Pith. sign in

REVIEW 4 major objections 8 minor 88 references

Event Stream based Multi-Modal Video Anomaly Detection: A Benchmark Dataset and Algorithms

T0 review · 4 major / 8 minor · reviewed 2026-07-13 · grok-4.5

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

Pith's one-line read Real event cameras plus video and text alignment make surveillance anomaly detection far more reliable under bad light and fast motion.

desk verdict Real visible–event VAD dataset is the real contribution; the pipeline works on it, but “events are essential” is oversold and public-benchmark events are mostly non-physical. read the letter →

arxiv 2607.09114 v1 pith:FZ5C3BWR submitted 2026-07-10 cs.CV cs.AIcs.MM

Peipei Zhu , Yueqing Niu , Lin Zhu , Guanchong Niu , Yang Yu , Zheng Li This is my paper
classification cs.CVcs.AIcs.MM
keywords videoanomalydetectioneventcameramulti-modallearningfeaturefusioncontrastivepretrainingweaklysupervisedTJUTCMPha
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

Standard video anomaly detection fails when lighting changes, motion is fast, or backgrounds are cluttered because ordinary cameras smear or wash out the cues. This paper argues that pairing ordinary video with genuine event-camera streams fixes that: events fire only on brightness changes at microsecond resolution, so they keep clean motion signals when video collapses. The authors release TJUTCM Pha, the first large real-world visible-plus-event dataset recorded in pharmaceutical labs and production lines (6.3 billion events, 376k frames, 13 anomaly types). On top of it they build E-VAD: a contrastive pretraining stage that aligns event, video and text embeddings so sparse event features become semantically useful, then an adaptive weighting module that trusts the more reliable modality moment by moment. Across public benchmarks and their own dataset the method beats strong video-only and multi-modal baselines, with especially large gains on the real industrial data, supporting the claim that event sensing is not optional decoration but essential for robust real-world anomaly detection.

What carries the argument

E-VAD's two-stage pipeline: contrastive multi-modal pretraining that projects event frames into a shared CLIP-anchored space with video and ConceptNet-expanded text, followed by adaptive weighting fusion (a learned per-feature gate α that mixes video and event features) that supplies the anomaly classifier.

What would settle it

Run the identical E-VAD pipeline on a held-out real hybrid-camera capture set where events are never accumulated into frames (or where event frames are deliberately degraded to match the noise profile of monitor-re-recorded events) and check whether the large AUC gap over video-only and re-trained multi-modal baselines disappears.

Watch

Extended reading notes

Core claim

Jointly using real event streams with visible video, via contrastive multi-modal pretraining that aligns event-video-text embeddings and an adaptive weighting fusion that balances temporal event cues against spatial video features, yields consistently higher weakly-supervised anomaly detection performance than video-only or naively fused baselines, establishing that authentic event sensing is essential for robust real-world VAD.

Load-bearing premise

That turning raw asynchronous events into ordinary image-like frames and projecting them into a vision-language space trained on everyday photos still keeps enough of the original high-speed, high-dynamic-range motion signal for the claimed complementarity to hold.

Editorial extensions

If this is right

  • Surveillance systems that add a real event camera can keep detecting anomalies under low light, glare and rapid motion that currently force video-only detectors to fail.
  • New VAD benchmarks and methods must treat genuine event streams as a first-class modality rather than as a derived or simulated add-on.
  • The same contrastive alignment-plus-adaptive-fusion recipe can be reused for other safety-critical perception tasks that already have hybrid event-video sensors.
  • Industrial settings with subtle, sparse anomalies (labs, production lines) become practical targets for weakly supervised multi-modal detectors once real event data are available.

Reading between the lines

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

  • If the event-frame + CLIP-projection step is the weakest link, later work that keeps events in native sparse or spiking form should widen the gap still further on real industrial data.
  • The large performance jump on TJUTCM Pha versus the more modest gains on simulated or re-recorded public sets already hints that dataset realism, not just model architecture, is the main driver of the reported advance.
  • The adaptive gate itself could become a diagnostic: high event weight regions may serve as automatic pointers to lighting or blur failures in deployed video systems.
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 / 8 minor

Summary. The paper proposes E-VAD, a weakly supervised multi-modal video anomaly detection framework that fuses conventional RGB video with asynchronous event streams. It introduces TJUTCM Pha, claimed as the first real-scene visible–event VAD benchmark (6.3B events, ~376k frames) collected with a DAVIS346 in pharmaceutical lab/production settings, and a two-stage method: (i) contrastive multi-modal pretraining that aligns event, video, and ConceptNet-expanded text embeddings in a CLIP-anchored space, and (ii) an adaptive weighting fusion (AWF) module that gates video vs. event features before temporal modeling and MIL-style snippet scoring. Experiments on ShanghaiTech-Event (simulated events), UCF-Crime-DVS (monitor re-recorded events), and TJUTCM Pha report strong gains, especially 82.25% AUC / 0.00% FAR on TJUTCM Pha versus re-trained multi-modal baselines around 72–74%, and the authors conclude that event sensing is complementary and essentially necessary for robust real-world VAD.

Significance. If the sensing-complementarity claim holds under fair controls, the work is significant for VAD: it reframes robustness as a sensing problem rather than purely architectural, and TJUTCM Pha would be a useful community resource because it is real hybrid capture rather than screen-replay or pure simulation. Strengths that should be credited include the scale and industrial realism of TJUTCM Pha, the explicit acknowledgment of limitations of UCF-Crime-DVS and ShanghaiTech-Event, thorough ablations of fusion variants / temporal modules / event representations on TJUTCM Pha, and qualitative score-curve and t-SNE analyses that support the design narrative. The free parameters (contrastive weights, μ, λ, top-k rule, etc.) are standard for this literature and do not by themselves invalidate the results. The main open question is attribution: how much of the large TJUTCM gap is physical event sensing versus the new CLIP+TM+AWF+text pipeline.

major comments (4)
  1. Central attribution claim vs. missing matched control (Abstract; §I contributions; Table V; §V.C–D). The paper asserts that event sensing is not merely beneficial but “essential,” with the 82.25% AUC on TJUTCM Pha (vs. ~73.8% best re-trained baseline) as primary evidence. Table V shows video-only 75.30, event-only 78.15, video+event 80.91, and full model 82.25, but there is no control that runs the identical CLIP encoder + temporal module + AWF-style head + MIL/KL training on video alone (or with a non-event second stream) on TJUTCM Pha. Without that matched architecture control, the jump from ~73–75% external video baselines to 82.25% cannot be cleanly attributed to physical event complementarity rather than the new pretraining/fusion pipeline. Please add this control (and ideally a non-event auxiliary stream) and revise the “essential” language to match what the controls support.
  2. Overstated “consistently outperforms” relative to public benchmarks (Abstract; §I; Tables II–III). On ShanghaiTech-Event, E-VAD reaches 98.67% AUC and is best among reported methods; on UCF-Crime-DVS, E-VAD AUC is 88.75%, below ITC’s 89.04% and only modestly above VADCLIP (88.02). Several re-trained video+event baselines also drop or barely move (e.g., VADCLIP* 87.94, UML* 85.95), which the paper itself links to non-physical event generation. The abstract and contribution bullets should be revised to state dataset-specific outcomes accurately rather than a blanket “consistently outperforms,” and to separate gains on real hybrid capture from gains on simulated/re-recorded events.
  3. Public-benchmark event validity and self-constructed evaluation risk (§IV Table I; §V.A; ShanghaiTech-Event construction). ShanghaiTech-Event is the authors’ DVS-Voltmeter simulation extension, and UCF-Crime-DVS is monitor re-recording; both inherit video temporal/photometric limits, as the manuscript notes. Using a self-built simulation extension as a primary SOTA table (Table II) while claiming event-driven superiority creates a mild circularity risk. Please (i) move ShanghaiTech-Event results to a clearly labeled “controlled / simulated” analysis, (ii) report any available statistics on event noise/sparsity differences vs. TJUTCM Pha, and (iii) avoid treating simulated/re-recorded gains as equivalent evidence for “real event sensing” in the conclusion.
  4. Event-frame representation vs. claimed microsecond / HDR advantages (§III.B.1; Table VIII; weakest assumption in the design). Asynchronous events are accumulated into CLIP-compatible event frames and projected into a natural-image CLIP space. Table VIII shows event frames beat voxel grids and time surfaces under the authors’ protocol, but this does not demonstrate that microsecond timing and high dynamic range survive the accumulation+CLIP pathway—especially when public events are already non-physical. Please quantify temporal binning (interval length, polarity handling) and discuss, with evidence, which claimed event advantages remain after framing; if the benefit is mainly motion-salient 2D patterns rather than true asynchronous sensing, the introduction and conclusion should say so.
minor comments (8)
  1. Abstract and elsewhere: “E VAD” / “EVAD” / “E-V AD” spacing is inconsistent; standardize to E-VAD.
  2. Table I: “#Frame” for TJUTCM Pha is listed as 377k while the abstract/text use 376,368; reconcile the count.
  3. Eq. (5)–(6): notation mixes L_ρq / L_et / L_ve / L_vt and “ρ! =q”; clarify that the three pairwise losses are the only terms and give default θ, β, γ values used in experiments.
  4. Eq. (12): L_kd is written with p_v2t / q_v2t but the surrounding text does not fully specify how semantic consistency labels are built for multi-class anomaly categories under weak labels.
  5. Fig. 2 caption and §III: “CLIP is not used as a pretrained event encoder” is important; consider elevating this clarification earlier to avoid reader confusion with EventCLIP/EventBind.
  6. §V.B: batch size 512 for 1000 epochs with max 200-frame sequences is heavy; a short note on wall-clock cost or hardware would aid reproducibility.
  7. Project link for TJUTCM Pha is deferred (“will be release later”); for a dataset paper this should be concrete before acceptance, including split files and event format.
  8. Minor typos: “outperforms methods” (abstract), “dimen-sion” (Eq. 9 context), “A WF” spacing in Table VI, “V oxel” in Table VIII.

Circularity Check

0 steps flagged · score 1.0 of 10

No derivation circularity; standard contrastive/MIL training and empirical SOTA claims on a new real event dataset plus public benchmarks (with simulated/re-recorded events noted as limited).

full rationale

This is an empirical multi-modal VAD systems paper. The load-bearing claims (E-VAD outperforms via event–video–text contrastive pretraining + adaptive fusion; event sensing is essential) rest on standard losses (pairwise InfoNCE-style L_et/L_ve/L_vt, MIL top-k classification L_cs, and KL L_kd) and measured AUC/FAR on held-out test splits. No equation defines a quantity in terms of the target it then “predicts,” no parameter is fitted to a subset and then reported as an independent prediction of a near-identical quantity, and no uniqueness theorem or ansatz is imported via self-citation to force the architecture. ShanghaiTech-Event is an author-constructed simulation extension (via DVS-Voltmeter) used as one controlled benchmark, and TJUTCM Pha is their new real capture set; both are normal dataset contributions, not algebraic self-definitions of the reported scores. Public UCF-Crime-DVS results and ablations (video-only, event-only, fusion variants) supply independent grounding. Mild self-construction risk exists only in the sense that the largest absolute gains appear on the authors’ own industrial set, but that does not reduce any claimed result to its inputs by construction. Score 1 reflects that residual dataset-construction proximity without elevating it to circular derivation.

Assumptions & free parameters 7 free parameters · 6 assumptions · 3 invented entities

The central empirical claim rests on standard deep-learning and weak-supervision machinery plus several domain choices: event-to-frame conversion, CLIP as semantic anchor, ConceptNet text expansion, and hand-weighted multi-loss training. Free parameters are the usual loss/fusion coefficients and training hyperparameters. No new physical entities are postulated; the ‘invented’ objects are engineering modules and a dataset.

free parameters (7)
  • contrastive loss weights θ, β, γ (L_totalpre)
    Hand-chosen coefficients balancing event–text, video–event, and video–text contrastive terms; not derived from first principles.
  • detection loss weight λ (L = L_ce + λ L_kd)
    Balances MIL classification against KL semantic consistency; fitted/selected for training stability and accuracy.
  • temporal mix μ (F_o = μ X_g + (1-μ) X_l)
    Learnable interpolation between global and local attention; free capacity that shapes temporal features.
  • contrastive temperature τ
    Standard InfoNCE temperature controlling softness of alignment; affects representation geometry.
  • fusion gate initialization α≈0.5 and linear gate parameters W_α, b_α
    Initial modality balance and learned reliability weights; central to claimed adaptive fusion behavior.
  • top-k snippet rule (k=⌊T/16+1⌋ anomalous, k=1 normal)
    Heuristic from prior MIL VAD work that defines how video-level labels supervise snippet scores.
  • training hyperparameters (lr 1e-3, batch 512, 1000 epochs, max 200 frames, local interval 5)
    Chosen for optimization; can move reported AUC/FAR without changing the architecture story.
assumptions (6)
  • domain assumption Frame-like accumulation of events over intervals preserves enough temporal dynamics for VAD (inter-frame evolution dominates fine intra-frame timing).
    Stated in §III.A; justifies discarding full asynchronous event modeling in favor of CLIP-friendly event frames.
  • domain assumption CLIP image/text embedding space is a valid semantic anchor for projecting event features via a learnable projector.
    §III.B; core of multi-modal pretraining. CLIP was trained on natural images, not events.
  • domain assumption Video-level binary labels plus multiple-instance learning (top-k mean) suffice to localize snippet anomalies under weak supervision.
    Standard VAD assumption adopted in §III.D–E following [45,47,48].
  • domain assumption ConceptNet-expanded category text provides useful supervisory semantics for anomaly subclasses.
    §III.B.1; language side of contrastive alignment.
  • ad hoc to paper Simulated (ShanghaiTech-Event) and monitor-rerecorded (UCF-Crime-DVS) events are informative enough for multi-modal comparison, even if not fully realistic.
    §V.A acknowledges limitations but still uses these sets for SOTA tables that support the ‘consistently outperforms’ narrative.
  • standard math Scaled dot-product attention and causal convolution are appropriate temporal models for snippet scoring.
    Standard neural components used in §III.B.2 and §III.D.
invented entities (3)
  • TJUTCM Pha dataset
    purpose: Provide the first real-world synchronized visible–event benchmark for pharmaceutical VAD (6.3B events, 376k frames, 13 anomaly classes).
    Core empirical resource; independent_evidence is false until public release and third-party use.
  • E-VAD adaptive weighting fusion (AWF) module
    purpose: Dynamically gate video vs event features via a learned sigmoid weight α before anomaly scoring.
    Engineering module, not a physical entity; value is only through reported ablations on the authors’ data.
  • ShanghaiTech-Event extension
    purpose: Simulated event streams for the ShanghaiTech campus VAD benchmark to enable multi-modal comparison.
    Author-constructed simulation resource; inherits clean video limitations noted in §V.A.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Event Stream based Multi-Modal Video Anomaly Detection: A Benchmark Dataset and Algorithms." pith.science (2026). https://pith.science/paper/FZ5C3BWR

@misc{pith2026260709114,
  author       = {Pith},
  title        = {Pith review of: Event Stream based Multi-Modal Video Anomaly Detection: A Benchmark Dataset and Algorithms},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FZ5C3BWR}},
  note         = {Machine review of arXiv:2607.09114}
}
read the original abstract

Video anomaly detection (VAD) is critical for automated surveillance but remains fragile under challenging conditions such as illumination variations, fast motion, and complex backgrounds when relying solely on visible light videos. To address these limitations, we propose EVAD, an event enhanced VAD framework that jointly exploits conventional video and event streams captured by bio inspired event cameras. Event sensors asynchronously capture brightness changes with high temporal resolution, offering robustness to motion blur and extreme lighting, and providing motion salient cues complementary to video based visual information. To support multi modal VAD research, we construct a large scale visible event benchmark comprising 6.3 billion events and 376,368 video frames collected under diverse illumination levels, motion patterns, and background complexities, filling the gap of realistic and scalable datasets for event based anomaly detection. Building upon this dataset, we design a contrastive multi modal pretraining framework to learn discriminative event representations by aligning semantic embeddings across event streams, visible videos, and textual descriptions. An adaptive fusion module then dynamically integrates event based temporal cues with video based spatial semantics, improving robustness to environmental disturbances. Experiments on benchmarks and the proposed TJUTCM Pha dataset demonstrate that E VAD consistently outperforms methods, validating the effectiveness of event-based sensing for VAD in real world scenarios.

Figures

Figures reproduced from arXiv: 2607.09114 by the authors.

Figure 1
Figure 1. Comparison of sensing characteristics between visible-light cameras [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. An overview of the proposed E-VAD framework via collaboration of visible and event frames, which comprises two key modules: (a) a multi-modal [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Architecture of the proposed adaptive weighting fusion module, which ˆˆ [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Representative synchronized visible and event frames from the [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Visualization of feature embeddings before and after multi-modal pretraining via t-SNE [87] on the proposed TJUTCM Pha dataset. Post-training [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: The frame-level visualizations for the abnormal detection of the proposed E-VAD method. Each example includes sampled video frames, event frames, [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: Radar charts showing per-class anomaly detection performance across [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: Radar charts showing anomaly detection performance under four [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

88 extracted references · 3 linked inside Pith

  1. [1]

    Narwade, R

    P. Narwade, R. Kawamura, G. Gajbhiye, K. Niinuma, Synthetic video generation for weakly supervised cross-domain video anomaly detection, in: International Conference on Pattern Recognition, Springer, 2024, pp. 375–391

    2024

  2. [2]

    Zhang, J

    M. Zhang, J. Wang, Q. Qi, H. Sun, Z. Zhuang, P. Ren, R. Ma, J. Liao, Multi-scale video anomaly detection by multi-grained spatio-temporal representation learning, in: 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, 2024, pp. 17385–17394

    2024

  3. [3]

    J. Ahn, J. Park, S. S. Lee, K.-H. Lee, H. Do, J. Ko, Safefac: Video-based smart safety monitoring for preventing industrial work accidents, Expert Systems with Applications 215 (2023) 119397

    2023

  4. [4]

    Z. Chen, H. Wang, C. Li, C. Liu, F. Yang, D. Zhang, A. J. Fauci, J. Zhang, Large language models in traditional chinese medicine: a systematic review, Acupuncture and Herbal Medicine 5 (1) (2025) 57– 67

    2025

  5. [5]

    X. Liu, T. Gong, Artificial intelligence and evidence-based research will promote the development of traditional medicine, Acupuncture and Herbal Medicine 4 (1) (2024) 134–135

    2024

  6. [6]

    Bogdoll, M

    D. Bogdoll, M. Nitsche, J. M. Z ¨ollner, Anomaly detection in autonomous driving: A survey, in: 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), IEEE, 2022, pp. 4487– 4498

    2022

  7. [7]

    F. U. M. Ullah, M. S. Obaidat, A. Ullah, K. Muhammad, M. Hijji, S. W. Baik, A comprehensive review on vision-based violence detection in surveillance videos, ACM Computing Surveys 55 (10) (2023) 1–44

    2023

  8. [8]

    P. Zhu, X. Wang, Y . Luo, Z. Sun, W.-S. Zheng, Y . Wang, C. Chen, Unpaired image captioning by image-level weakly-supervised visual concept recognition, IEEE Transactions on Multimedia 25 (2022) 6702– 6716

    2022

Show all 88 references
  1. [9]

    Z. Ren, L. He, P. Zhu, Super-resolution learning strategy based on expert knowledge supervision, Remote Sensing 16 (16) (2024) 2888

    2024

  2. [10]

    F. Liu, Y . Wen, J. Sun, P. Zhu, L. Mao, G. Niu, J. Li, Iterative mamba diffusion change-detection model for remote sensing, Remote Sensing 16 (19) (2024) 3651

    2024

  3. [11]

    Y . Bao, H. Ding, Z. Zhang, K. Yang, Q. Tran, Q. Sun, T. Xu, Intelligent acupuncture: data-driven revolution of traditional chinese medicine, Acupuncture and Herbal Medicine 3 (4) (2023) 271–284

    2023

  4. [12]

    Zhong, G

    Y . Zhong, G. Yan, Y . Hu, D. Zhu, R. Zhu, A two-stage framework with memory for anomaly detection via video decomposition and bidirectional consistency, IEEE Transactions on Circuits and Systems for Video Technology (2025)

    2025

  5. [13]

    Xie, Y .-W

    X.-D. Xie, Y .-W. Zhan, Z.-X. Ma, H.-M. Liu, Z.-D. Chen, X. Luo, X.-S. Xu, Distributed learning for privacy-preserving semi-supervised video anomaly detection, IEEE Transactions on Circuits and Systems for Video Technology (2025)

    2025

  6. [14]

    Zhong, R

    Y . Zhong, R. Zhu, G. Yan, P. Gan, X. Shen, D. Zhu, Inter-clip feature similarity based weakly supervised video anomaly detection via multi- scale temporal mlp, IEEE Transactions on Circuits and Systems for Video Technology (2024)

    2024

  7. [15]

    Gehrig, D

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

    2024

  8. [16]

    Z. Wu, M. Gehrig, Q. Lyu, X. Liu, I. Gilitschenski, Leod: Label-efficient object detection for event cameras, in: 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, 2024, pp. 16933–16942

    2024

  9. [17]

    X. Wang, S. Wang, C. Tang, L. Zhu, B. Jiang, Y . Tian, J. Tang, Event stream-based visual object tracking: A high-resolution benchmark dataset and a novel baseline, in: Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition, 2024, pp. 19248– 19257

    2024

  10. [18]

    Gallego, T

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

    2020

  11. [19]

    Y . Qian, S. Ye, C. Wang, X. Cai, J. Qian, J. Wu, Ucf-crime-dvs: A novel event-based dataset for video anomaly detection with spiking neural networks, in: Proceedings of the AAAI Conference on Artificial Intelligence, V ol. 39, 2025, pp. 6577–6585

    2025

  12. [20]

    Flaborea, L

    A. Flaborea, L. Collorone, G. M. D. Di Melendugno, S. D’Arrigo, B. Prenkaj, F. Galasso, Multimodal motion conditioned diffusion model for skeleton-based video anomaly detection, in: Proceedings of the IEEE/CVF international conference on computer vision, 2023, pp. 10318–10329

    2023

  13. [21]

    G. Wang, Y . Wang, J. Qin, D. Zhang, X. Bao, D. Huang, Video anomaly detection by solving decoupled spatio-temporal jigsaw puzzles, in: European Conference on Computer Vision, Springer, 2022, pp. 494– 511

    2022

  14. [22]

    H. Liu, L. He, M. Zhang, F. Li, Vadiffusion: Compressed domain information guided conditional diffusion for video anomaly detection, IEEE Transactions on Circuits and Systems for Video Technology 34 (9) (2024) 8398–8411

    2024

  15. [23]

    C. Guo, L. Li, Y . Ren, X. Zhang, G. Feng, Aligning normal representa- tions in diffusion model for video anomaly detection, IEEE Transactions on Circuits and Systems for Video Technology (2025)

    2025

  16. [24]

    C. Guo, H. Wang, Y . Xia, G. Feng, Learning appearance-motion synergy via memory-guided event prediction for video anomaly detection, IEEE Transactions on Circuits and Systems for Video Technology 34 (3) (2023) 1519–1531

    2023

  17. [25]

    Z. Yang, J. Liu, P. Wu, Text prompt with normality guidance for weakly supervised video anomaly detection, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2024, pp. 18899– 18908

    2024

  18. [26]

    M. I. Georgescu, R. T. Ionescu, F. S. Khan, M. Popescu, M. Shah, A background-agnostic framework with adversarial training for abnormal 13 event detection in video, IEEE transactions on pattern analysis and machine intelligence 44 (9) (2021) 4505–4523

    2021

  19. [27]

    Georgescu, A

    M.-I. Georgescu, A. Barbalau, R. T. Ionescu, F. S. Khan, M. Popescu, M. Shah, Anomaly detection in video via self-supervised and multi- task learning, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2021, pp. 12742–12752

    2021

  20. [28]

    P. Wu, X. Zhou, G. Pang, Y . Sun, J. Liu, P. Wang, Y . Zhang, Open- vocabulary video anomaly detection, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 18297–18307

    2024

  21. [29]

    X. Wang, Y . Jin, W. Wu, W. Zhang, L. Zhu, B. Jiang, Y . Tian, Object detection using event camera: A moe heat conduction based detector and a new benchmark dataset, in: Proceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 29321–29330

    2025

  22. [30]

    Chakravarthi, A

    B. Chakravarthi, A. A. Verma, K. Daniilidis, C. Fermuller, Y . Yang, Recent event camera innovations: A survey, in: European Conference on Computer Vision, Springer, 2024, pp. 342–376

    2024

  23. [31]

    C. Feng, W. Yu, X. Cheng, Z. Tang, J. Zhang, L. Yuan, Y . Tian, Ae-nerf: Augmenting event-based neural radiance fields for non-ideal conditions and larger scenes, in: Proceedings of the AAAI Conference on Artificial Intelligence, V ol. 39, 2025, pp. 2924–2932

    2025

  24. [32]

    Y . Yang, L. Pan, L. Liu, Event camera data pre-training, in: Proceedings of the IEEE/CVF international conference on computer vision, 2023, pp. 10699–10709

    2023

  25. [33]

    Zubic, M

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

    2024

  26. [34]

    P. Zhu, X. Wang, L. Zhu, Z. Sun, W.-S. Zheng, Y . Wang, C. Chen, Prompt-based learning for unpaired image captioning, IEEE Transac- tions on Multimedia 26 (2023) 379–393

    2023

  27. [35]

    Radford, J

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clark, et al., Learning transferable visual models from natural language supervision, in: International con- ference on machine learning, PmLR, 2021, pp. 8748–8763

    2021

  28. [36]

    Chen, D.-Z

    F.-L. Chen, D.-Z. Zhang, M.-L. Han, X.-Y . Chen, J. Shi, S. Xu, B. Xu, Vlp: A survey on vision-language pre-training, Machine Intelligence Research 20 (1) (2023) 38–56

    2023

  29. [37]

    Zhang, Z

    R. Zhang, Z. Guo, W. Zhang, K. Li, X. Miao, B. Cui, Y . Qiao, P. Gao, H. Li, Pointclip: Point cloud understanding by clip, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 8552–8562

    2022

  30. [38]

    S. T. Wasim, M. Naseer, S. Khan, F. S. Khan, M. Shah, Vita-clip: Video and text adaptive clip via multimodal prompting, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 23034–23044

    2023

  31. [39]

    Zhang, Z

    R. Zhang, Z. Zeng, Z. Guo, Y . Li, Can language understand depth?, in: Proceedings of the 30th ACM International Conference on Multimedia, 2022, pp. 6868–6874

    2022

  32. [40]

    Z. Wu, X. Liu, I. Gilitschenski, Eventclip: Adapting clip for event-based object recognition, arXiv preprint arXiv:2306.06354 (2023)

    2023 arXiv

  33. [41]

    J. Zhou, X. Zheng, Y . Lyu, L. Wang, Eventbind: Learning a unified representation to bind them all for event-based open-world understand- ing, in: European Conference on Computer Vision, Springer, 2024, pp. 477–494

    2024

  34. [42]

    Speer, C

    R. Speer, C. Havasi, Conceptnet 5: A large semantic network for relational knowledge, in: The People’s Web Meets NLP: Collaboratively Constructed Language Resources, Springer, 2013, pp. 161–176

    2013

  35. [43]

    Y . Fan, Y . Yu, W. Lu, Y . Han, Weakly-supervised video anomaly de- tection with snippet anomalous attention, IEEE Transactions on Circuits and Systems for Video Technology 34 (7) (2024) 5480–5492

    2024

  36. [44]

    S. Yan, R. Zhang, Z. Guo, W. Chen, W. Zhang, H. Li, Y . Qiao, H. Dong, Z. He, P. Gao, Referred by multi-modality: A unified temporal transformer for video object segmentation, in: Proceedings of the AAAI Conference on Artificial Intelligence, V ol. 38, 2024, pp. 6449–6457

    2024

  37. [45]

    Y . Pu, X. Wu, L. Yang, S. Wang, Learning prompt-enhanced context features for weakly-supervised video anomaly detection, IEEE Transac- tions on Image Processing (2024)

    2024

  38. [46]

    Z.-Y . Hu, Y . Zhong, S. Huang, M. Lyu, L. Wang, Enhancing temporal modeling of video llms via time gating, in: Findings of the Association for Computational Linguistics: EMNLP 2024, 2024, pp. 2845–2856

    2024

  39. [47]

    P. Wu, J. Liu, Learning causal temporal relation and feature discrimina- tion for anomaly detection, IEEE Transactions on Image Processing 30 (2021) 3513–3527

    2021

  40. [48]

    Sultani, C

    W. Sultani, C. Chen, M. Shah, Real-world anomaly detection in surveil- lance videos, in: Proceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 6549–6558

    2018

  41. [49]

    A. Adam, E. Rivlin, I. Shimshoni, R. Reinitz, Robust real-time unusual event detection using multiple fixed-location monitors, IEEE Transac- tions on Pattern Analysis and Machine Intelligence 30 (3) (2008) 555– 560

    2008

  42. [50]

    C. Lu, J. Shi, J. Jia, Abnormal event detection at 150 fps in matlab, in: Proceedings of the IEEE International Conference on Computer Vision, 2013, pp. 2720–2727

    2013

  43. [51]

    Mahadevan, W

    V . Mahadevan, W. Li, V . Bhalodia, N. Vasconcelos, Anomaly detection in crowded scenes, in: 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, IEEE, 2010, pp. 1976–1983

    2010

  44. [52]

    P. Wu, J. Liu, Y . Shi, Y . Sun, F. Shao, Z. Wu, Z. Yang, Not only look, but also listen: Learning multimodal violence detection under weak supervision, in: European conference on computer vision, Springer, 2020, pp. 322–339

    2020

  45. [53]

    A. Amir, B. Taba, D. Berg, T. Melano, J. McKinstry, C. Di Nolfo, T. Nayak, A. Andreopoulos, G. Garreau, M. Mendoza, et al., A low power, fully event-based gesture recognition system, in: Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 7243–7252

    2017

  46. [54]

    Y . Bi, A. Chadha, A. Abbas, E. Bourtsoulatze, Y . Andreopoulos, Graph-based object classification for neuromorphic vision sensing, in: Proceedings of the IEEE/CVF international conference on computer vision, 2019, pp. 491–501

    2019

  47. [55]

    Y . Dong, Y . Li, D. Zhao, G. Shen, Y . Zeng, Bullying10k: a large-scale neuromorphic dataset towards privacy-preserving bullying recognition, Advances in Neural Information Processing Systems 36 (2023) 1923– 1937

    2023

  48. [56]

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

    2018

  49. [57]

    S. Lin, Y . Ma, Z. Guo, B. Wen, Dvs-voltmeter: Stochastic process-based event simulator for dynamic vision sensors, in: European Conference on Computer Vision, Springer, 2022, pp. 578–593

    2022

  50. [58]

    Zhang, L

    J. Zhang, L. Qing, J. Miao, Temporal convolutional network with complementary inner bag loss for weakly supervised anomaly detection, in: 2019 IEEE International Conference on Image Processing (ICIP), IEEE, 2019, pp. 4030–4034

    2019

  51. [59]

    Zhong, N

    J.-X. Zhong, N. Li, W. Kong, S. Liu, T. H. Li, G. Li, Graph convolutional label noise cleaner: Train a plug-and-play action classifier for anomaly detection, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2019, pp. 1237–1246

    2019

  52. [60]

    M. Z. Zaheer, A. Mahmood, M. Astrid, S.-I. Lee, Claws: Clustering assisted weakly supervised learning with normalcy suppression for anomalous event detection, in: European Conference on Computer Vision, Springer, 2020, pp. 358–376

    2020

  53. [61]

    Feng, F.-T

    J.-C. Feng, F.-T. Hong, W.-S. Zheng, Mist: Multiple instance self- training framework for video anomaly detection, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2021, pp. 14009–14018

    2021

  54. [62]

    Y . Tian, G. Pang, Y . Chen, R. Singh, J. W. Verjans, G. Carneiro, Weakly-supervised video anomaly detection with robust temporal feature magnitude learning, in: Proceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 4975–4986

    2021

  55. [63]

    S. Li, F. Liu, L. Jiao, Self-training multi-sequence learning with trans- former for weakly supervised video anomaly detection, in: Proceedings of the AAAI Conference on Artificial Intelligence, V ol. 36, 2022, pp. 1395–1403

    2022

  56. [64]

    S. Park, H. Kim, M. Kim, D. Kim, K. Sohn, Normality guided multiple instance learning for weakly supervised video anomaly detection, in: Proceedings of the IEEE/CVF winter conference on applications of computer vision, 2023, pp. 2665–2674

    2023

  57. [65]

    Wu, H.-Y

    J.-C. Wu, H.-Y . Hsieh, D.-J. Chen, C.-S. Fuh, T.-L. Liu, Self-supervised sparse representation for video anomaly detection, in: European Confer- ence on Computer Vision, Springer, 2022, pp. 729–745

    2022

  58. [66]

    M. Cho, M. Kim, S. Hwang, C. Park, K. Lee, S. Lee, Look around for anomalies: Weakly-supervised anomaly detection via context-motion relational learning, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2023, pp. 12137–12146

    2023

  59. [67]

    B. Wan, Y . Fang, X. Xia, J. Mei, Weakly supervised video anomaly detection via center-guided discriminative learning, in: 2020 IEEE International Conference on Multimedia and Expo (ICME), IEEE, 2020, pp. 1–6

    2020

  60. [68]

    H. Lv, Z. Yue, Q. Sun, B. Luo, Z. Cui, H. Zhang, Unbiased multiple instance learning for weakly supervised video anomaly detection, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2023, pp. 8022–8031. 14

    2023

  61. [69]

    T. Zhu, Q. Yu, X. Dong, S. Li, Y . Liu, J. Jiang, L. Shu, Prodisc-vad: An efficient system for weakly-supervised anomaly detection in video surveillance applications, arXiv e-prints (2025) arXiv–2505

    2025

  62. [70]

    Y . Zhu, S. Newsam, Motion-aware feature for improved video anomaly detection, arXiv preprint arXiv:1907.10211 (2019)

    1907 arXiv

  63. [71]

    Y . Zhen, Y . Guo, J. Wei, X. Bao, D. Huang, Multi-scale background suppression anomaly detection in surveillance videos, in: 2021 IEEE International Conference on Image Processing (ICIP), IEEE, 2021, pp. 1114–1118

    2021

  64. [72]

    Y . Pu, X. Wu, Locality-aware attention network with discriminative dynamics learning for weakly supervised anomaly detection, in: 2022 IEEE International Conference on Multimedia and Expo (ICME), IEEE, 2022, pp. 1–6

    2022

  65. [73]

    Zhang, G

    C. Zhang, G. Li, Q. Xu, X. Zhang, L. Su, Q. Huang, Weakly supervised anomaly detection in videos considering the openness of events, IEEE transactions on intelligent transportation systems 23 (11) (2022) 21687– 21699

    2022

  66. [74]

    Zhang, G

    C. Zhang, G. Li, Y . Qi, S. Wang, L. Qing, Q. Huang, M.-H. Yang, Exploiting completeness and uncertainty of pseudo labels for weakly supervised video anomaly detection, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 16271–16280

    2023

  67. [75]

    P. Wu, X. Zhou, G. Pang, L. Zhou, Q. Yan, P. Wang, Y . Zhang, Vadclip: Adapting vision-language models for weakly supervised video anomaly detection (2024)

    2024

  68. [76]

    Liu, K.-M

    T. Liu, K.-M. Lam, B.-K. Bao, Injecting text clues for improving anoma- lous event detection from weakly labeled videos, IEEE Transactions on Image Processing (2024)

    2024

  69. [77]

    Y . Chen, Z. Liu, B. Zhang, W. Fok, X. Qi, Y .-C. Wu, Mgfn: Magnitude-contrastive glance-and-focus network for weakly-supervised video anomaly detection, in: Proceedings of the AAAI conference on artificial intelligence, V ol. 37, 2023, pp. 387–395

    2023

  70. [78]

    H. K. Joo, K. V o, K. Yamazaki, N. Le, Clip-tsa: Clip-assisted temporal self-attention for weakly-supervised video anomaly detection, in: 2023 IEEE International Conference on Image Processing (ICIP), IEEE, 2023, pp. 3230–3234

    2023

  71. [79]

    Y . Zhou, Y . Qu, X. Xu, F. Shen, J. Song, H. T. Shen, Batchnorm- based weakly supervised video anomaly detection, IEEE Transactions on Circuits and Systems for Video Technology (2024)

    2024

  72. [80]

    H. Zhou, J. Yu, W. Yang, Dual memory units with uncertainty regulation for weakly supervised video anomaly detection, in: Proceedings of the AAAI Conference on Artificial Intelligence, V ol. 37, 2023, pp. 3769– 3777

    2023

  73. [81]

    S. Tu, Q. Dai, Z. Wu, Z.-Q. Cheng, H. Hu, Y .-G. Jiang, Implicit temporal modeling with learnable alignment for video recognition, in: Proceedings of the ieee/cvf international conference on computer vision, 2023, pp. 19936–19947

    2023

  74. [82]

    Y . Tang, L. Qi, F. Xie, X. Li, C. Ma, M.-H. Yang, Video predic- tion transformers without recurrence or convolution, in: arXiv preprint arxiv:2410.04733, 2024

    2024

  75. [83]

    H. Lu, G. Yang, N. Fei, Y . Huo, Z. Lu, P. Luo, M. Ding, Vdt: General- purpose video diffusion transformers via mask modeling, arXiv preprint arXiv:2305.13311 (2023)

    2023 arXiv

  76. [84]

    S. T. Wasim, M. U. Khattak, M. Naseer, S. Khan, M. Shah, F. S. Khan, Video-focalnets: Spatio-temporal focal modulation for video action recognition, in: Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 13778–13789

    2023

  77. [85]

    S. Yang, Y . Zhou, Z. Liu, C. C. Loy, Fresco: Spatial-temporal correspon- dence for zero-shot video translation, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 8703–8712

    2024

  78. [86]

    L. Qian, J. Li, Y . Wu, Y . Ye, H. Fei, T.-S. Chua, Y . Zhuang, S. Tang, Momentor: advancing video large language model with fine-grained temporal reasoning, in: Proceedings of the 41st International Conference on Machine Learning, 2024, pp. 41340–41356

    2024

  79. [87]

    L. v. d. Maaten, G. Hinton, Visualizing data using t-sne, Journal of machine learning research 9 (Nov) (2008) 2579–2605. Peipei Zhureceived the B.E. degree from Hebei University of Technology, Tianjin, China, in 2014, the M.E. degree from Beijing University of Posts and Teleco...

    2008

  80. [88]

    He is currently an Associate Professor at Guangzhou Institute of Technology, Xi- dian University

    From January to March 2020, he was an Intern at Bell Labs France, where he worked on indoor localization systems. He is currently an Associate Professor at Guangzhou Institute of Technology, Xi- dian University. His research interests include UA V– UGV cooperative systems, mil...

    2020

Pith tools

Reviewed July 13, 2026 · model on record in the stance chip above.