and temporal radius (up to 16 neighbors), and edge features encode relative spatial offsets. A deep asynchronous graph network with spline convolution aggregates these nodes, with RGB features from the ResNet branch sampled at each node's location and concatenated in, and the fusion is deliberately unidirectional so the GNN gains appearance context without making the CNN wait for event processing. Because events carry sub-frame timestamps, this graph can produce motion features between video frames, which is the mechanism behind the paper's inter-frame anomaly detection and its low response latency; downstream, GRUs accumulate object-level temporal evidence and an attention mechanism weights anomalous objects.

What would settle it

Mount a real event camera alongside an RGB camera on a vehicle, drive or replay the ROL/DoTA-type sudden-appearance scenarios, and compare anomaly delay (mResponse and mTTA) and accuracy against the RGB-only baseline and against the same model fed v2e-converted video. If real event streams do not reproduce the inter-frame detection advantage, for instance if mResponse on real events is no better than frame-only processing or accuracy drops because the network trained on synthetic events does not generalize, then the central claim about asynchronous event processing fails.

Watch

Extended reading notes

Core claim

The paper's central claim is that asynchronous event data and frame-based imagery are complementary enough, when fused in one network, to make anomaly detection both more accurate and much faster than frame-only methods. The architecture is a ResNet appearance branch plus a deep asynchronous graph network, DAGr, whose nodes are individual brightness-change events with normalized coordinates and scaled timestamps and whose edges join events that are close in space and time; each node's features are enriched with the RGB feature sampled at its location, and a GRU with object-level attention turns the fused features into risk scores. On ROL and DoTA the method reports AUC of 0.879 and 0.896, mean time-to-accident of 2.80 and 2.78 seconds, inference near 579 FPS, and mResponse of 1.17 and 1.21 seconds, outperforming all prior baselines on response time while staying at or near the top on accuracy; the one metric where a rival (TTHF) leads is frame-level AUC. The authors additionally show the event branch detects anomalies between consecutive frames, flagging fast-appearing objects before they are visible in the next image.

Load-bearing premise

The model is trained and evaluated on event streams generated from ordinary video by the v2e converter rather than recorded by a real event camera, and the paper's own validation of that synthetic fidelity covers normal driving scenes, so the millisecond-level inter-frame detection advantage is assumed to transfer to real hardware in actual anomaly moments.

Editorial extensions

If this is right

  • A detector that runs near 579 FPS and reports anomalies with sub-second mean response leaves room in the latency budget for braking and evasive planning, which is the paper's stated reason for measuring response time at all.
  • Because event streams carry timestamps between video frames, the model can raise an anomaly score before an abruptly appearing object is visible in the next RGB frame, an inter-frame detection capability the paper demonstrates on rush-out scenarios.
  • The mResponse metric, averaging detection delay across thresholds, gives a threshold-independent way to compare timeliness, and the experiments report it alongside FPS for every method.
  • In adverse weather and low-light subsets, the event branch keeps detection performance well above the best frame-based baselines, since events encode brightness change rather than absolute illumination.
  • The network is sized for on-board compute, with about 8.7 MFLOPs per event and roughly 87 TFLOPs worst case at high event rates, which the authors argue fits within Orin-class chip capability.

Reading between the lines

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

  • The decisive test the paper does not run is synthetic-to-real transfer: v2e fidelity is validated on classification and normal driving scenes only, so if a physical event camera's noise, bandwidth, or calibration differs in anomaly moments, the headline latency advantage could shrink on real hardware.
  • If the inter-frame advantage survives real hardware, event cameras look most plausible as a low-cost early-warning add-on to existing RGB-only stacks, targeted at sudden-appearance hazards, rather than as a replacement modality.
  • The mResponse idea generalizes: any time-critical perception output, from pedestrian warning to collision trigger, could report multi-threshold latency instead of a single time-to-event number, which would make latency comparisons meaningful across systems.
  • The 579 FPS figure assumes sparse event graphs; dense traffic at one to ten million events per second drives worst-case compute to roughly 87 TFLOPs, so the headline speed depends on event sparsity holding in congested scenes.
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 / 5 minor

Summary. The paper proposes a multimodal asynchronous hybrid network for real-time traffic anomaly detection in autonomous driving. The method fuses RGB frames (processed with a ResNet) and event streams (processed with an asynchronous graph neural network using spline convolution), then applies GRUs and attention at the object level to produce risk scores. The authors introduce a mean response time metric (mResponse) and claim millisecond-level real-time performance, reporting improvements in AUC, AP, mTTA, and response time on the ROL and DoTA datasets. They also present ablations, a new Rush-Out subset, comparisons with other real-time methods, and appendices on synthetic event validation and hardware deployment. The event data are generated from RGB videos using v2e because the datasets lack real event camera recordings.

Significance. If the results are reproducible and the v2e-generated events are a valid proxy for real event camera signals, the paper would demonstrate a useful method for anomaly detection with lower response time than several frame-based baselines, which matters for safety-critical driving. The paper ships a code link, provides extensive ablation tables, and discusses hardware feasibility, which are strengths. However, the central 'millisecond-level real-time' claim is directly contradicted by the paper's own mResponse values (1.17 s and 1.21 s), and the event modality is simulated from the same RGB frames, so the claimed advantages of event cameras' high temporal resolution are not empirically established. The architecture may still be of interest, but the headline claims need substantial revision and additional evidence.

major comments (5)
  1. [Abstract / Sec. 3 / Table 1] The abstract and Section 3 claim 'millisecond-level real-time performance' and 'response times at the millisecond level,' but Table 1 reports mResponse = 1.17 s on ROL and 1.21 s on DoTA. Since Eq. (3) defines R = ΔT_detection + T_inference, these are total response times; 1.17 s is 1170 ms. Section 5.3 then calls this 'exceptionally low mean response time,' which is internally inconsistent. The 579 FPS figure supports an inference latency of roughly 1.7 ms, but the paper does not limit its claim to inference latency. Please recalibrate the headline claims to the reported metric or re-report the metric in milliseconds with a clear breakdown of inference time and detection delay.
  2. [Sec. 5.1 / Appendix H / Sec. 5.5] The event stream is generated from the RGB video using v2e, as stated in Section 5.1 and Appendix H. Because the events are a deterministic transformation of the same RGB frames, they do not constitute an independent sensory modality; any information in the events is already present in the RGB input. Claims in Section 5.5 and Appendix D that the method performs 'inter-frame anomaly detection' and exploits the 'high temporal resolution' of event cameras are not supported by experiments with simulated events. The DSEC comparison in Appendix H validates v2e only on normal driving scenes, not on anomaly scenes, so it does not justify transfer to anomaly detection. The authors should either test with real event data on at least a subset of anomaly videos, or substantially soften the conclusions about event-camera advantages and add an explicit argument that v2e-generated events provide a meaningful proxy.
  3. [Sec. 5.2, Eq. (20)] The mResponse metric, which is the primary latency metric for the paper's central claim, is not reproducibly defined. Eq. (20) gives mResponse = (1/n) Σ Response_j, but the paper does not specify how Response_j is computed, what threshold values are used, or the value of n. Without this information, Table 1's latency numbers cannot be independently checked, and the comparison to baselines is not verifiable. Please provide the full protocol, including the threshold grid and the definition of Response_j.
  4. [Abstract / Table 1 / Sec. 5.3] The claim of outperforming existing methods 'in both accuracy and response time' is not supported by Table 1. For ROL, TTHF achieves AP = 0.585, while OURS achieves 0.570; for DoTA, TTHF achieves AUC-Frame = 0.847, while OURS achieves 0.823. Section 5.3 states 'OURS achieves leading performance in both AUC and AP metrics,' which is contradicted by these numbers. Please revise the claims to specify the metrics on which the method is actually superior, or include error bars and significance tests to support the superiority claims.
  5. [Sec. 5.4, Table 2] The ablation analysis attributes specific improvements to individual modules in a way not supported by the table. The text states 'Integrating GRUs increases AUC from 0.805 to 0.817 and AP from 0.479 to 0.508,' but the row with 0.817/0.508 corresponds to adding several modules at once (row 2 has five checkmarks), not GRU alone. Without a row that adds only the GRU to the baseline, the observed gain cannot be attributed to the GRU. The same issue applies to the attention module discussion. Please add per-module ablations or revise the attribution statements.
minor comments (5)
  1. [Table 2 / Table 5] Table 2 and Table 5 (Appendix A) contain rows with only checkmarks; the column alignment is ambiguous, so it is impossible to determine exactly which modules are active in each row. Please add explicit row labels (e.g., '+GRU') or use check/cross marks for all columns.
  2. [Table 8] Table 8 reports OURS mResponse = 2.35 s on ROL, while Table 1 reports 1.17 s for the same configuration; if both tables refer to the same model, this discrepancy needs to be reconciled.
  3. [Appendix H, Table 9] The comparison 'DSEC' vs 'DSEC+V2E' in Appendix H is not clearly described; please clarify whether both rows use real DSEC event data or whether the second row uses v2e-generated events from DSEC frames.
  4. [Appendix J] The statement 'requires only 8.732 MFLOPs per event' is confusing because MFLOPs is not naturally a per-event unit when combined with the event rate of 560k events/s; please specify the total computational cost per second or per frame.
  5. [Table 1] Table 1 reports single-point numbers for all metrics without error bars or multiple runs, which makes it difficult to assess whether the differences between OURS and the closest baselines are statistically meaningful.

Circularity Check

0 steps flagged · score 1.0 of 10

No meaningful circularity: the fusion network is built from external components and evaluated on external benchmarks; the v2e-based event modality is a data-validity caveat, not a derivation step that reduces to its input.

full rationale

The paper's derivation chain is not circular. The network combines externally cited building blocks (ResNet, DAGr, GRU, attention) with an event stream generated by the external v2e simulator from the same RGB video; no equation makes the predicted anomaly score equal to a fitted input or to the v2e transformation. The event-modality contribution is weakened by the fact that v2e events are a deterministic transform of RGB frames and Appendix H admits DSEC 'only covers normal driving and lacks anomaly events', so the real-event benefit is not independently validated; however, that is an external-validity limitation, not a self-referential derivation. The mResponse metric (Sec. 5.2, Eq. 20) is underspecified (Response_j and the threshold set are not given), and the abstract's 'millisecond-level' claim conflicts with mResponse = 1.17/1.21 s in Table 1; these are internal-consistency/reproducibility problems rather than circularity. Self-citations (e.g., STFE of Zhou et al. 2022, which shares author Dong Xiao, and related-work citations by Chen/Tian) appear only as baselines or background and are not load-bearing for the central claim, so at most a trivial score of 1 is warranted.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

The central claim rests on two external assumptions: that synthetic v2e events capture the real benefit of event cameras, and that upstream object detection is reliable. Both are flagged by the authors as limitations in the appendix. The method itself introduces no new physical entities; the mResponse metric is a new evaluation measure, not an entity.

free parameters (3)
  • Negative class weight = 0.27
    Used to balance classes in ROL training loss; chosen by the authors and likely tuned on the validation set.
  • Event graph construction parameters (radius R, max neighbors 16) = radius not given, neighbors = 16
    Chosen by hand to balance efficiency and accuracy; directly controls the graph structure fed to the GNN.
  • mResponse threshold set = undefined
    The metric averages detection delay over n thresholds, but the paper never specifies how many thresholds or their values, making the reported mResponse numbers hard to interpret or reproduce.
assumptions (3)
  • domain assumption v2e-generated event streams faithfully represent real DVS event camera outputs for traffic anomaly detection, including the temporal resolution that enables inter-frame detection.
    This is load-bearing because all event data in the experiments is synthesized from RGB video. The paper only validates v2e on normal driving scenes (DSEC), not on anomaly scenarios.
  • domain assumption The YOLOX object detector provides reliable bounding boxes for all relevant agents, including small and blurred objects.
    The anomaly pipeline consumes detected bounding boxes; the paper's own failure analysis in Appendix E shows detection failures cause anomaly detection failures.
  • domain assumption The event camera model in Eq. 4, where events trigger when |ΔL| > C, is a valid representation of real event sensors.
    This is the standard DVS model, but the paper relies on v2e's approximation of it rather than real sensor data.

how reviews work

0 comments
Cite this review

Pith. "Pith review of When Every Millisecond Counts: Real-Time Anomaly Detection via the Multimodal Asynchronous Hybrid Network." pith.science (2026). https://pith.science/paper/LUSYNCTP

@misc{pith2026250617457,
  author       = {Pith},
  title        = {Pith review of: When Every Millisecond Counts: Real-Time Anomaly Detection via the Multimodal Asynchronous Hybrid Network},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LUSYNCTP}},
  note         = {Machine review of arXiv:2506.17457}
}
read the original abstract

Anomaly detection is essential for the safety and reliability of autonomous driving systems. Current methods often focus on detection accuracy but neglect response time, which is critical in time-sensitive driving scenarios. In this paper, we introduce real-time anomaly detection for autonomous driving, prioritizing both minimal response time and high accuracy. We propose a novel multimodal asynchronous hybrid network that combines event streams from event cameras with image data from RGB cameras. Our network utilizes the high temporal resolution of event cameras through an asynchronous Graph Neural Network and integrates it with spatial features extracted by a CNN from RGB images. This combination effectively captures both the temporal dynamics and spatial details of the driving environment, enabling swift and precise anomaly detection. Extensive experiments on benchmark datasets show that our approach outperforms existing methods in both accuracy and response time, achieving millisecond-level real-time performance.

Figures

Figures reproduced from arXiv: 2506.17457 by the authors.

Figure 1
Figure 1. For real-time anomaly detection with an emphasis on response time, the overall response time is primarily influenced by the model’s inference duration and the time taken to identify anomalies. ciency (Huang et al., 2018; Caesar et al., 2020). The core of autonomous vehicles lies in their ability to perceive and interpret complex and dynamic driving environments accu￾rately and efficiently. Among the various percepti… view at source ↗
Figure 2
Figure 2. Overview of the proposed multimodal asynchronous hybrid network. (a) The framework integrates RGB images and event streams as inputs. Appearance features are extracted from RGB images using a ResNet architecture, and event features are derived from event streams through an asynchronous graph neural network (GNN) utilizing spline convolution. These features are then fused and processed by a detection head to generate… view at source ↗
Figure 3
Figure 3. In various traffic anomaly scenarios, the actions of hazardous vehicles significantly threaten the autonomous vehicle, causing elevated anomaly scores [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: When a vehicle suddenly cuts in, its attention score in￾creases as it approaches, indicating its growing importance as a potential anomaly. Oncoming Collision Risk [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 6
Figure 6. Figure 6: In scenarios where high-speed objects suddenly emerge, a continuous stream of events helps the model perform inter-frame anomaly detection, allowing for earlier and more timely anomaly detection. and assign them higher anomaly scores. Together, the GRU and attention mo…
Figure 7
Figure 7. Figure 7: Examples from the Rush-Out dataset demonstrate the effectiveness of our approach. At a tunnel exit under intense backlighting, a vehicle suddenly emerges from the tunnel edge. Leveraging the advantages of event data, our method enables rapid and accurate anomaly detect…
Figure 8
Figure 8. Figure 8: The anomaly detection results of suddenly rushing out of the scene, the objects moving at high speed in two consecutive frames can use the characteristics of event streams to help detect abnormalities between two frames [PITH_FULL_IMAGE:figures/full_fig_p018_8.png]
Figure 9
Figure 9. Figure 9: Additional traffic anomaly scenarios for real-time anomaly detection primarily involve abnormal vehicles obstructing the autonomous vehicle’s path, posing significant risks to its operation. 19 [PITH_FULL_IMAGE:figures/full_fig_p019_9.png]
Figure 10
Figure 10. Figure 10: Our approach is limited by the reliability of the object detector, and improvements in detection accuracy are essential. F. Sensitivity Analysis Our model remains robust across different settings. At the standard 0.5 IOU and 0.5 confidence thresholds, the model achiev…
Figure 11
Figure 11. Figure 11: Sensitivity analysis of the model’s hyperparameters IOU threshold and target detection confidence threshold. 20 [PITH_FULL_IMAGE:figures/full_fig_p020_11.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Few-shot Human Action Anomaly Detection via a Unified Contrastive Learning Framework

    cs.CV 2025-08 reject novelty 6.0 of 10

    A single contrastive action encoder with diffusion-based motion augmentation performs few-shot human action anomaly detection on HumanAct12 with an average AUC of 86.9%, including three categories excluded from encode...

Reference graph

Works this paper leans on

65 extracted references · 57 canonical work pages · cited by 1 Pith paper

  1. [1]

    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 format.date year duplicate empty "emp...

  2. [2]

    Drive: Deep reinforced accident anticipation with visual explanation

    Bao, W., Yu, Q., and Kong, Y. Drive: Deep reinforced accident anticipation with visual explanation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp.\ 7619--7628, 2021

    work page 2021

  3. [3]

    Batanina, E., Bekkouch, I. E. I., Youssry, Y., Khan, A., Khattak, A. M., and Bortnikov, M. Domain adaptation for car accident detection in videos. In 2019 ninth international conference on image processing theory, tools and applications (IPTA), pp.\ 1--6. IEEE, 2019

    work page 2019

  4. [4]

    Efficientad: Accurate visual anomaly detection at millisecond-level latencies

    Batzner, K., Heckler, L., and K \"o nig, R. Efficientad: Accurate visual anomaly detection at millisecond-level latencies. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pp.\ 128--138, 2024

    work page 2024

  5. [5]

    H., Vora, S., Liong, V

    Caesar, H., Bankiti, V., Lang, A. H., Vora, S., Liong, V. E., Xu, Q., Krishnan, A., Pan, Y., Baldan, G., and Beijbom, O. nuscenes: A multimodal dataset for autonomous driving. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.\ 11621--11631, 2020

    2020

  6. [6]

    Freeway traffic incident detection from cameras: A semi-supervised learning approach

    Chakraborty, P., Sharma, A., and Hegde, C. Freeway traffic incident detection from cameras: A semi-supervised learning approach. In 2018 21st International Conference on Intelligent Transportation Systems (ITSC), pp.\ 1840--1845. IEEE, 2018

    work page 2018

  7. [7]

    Anticipating accidents in dashcam videos

    Chan, F.-H., Chen, Y.-T., Xiang, Y., and Sun, M. Anticipating accidents in dashcam videos. In Computer Vision--ACCV 2016: 13th Asian Conference on Computer Vision, Taipei, Taiwan, November 20-24, 2016, Revised Selected Papers, Part IV 13, pp.\ 136--153. Springer, 2017

    work page 2016

  8. [8]

    Adaptive discovering and merging for incremental novel class discovery

    Chen, G., Peng, P., Huang, Y., Geng, M., and Tian, Y. Adaptive discovering and merging for incremental novel class discovery. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pp.\ 11276--11284, 2024

    work page 2024

Show all 65 references
  1. [9]

    Automated essential concept discovery for few-shot out-of-distribution detection

    Chen, G., Horstmann, K., Wang, Z., and You, F. Automated essential concept discovery for few-shot out-of-distribution detection. In The IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2025

    2025

  2. [10]

    Fblnet: Feedback loop network for driver attention prediction

    Chen, Y., Nan, Z., and Xiang, T. Fblnet: Feedback loop network for driver attention prediction. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp.\ 13371--13380, 2023

    2023

  3. [11]

    Chong, Y. S. and Tay, Y. H. Abnormal event detection in videos using spatiotemporal autoencoder. In Advances in Neural Networks-ISNN 2017: 14th International Symposium, ISNN 2017, Sapporo, Hakodate, and Muroran, Hokkaido, Japan, June 21--26, 2017, Proceedings, Part II 14, pp.\...

    2017

  4. [12]

    Gorela: Go relative for viewpoint-invariant motion forecasting

    Cui, A., Casas, S., Wong, K., Suo, S., and Urtasun, R. Gorela: Go relative for viewpoint-invariant motion forecasting. In 2023 IEEE International Conference on Robotics and Automation (ICRA), pp.\ 7801--7807. IEEE, 2023

    2023

  5. [13]

    K., Winn, J., and Zisserman, A

    Everingham, M., Van Gool, L., Williams, C. K., Winn, J., and Zisserman, A. The pascal visual object classes (voc) challenge. International journal of computer vision, 88: 0 303--338, 2010

    2010

  6. [14]

    Traffic accident detection via self-supervised consistency learning in driving scenarios

    Fang, J., Qiao, J., Bai, J., Yu, H., and Xue, J. Traffic accident detection via self-supervised consistency learning in driving scenarios. IEEE Transactions on Intelligent Transportation Systems, 23 0 (7): 0 9601--9614, 2022

    2022

  7. [15]

    Vision-based traffic accident detection and anticipation: A survey

    Fang, J., Qiao, J., Xue, J., and Li, Z. Vision-based traffic accident detection and anticipation: A survey. IEEE Transactions on Circuits and Systems for Video Technology, 2023

    2023

  8. [16]

    Abductive ego-view accident video understanding for safe driving perception

    Fang, J., Li, L.-l., Zhou, J., Xiao, J., Yu, H., Lv, C., Xue, J., and Chua, T.-S. Abductive ego-view accident video understanding for safe driving perception. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 22030--22040, 2024

    2024

  9. [17]

    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. Event-based vision: A survey. IEEE transactions on pattern analysis and machine intelligence, 44 0 (1): 0 154--180, 2020

    2020

  10. [18]

    and Scaramuzza, D

    Gehrig, D. and Scaramuzza, D. Low-latency automotive vision with event cameras. Nature, 629 0 (8014): 0 1034--1040, 2024

    2024

  11. [19]

    Dsec: A stereo event camera dataset for driving scenarios

    Gehrig, M., Aarents, W., Gehrig, D., and Scaramuzza, D. Dsec: A stereo event camera dataset for driving scenarios. IEEE Robotics and Automation Letters, 6 0 (3): 0 4947--4954, 2021

    2021

  12. [20]

    R., Venkatesh, S., and Hengel, A

    Gong, D., Liu, L., Le, V., Saha, B., Mansour, M. R., Venkatesh, S., and Hengel, A. v. d. Memorizing normality to detect anomaly: Memory-augmented deep autoencoder for unsupervised anomaly detection. In Proceedings of the IEEE/CVF international conference on computer vision, pp...

    2019

  13. [21]

    and Pedraza, C

    Gutierrez-Osorio, C. and Pedraza, C. Modern data sources and techniques for analysis and forecast of road accidents: A review. Journal of traffic and transportation engineering (English edition), 7 0 (4): 0 432--446, 2020

    2020

  14. [22]

    Hanley, J. A. and McNeil, B. J. The meaning and use of the area under a receiver operating characteristic (roc) curve. Radiology, 143 0 (1): 0 29--36, 1982

    1982

  15. [23]

    A., Rehman, F

    Haris, M., Moin, M. A., Rehman, F. A., and Farhan, M. Vehicle crash prediction using vision. In 2021 55th Annual Conference on Information Sciences and Systems (CISS), pp.\ 1--5. IEEE, 2021

    2021

  16. [24]

    K., and Davis, L

    Hasan, M., Choi, J., Neumann, J., Roy-Chowdhury, A. K., and Davis, L. S. Learning temporal regularity in video sequences. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.\ 733--742, 2016

    2016

  17. [25]

    Deep residual learning for image recognition

    He, K., Zhang, X., Ren, S., and Sun, J. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.\ 770--778, 2016

    2016

  18. [26]

    Cost-sensitive semi-supervised deep learning to assess driving risk by application of naturalistic vehicle trajectories

    Hu, H., Wang, Q., Cheng, M., and Gao, Z. Cost-sensitive semi-supervised deep learning to assess driving risk by application of naturalistic vehicle trajectories. Expert Systems with Applications, 178: 0 115041, 2021 a

    2021

  19. [27]

    v2e: From video frames to realistic dvs events

    Hu, Y., Liu, S.-C., and Delbruck, T. v2e: From video frames to realistic dvs events. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.\ 1312--1321, 2021 b

    2021

  20. [28]

    The apolloscape dataset for autonomous driving

    Huang, X., Cheng, X., Geng, Q., Cao, B., Zhou, D., Wang, P., Lin, Y., and Yang, R. The apolloscape dataset for autonomous driving. In Proceedings of the IEEE conference on computer vision and pattern recognition workshops, pp.\ 954--960, 2018

    2018

  21. [29]

    M., Li, Y., Qin, R., and Yin, Z

    Karim, M. M., Li, Y., Qin, R., and Yin, Z. A dynamic spatial-temporal attention network for early anticipation of traffic accidents. IEEE Transactions on Intelligent Transportation Systems, 23 0 (7): 0 9590--9600, 2022

    2022

  22. [30]

    M., Yin, Z., and Qin, R

    Karim, M. M., Yin, Z., and Qin, R. An attention-guided multistream feature fusion network for early localization of risky traffic agents in driving videos. IEEE Transactions on Intelligent Vehicles, 9 0 (1): 0 1792--1803, 2023

    2023

  23. [31]

    Attention r-cnn for accident detection

    Le, T.-N., Ono, S., Sugimoto, A., and Kawasaki, H. Attention r-cnn for accident detection. In 2020 IEEE intelligent vehicles symposium (IV), pp.\ 313--320. IEEE, 2020

    2020

  24. [32]

    Graph-based asynchronous event processing for rapid object recognition

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

    2021

  25. [33]

    A memory-augmented multi-task collaborative framework for unsupervised traffic accident detection in driving videos

    Liang, R., Li, Y., Yi, Y., Zhou, J., and Li, X. A memory-augmented multi-task collaborative framework for unsupervised traffic accident detection in driving videos. arXiv preprint arXiv:2307.14575, 2023

    2023 arXiv

  26. [34]

    Text-driven traffic anomaly detection with temporal high-frequency modeling in driving videos

    Liang, R., Li, Y., Zhou, J., and Li, X. Text-driven traffic anomaly detection with temporal high-frequency modeling in driving videos. IEEE Transactions on Circuits and Systems for Video Technology, 2024

    2024

  27. [35]

    Lin, T.-Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., Doll \'a r, P., and Zitnick, C. L. Microsoft coco: Common objects in context. In Computer Vision--ECCV 2014: 13th European Conference, Zurich, Switzerland, September 6-12, 2014, Proceedings, Part V 13, pp....

    2014

  28. [36]

    Future frame prediction for anomaly detection--a new baseline

    Liu, W., Luo, W., Lian, D., and Gao, S. Future frame prediction for anomaly detection--a new baseline. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.\ 6536--6545, 2018

    2018

  29. [37]

    H., and Li, J

    Malla, S., Choi, C., Dwivedi, I., Choi, J. H., and Li, J. Drama: Joint risk localization and captioning in driving. In Proceedings of the IEEE/CVF winter conference on applications of computer vision, pp.\ 1043--1052, 2023

    2023

  30. [38]

    Towards explainable artificial intelligence (xai) for early anticipation of traffic accidents

    Monjurul Karim, M., Li, Y., and Qin, R. Towards explainable artificial intelligence (xai) for early anticipation of traffic accidents. arXiv e-prints, pp.\ arXiv--2108, 2021

    2021

  31. [39]

    T., Popescu, M., Khan, F

    Ristea, N.-C., Croitoru, F.-A., Ionescu, R. T., Popescu, M., Khan, F. S., Shah, M., et al. Self-distilled masked auto-encoders are efficient video anomaly detectors. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 15984--15995, 2024

    2024

  32. [40]

    Memory-augmented online video anomaly detection

    Rossi, L., Bernuzzi, V., Fontanini, T., Bertozzi, M., and Prati, A. Memory-augmented online video anomaly detection. In ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp.\ 6590--6594. IEEE, 2024

    2024

  33. [41]

    K., and Fukuda, A

    Roy, D., Ishizaka, T., Mohan, C. K., and Fukuda, A. Detection of collision-prone vehicle behavior at intersections using siamese interaction lstm. IEEE transactions on intelligent transportation systems, 23 0 (4): 0 3137--3147, 2020

    2020

  34. [42]

    K., Dogra, D

    Santhosh, K. K., Dogra, D. P., Roy, P. P., and Mitra, A. Vehicular trajectory classification and traffic anomaly detection in videos using a hybrid cnn-vae architecture. IEEE Transactions on Intelligent Transportation Systems, 23 0 (8): 0 11891--11902, 2021

    2021

  35. [43]

    Potential risk localization via weak labeling out of blind spot

    Shimomura, K., Hirakawa, T., Yamashita, T., and Fujiyoshi, H. Potential risk localization via weak labeling out of blind spot. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 4504--4513, 2024

    2024

  36. [44]

    Classification of crash and near-crash events from dashcam videos and telematics

    Taccari, L., Sambo, F., Bravi, L., Salti, S., Sarti, L., Simoncini, M., and Lori, A. Classification of crash and near-crash events from dashcam videos and telematics. In 2018 21st International Conference on intelligent transportation systems (ITSC), pp.\ 2460--2465. IEEE, 2018

    2018

  37. [45]

    G., Huynh, M

    Tamagusko, T., Correia, M. G., Huynh, M. A., and Ferreira, A. Deep learning applied to road accident detection with transfer learning and synthetic images. Transportation research procedia, 64: 0 90--97, 2022

    2022

  38. [46]

    Graph (graph): A nested graph-based framework for early accident anticipation

    Thakur, N., Gouripeddi, P., and Li, B. Graph (graph): A nested graph-based framework for early accident anticipation. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pp.\ 7533--7541, 2024

    2024

  39. [47]

    Latency-aware road anomaly segmentation in videos: A photorealistic dataset and new metrics

    Tian, B., Gao, H.-a., Cui, L., Zheng, Y., Luo, L., Wang, B., Zhi, R., Zhou, G., and Zhao, H. Latency-aware road anomaly segmentation in videos: A photorealistic dataset and new metrics. arXiv preprint arXiv:2401.04942, 2024

    2024 arXiv

  40. [48]

    W., and Chanda, P

    Veeramani, B., Raymond, J. W., and Chanda, P. Deepsort: deep convolutional networks for sorting haploid maize seeds. BMC bioinformatics, 19: 0 1--9, 2018

    2018

  41. [49]

    K., Dogra, D

    Vijay, T. K., Dogra, D. P., Choi, H., Nam, G., and Kim, I.-J. Detection of road accidents using synthetically generated multi-perspective accident videos. IEEE Transactions on Intelligent Transportation Systems, 24 0 (2): 0 1926--1935, 2022

    1926

  42. [50]

    Prophnet: Efficient agent-centric motion forecasting with anchor-informed proposals

    Wang, X., Su, T., Da, F., and Yang, X. Prophnet: Efficient agent-centric motion forecasting with anchor-informed proposals. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 21995--22003, 2023

    2023

  43. [51]

    A new framework of vehicle collision prediction by combining svm and hmm

    Xiong, X., Chen, L., and Liang, J. A new framework of vehicle collision prediction by combining svm and hmm. IEEE Transactions on Intelligent Transportation Systems, 19 0 (3): 0 699--710, 2017

    2017

  44. [52]

    Classifying near-miss traffic incidents through video, sensor, and object features

    Yamamoto, S., Kurashima, T., and Toda, H. Classifying near-miss traffic incidents through video, sensor, and object features. IEICE TRANSACTIONS on Information and Systems, 105 0 (2): 0 377--386, 2022

    2022

  45. [53]

    J., and Atkins, E

    Yao, Y., Xu, M., Wang, Y., Crandall, D. J., and Atkins, E. M. Unsupervised traffic accident detection in first-person videos. In 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp.\ 273--280. IEEE, 2019

    2019

  46. [54]

    Yao, Y., Wang, X., Xu, M., Pu, Z., Wang, Y., Atkins, E., and Crandall, D. J. Dota: Unsupervised detection of traffic anomaly in driving videos. IEEE transactions on pattern analysis and machine intelligence, 45 0 (1): 0 444--459, 2022

    2022

  47. [55]

    and Han, B

    You, T. and Han, B. Traffic accident benchmark for causality recognition. In Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part VII 16, pp.\ 540--556. Springer, 2020

    2020

  48. [56]

    Agent-centric risk assessment: Accident anticipation and risky region localization

    Zeng, K.-H., Chou, S.-H., Chan, F.-H., Carlos Niebles, J., and Sun, M. Agent-centric risk assessment: Accident anticipation and risky region localization. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.\ 2222--2230, 2017

    2017

  49. [57]

    Systems and methods for actor motion forecasting within a surrounding environment of an autonomous vehicle, November 2 2023

    Zeng, W., Liao, R., Urtasun, R., and Liang, M. Systems and methods for actor motion forecasting within a surrounding environment of an autonomous vehicle, November 2 2023. US Patent App. 18/346,518

    2023

  50. [58]

    Training-free anomaly event detection via llm-guided symbolic pattern discovery

    Zeng, Y., Wu, H., Nie, W., Chen, G., Zheng, X., Shen, Y., Li, G., Zou, Y., Tian, Y., and Ji, R. Training-free anomaly event detection via llm-guided symbolic pattern discovery. arXiv preprint arXiv:2502.05843, 2025

    2025 arXiv

  51. [59]

    Anonymous model pruning for compressing deep neural networks

    Zhang, L., Chen, G., Shi, Y., Zhang, Q., Tan, M., Wang, Y., Tian, Y., and Huang, T. Anonymous model pruning for compressing deep neural networks. In 2020 IEEE Conference on Multimedia Information Processing and Retrieval (MIPR), pp.\ 157--160. IEEE, 2020

    2020

  52. [60]

    Micm: Rethinking unsupervised pretraining for enhanced few-shot learning

    Zhang, Z., Chen, G., Zou, Y., Huang, Z., Li, Y., and Li, R. Micm: Rethinking unsupervised pretraining for enhanced few-shot learning. In Proceedings of the 32nd ACM International Conference on Multimedia, pp.\ 7686--7695, 2024 a

    2024

  53. [61]

    Learning unknowns from unknowns: Diversified negative prototypes generator for few-shot open-set recognition

    Zhang, Z., Chen, G., Zou, Y., Li, Y., and Li, R. Learning unknowns from unknowns: Diversified negative prototypes generator for few-shot open-set recognition. In Proceedings of the 32nd ACM International Conference on Multimedia, pp.\ 6053--6062, 2024 b

    2024

  54. [62]

    Spatio-temporal autoencoder for video anomaly detection

    Zhao, Y., Deng, B., Shen, C., Liu, Y., Lu, H., and Hua, X.-S. Spatio-temporal autoencoder for video anomaly detection. In Proceedings of the 25th ACM international conference on Multimedia, pp.\ 1933--1941, 2017

    1933

  55. [63]

    and Jiang, J

    Zhou, B. and Jiang, J. Deep event-based object detection in autonomous driving: A survey. arXiv preprint arXiv:2405.03995, 2024

    2024 arXiv

  56. [64]

    Spatio-temporal feature encoding for traffic accident detection in vanet environment

    Zhou, Z., Dong, X., Li, Z., Yu, K., Ding, C., and Yang, Y. Spatio-temporal feature encoding for traffic accident detection in vanet environment. IEEE Transactions on Intelligent Transportation Systems, 23 0 (10): 0 19772--19781, 2022

    2022

  57. [65]

    Zou, Y., Zhang, S., Chen, G., Tian, Y., Keutzer, K., and Moura, J. M. Annotation-efficient untrimmed video action recognition. In Proceedings of the 29th ACM International Conference on Multimedia, pp.\ 487--495, 2021

    2021

Pith tools

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