Pith. sign in

REVIEW 2 major objections 5 minor 43 references

Just Dance with $\pi$! A Poly-modal Inductor for Weakly-supervised Video Anomaly Detection

T0 review · 2 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 PI-VAD shows that a video anomaly detector can be trained on five auxiliary modalities and then run on RGB alone, reaching state-of-the-art scores on UCF-Crime, XD-Violence, and MSAD.

desk verdict Solid multi-modal WSVAD paper with real but possibly over-interpreted gains; the missing random-target control is the key thing a referee should demand. read the letter →

arxiv 2505.13123 v1 pith:HDBZFPER submitted 2025-05-19 cs.CV cs.AIcs.LG

Snehashis Majhi , Giacomo D'Amicantonio , Antitza Dantcheva , Quan Kong , Lorenzo Garattoni , Gianpiero Francesca , Egor Bondarev , Francois Bremond This is my paper · ORCID
classification cs.CVcs.AIcs.LG
keywords weakly-supervisedvideoanomalydetectionmulti-modalrepresentationlearningpseudo-modalitygenerationcross-modalinductionRGB-onlyinferencebenchmarksteacher-studentdistillation
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 argues that weakly-supervised video anomaly detection is limited by relying on RGB appearance alone, because subtle anomalies such as shoplifting can look similar to normal actions. To fix this, PI-VAD trains a video transformer with five extra modalities—pose, depth, panoptic masks, optical flow, and text—and at the same time learns to synthesize those modalities from RGB features. The extra modality backbones are used only during training; at inference the model needs only RGB and still reports top accuracy on UCF-Crime, XD-Violence, and MSAD. If correct, this means multi-modal training can enrich an RGB detector without the cost of running several heavy networks per video.

What carries the argument

The load-bearing mechanism is the Poly-modal Inductor (PI), a training-only plug-in placed at early and late transformer blocks of a teacher-student architecture. It has two parts: PMG, a shared encoder with five decoders that predicts each modality's embedding from the student's RGB features under an MSE reconstruction loss, and CMI, which applies a snippet-level bi-directional InfoNCE contrastive loss between each pseudo-modality and RGB, fuses the aligned embeddings with transformer blocks, and distills the fused output toward the frozen teacher using an MSE distillation loss. PI is what lets the heavy modality encoders be removed at inference while their signal remains present in the RGB student.

What would settle it

Take the trained PI-VAD and, at inference, replace the generated pseudo-modalities with random noise of the same shape; if UCF-Crime AUC stays close to 90.33, the gains come from the alignment and distillation machinery rather than from modality content. A second test is to train on UCF-Crime and evaluate on MSAD without retuning; if accuracy falls to the level of the RGB-only teacher on MSAD, the pseudo-modalities are not transferring across domains.

Watch

Extended reading notes

Core claim

The central claim is that the latent embeddings of five auxiliary modalities can be approximated from RGB features well enough to improve weakly-supervised anomaly detection. A Pseudo-modality Generation (PMG) module maps the student's intermediate RGB representation into five pseudo-embeddings by reconstructing the outputs of pretrained pose, depth, panoptic, flow, and text models. A Cross Modal Induction (CMI) module aligns each pseudo-modality to its RGB snippet with a contrastive loss, lets transformer blocks select the relevant cues, and distills the result toward a frozen teacher's anomaly-aware features. With this training-only module inserted at early and late stages of the student, the method reaches UCF-Crime AUC 90.33, XD-Violence AP 85.37, and MSAD AUC 88.68 while processing about 30 frames per second at inference. The paper's ablation also shows that using the true modality features instead of generated pseudo-modalities raises UCF-Crime AUC by only about 0.25 points, indicating the reconstruction is nearly lossless.

Load-bearing premise

The load-bearing premise is that a small network can reproduce each auxiliary modality's learned representation from RGB features closely enough to keep the anomaly signal; if the reproduced versions lose too much, the reported gains shrink toward the RGB-only baseline.

Editorial extensions

If this is right

  • The method needs only video-level labels, so the same pseudo-modality recipe applies to other weakly supervised video tasks without frame-level annotation.
  • At test time the model runs at about 30 frames per second, making it plausible for live surveillance, whereas running all five modality backbones would be far more expensive.
  • Because different anomalies activate different modalities, the framework can also serve as an explanation tool: a high pose-modality activation flags human-centered anomalies, while text and panoptic activations flag scene-level events.
  • The paper's sequential-modality ablation improves accuracy at every step, suggesting the poly-modal contributions are complementary rather than redundant.
  • Reconstruction costs only about 0.25 AUC compared with using true modality features, so the efficiency gain comes at a very small measured accuracy cost on UCF-Crime.

Reading between the lines

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

  • Inference: if the 0.25 AUC reconstruction gap is stable across datasets, the same pseudo-modality trick could transfer to other label-sparse video tasks such as temporal action localization, where multi-modal ground truth is even scarcer.
  • Inference: the class-wise results suggest a budget-driven choice a practitioner could test: two or three modalities such as depth, text, and pose may capture most of the gain for human-centered anomaly categories.
  • Inference: the method's ceiling is set by the frozen per-modality backbones; swapping those backbones for stronger or surveillance-tuned versions is a direct test of how much headroom remains.
  • Inference: because text and panoptic masks dominate scene-based anomalies while pose and depth dominate human-based ones, a modular deployment could select modalities according to the anomaly types a given camera site prioritizes.
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

2 major / 5 minor

Summary. This paper introduces PI-VAD, a weakly-supervised video anomaly detection framework that augments RGB snippet features with five pseudo-modalities (pose, depth, panoptic masks, optical flow, and text) which are generated from the RGB features themselves by a Pseudo Modality Generation (PMG) module and integrated by a Cross Modal Induction (CMI) module through contrastive alignment and distillation from a frozen UR-DMU teacher. The modality backbones are used only during training; at inference, only the student RGB encoder and the lightweight PI modules are needed. Experiments on UCF-Crime, XD-Violence, and MSAD report high AUC/AP values, and ablations show contributions from the alignment and distillation losses and from individual modalities. The paper claims state-of-the-art accuracy while maintaining RGB-only inference and reports a favorable computational profile.

Significance. The central idea is attractive and practically relevant: frozen external modality embeddings can serve as auxiliary training targets that transfer multi-modal knowledge into a model that is RGB-only at inference. The paper is commendable for releasing code, for ablating the two proposed modules (Table 3), and for providing a per-modality analysis (Table 6 and Figures 4-6). The efficiency result (30 FPS, 19.88 GFLOPs versus 2,561.40 GFLOPs for the five modality backbones) is a concrete practical strength. However, the state-of-the-art claim is weakened by a higher-scoring multi-modal method in the paper's own comparison table, and the causal role of the specific modality targets is not isolated by a control experiment. If these two issues are resolved, the work would constitute a solid contribution to WSVAD.

major comments (2)
  1. [§5.1, Table 1] The paper claims in §5.1 that 'Compared with current multi-modal SoTA approaches, π-VAD demonstrates superior capabilities in both UCF-Crime and XD-Violence datasets,' and the abstract claims unqualified 'state-of-the-art accuracy' on three datasets. However, Table 1 lists PEMIL (I3D+Text) under 'SoTA with multi-modality at inference' with an XD-Violence AP of 88.21, which is 2.84 points higher than the reported π-VAD AP of 85.37. This directly contradicts the stated claim. The authors should either benchmark against PEMIL under the same protocol, explicitly explain why PEMIL is excluded from the comparison, or restrict the claim to 'state-of-the-art among methods with RGB-only inference.' As written, the headline contribution is not supported by the reported numbers.
  2. [§4.2-4.3, Table 3] The load-bearing interpretation of the paper is that the five external modalities contribute salient anomaly-relevant cues beyond RGB. Yet no ablation replaces the modality targets with non-informative controls of the same shape, such as random targets, shuffled targets, or targets from a fixed untrained encoder. Table 3 shows that the full CMI pipeline (row 4) reaches 90.58 AUC versus 86.97 for the UR-DMU baseline, but Table 6 shows each individual modality improves AUC by only 0.68-0.95 points, and the differences between modalities are small. Without a control, the large CMI gain could be due to the extra transformer blocks, the contrastive alignment, and the distillation loss rather than to the semantic content of the pose/depth/panoptic/flow/text targets. Adding such a control is necessary to support the central claim that RGB features are insufficient and that the five modalities add salient cues.
minor comments (5)
  1. [§4.2] The text states that PMG has 'one encoder and five parallel decoders' and then says 'The six decoders operate in a mutually exclusive manner'; this count is inconsistent and should be corrected.
  2. [Eq. (2)] The InfoNCE denominator sums over k≠i, which omits the positive pair, and the temperature τ appears only in the numerator and not in the denominator. This is inconsistent with the standard InfoNCE formulation described in the text and should be fixed.
  3. [§4.4, Eq. (6)] The hyperparameters λ1, λ2, and τ are introduced but their values are not reported in the main text; Section A is referenced but not included in the manuscript text provided. The authors should ensure all training details, including these values, are fully specified in the appendix or supplement.
  4. [Table 6, §6.1] The text-only configuration in Table 6 yields AUC_A of 69.45, which is below the RGB baseline's 70.81. This contradicts the unqualified statement in §6.1 that 'each modality is able to enhance the RGB features' and the conclusion's claim that all five modalities are useful; the statement should be qualified to mention the AUC_A regression for text.
  5. [Tables 1-3, 6] No error bars or multiple-seed statistics are reported. Given that some comparisons in Table 3 differ by only 0.25 AUC points, reporting mean±std over at least three runs would strengthen the conclusions.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the pseudo-modalities are anchored to external frozen backbones and the reported gains are held-out evaluations, not fitted predictions.

full rationale

The paper's derivation chain is not circular. PMG (Sec. 4.2) regresses student RGB features onto intermediate embeddings of five frozen, externally pretrained backbones (YOLOv7-pose, DepthAnythingV2, SAM, RAFT, ViFiCLIP) via L_PMG in Eq. (1); these targets are fixed, not produced by the model itself. CMI's InfoNCE alignment (Eq. 2) does relate pseudo-modalities back to the same RGB stream, but those pseudo-modalities are simultaneously constrained by Eq. (1) to match external targets, so the loop is anchored by outside information rather than being a free self-consistency condition. The distillation loss (Eq. 4) uses a frozen UR-DMU teacher, and the student's anomaly scores are evaluated on standard held-out test protocols, so the reported AUC/AP numbers are genuine empirical outcomes rather than quantities recovered from the training targets by construction. Table 3 directly tests the pseudo-modality assumption: replacing pseudo-modalities with true backbone features changes AUC by only 0.25 points (90.33 vs 90.58), which is evidence about reconstruction quality, not circularity. The absence of a random-target control for the five modality embeddings is a legitimate scientific caveat about interpretation, but under the stated rules it is not circular because no equation or fitted parameter is being renamed as a prediction. Author self-citations appear only in related-work positions and are not load-bearing. Accordingly, the paper warrants a circularity score of 0.

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

The central claim rests on standard hyperparameters and on the reliability of external pretrained encoders and a pretrained teacher. No fundamentally new physical or model entities are introduced; pseudo-modalities are generated latent embeddings, not new sensors or signals.

free parameters (3)
  • lambda_1
    Weight of the alignment loss in the final objective (Eq. 6); no value or tuning range is reported in the main text.
  • lambda_2
    Weight of the distillation loss in Eq. 6; value not reported in the main text.
  • InfoNCE temperature tau
    Temperature in the contrastive loss (Eq. 2); not reported in the main text.
assumptions (3)
  • domain assumption Intermediate embeddings of YOLOv7-pose, DepthAnythingV2, SAM, RAFT, and ViFiCLIP are reliable, task-relevant targets for anomaly detection.
    Used as ground truth in Eq. (1). If these encoders are noisy or fail on surveillance scenes, the pseudo-modalities lose value.
  • domain assumption The pretrained UR-DMU teacher's features at early and late blocks are a good distillation target.
    The student is aligned to these features via Eq. (4); a poor teacher would bound the student.
  • domain assumption Same-snippet RGB and pseudo-modality embeddings should be contrastive positives, and different-snippet embeddings negatives.
    This pairing in Eq. (2) may be violated for visually similar normal snippets, but serves as an auxiliary objective.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Just Dance with $\pi$! A Poly-modal Inductor for Weakly-supervised Video Anomaly Detection." pith.science (2026). https://pith.science/paper/HDBZFPER

@misc{pith2026250513123,
  author       = {Pith},
  title        = {Pith review of: Just Dance with $\pi$! A Poly-modal Inductor for Weakly-supervised Video Anomaly Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HDBZFPER}},
  note         = {Machine review of arXiv:2505.13123}
}
read the original abstract

Weakly-supervised methods for video anomaly detection (VAD) are conventionally based merely on RGB spatio-temporal features, which continues to limit their reliability in real-world scenarios. This is due to the fact that RGB-features are not sufficiently distinctive in setting apart categories such as shoplifting from visually similar events. Therefore, towards robust complex real-world VAD, it is essential to augment RGB spatio-temporal features by additional modalities. Motivated by this, we introduce the Poly-modal Induced framework for VAD: "PI-VAD", a novel approach that augments RGB representations by five additional modalities. Specifically, the modalities include sensitivity to fine-grained motion (Pose), three dimensional scene and entity representation (Depth), surrounding objects (Panoptic masks), global motion (optical flow), as well as language cues (VLM). Each modality represents an axis of a polygon, streamlined to add salient cues to RGB. PI-VAD includes two plug-in modules, namely Pseudo-modality Generation module and Cross Modal Induction module, which generate modality-specific prototypical representation and, thereby, induce multi-modal information into RGB cues. These modules operate by performing anomaly-aware auxiliary tasks and necessitate five modality backbones -- only during training. Notably, PI-VAD achieves state-of-the-art accuracy on three prominent VAD datasets encompassing real-world scenarios, without requiring the computational overhead of five modality backbones at inference.

Figures

Figures reproduced from arXiv: 2505.13123 by the authors.

Figure 1
Figure 1. a): Illustration of abnormal frames and respective multi-modal saliencies in complex real-world scenes. Optical flow captures distinct abnormal motion in “Abuse” and “Arrest”, while depth and pose detect subtle movements that optical flow may miss. Panoptic masks and text provide overall scene context. b): Comparison of multi-modal methods with our PI-VAD. PI-VAD requires the five modalities only during training, si… view at source ↗
Figure 2
Figure 2. (a) Overview of Poly-modal Induced VAD (π-VAD): In the training phase, π-VAD uses a teacher-student approach, where a poly-modal inductor enhances the student’s RGB representation by generating and associating five distinct modalities. Note that the teacher’s weights remain fixed during training. At inference, the student and poly-modal inductor operate independently to detect video anomalies. (b) Poly-modal Inducto… view at source ↗
Figure 3
Figure 3. Class-wise [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Visualization of sample frames and ground truth (green shed) vs. prediction scores (red shed) for various cases in Row-1 and [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Class-wise AUC comparison between the RGB model and RGB with one additional modality model on UCF-Crime. in [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Comparison between the AUC scores of different mix￾tures of modalities in the π-VAD framework for the UCF-Crime dataset. In red, we highlight the classes on which the modalities have contrastive features, in green the classes where the modalities are complementary. Gen…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 27 canonical work pages

  1. [1]

    Quo vadis, action recognition? A new model and the kinetics dataset

    Jo ˜ao Carreira and Andrew Zisserman. Quo vadis, action recognition? A new model and the kinetics dataset. CoRR, abs/1705.07750, 2017. 3

    arXiv 2017

  2. [2]

    Prompt-enhanced multiple instance learning for weakly supervised video anomaly detection

    Junxi Chen, Liang Li, Li Su, Zheng-jun Zha, and Qing- ming Huang. Prompt-enhanced multiple instance learning for weakly supervised video anomaly detection. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 18319–18329, 2024. 6

    work page 2024

  3. [3]

    Tevad: Improved video anomaly de- tection with captions

    Weiling Chen, Keng Teck Ma, Zi Jian Yew, Minhoe Hur, and David Aik-Aun Khoo. Tevad: Improved video anomaly de- tection with captions. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 5549–5559, 2023. 6

    2023

  4. [4]

    Mgfn: Magnitude- contrastive glance-and-focus network for weakly-supervised video anomaly detection

    Yingxian Chen, Zhengzhe Liu, Baoheng Zhang, Wilton Fok, Xiaojuan Qi, and Yik-Chung Wu. Mgfn: Magnitude- contrastive glance-and-focus network for weakly-supervised video anomaly detection. In Proceedings of the AAAI Con- ference on Artificial Intelligence , pages 387–395, 2023. 2, 6

    work page 2023

  5. [5]

    Look around for anomalies: Weakly-supervised anomaly detection via context-motion relational learning

    MyeongAh Cho, Minjung Kim, Sangwon Hwang, Chae- won Park, Kyungjae Lee, and Sangyoun 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, pages 12137–12146, 2023. 2, 3, 6

    work page 2023

  6. [6]

    Learning an augmented rgb representation with cross-modal knowl- edge distillation for action detection

    Rui Dai, Srijan Das, and Franc ¸ois Bremond. Learning an augmented rgb representation with cross-modal knowl- edge distillation for action detection. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 13053–13064, 2021. 3

    work page 2021

  7. [7]

    Vpn: Learning video-pose embedding for activities of daily living, 2020

    Srijan Das, Saurav Sharma, Rui Dai, Francois Bremond, and Monique Thonnat. Vpn: Learning video-pose embedding for activities of daily living, 2020

    work page 2020

  8. [8]

    Vpn++: Rethinking video-pose embeddings for understand- ing activities of daily living

    Srijan Das, Rui Dai, Di Yang, and Francois Bremond. Vpn++: Rethinking video-pose embeddings for understand- ing activities of daily living. IEEE Transactions on Pat- tern Analysis and Machine Intelligence , 44(12):9703–9717,

    work page

Show all 43 references
  1. [9]

    Imagebind: One embedding space to bind them all

    Rohit Girdhar, Alaaeldin El-Nouby, Zhuang Liu, Mannat Singh, Kalyan Vasudev Alwala, Armand Joulin, and Ishan Misra. Imagebind: One embedding space to bind them all. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 15180–15190, 2023. 2

    2023

  2. [10]

    Clip-tsa: Clip-assisted temporal self-attention for weakly-supervised video anomaly detection

    Hyekang Kevin Joo, Khoa V o, Kashu Yamazaki, and Ngan Le. Clip-tsa: Clip-assisted temporal self-attention for weakly-supervised video anomaly detection. In 2023 IEEE International Conference on Image Processing (ICIP), pages 3230–3234. IEEE, 2023. 6

    2023

  3. [11]

    The kinetics hu- man action video dataset

    Will Kay, Joao Carreira, Karen Simonyan, Brian Zhang, Chloe Hillier, Sudheendra Vijayanarasimhan, Fabio Viola, Tim Green, Trevor Back, Paul Natsev, et al. The kinetics hu- man action video dataset. arXiv preprint arXiv:1705.06950,

    arXiv

  4. [12]

    Segment any- thing

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer White- head, Alexander C Berg, Wan-Yen Lo, et al. Segment any- thing. In Proceedings of the IEEE/CVF International Con- ference on Computer Vision, pages 4015–4026, 2023. 4

    2023

  5. [13]

    Scale-aware spatio-temporal relation learning for video anomaly detection

    Guoqiu Li, Guanxiong Cai, Xingyu Zeng, and Rui Zhao. Scale-aware spatio-temporal relation learning for video anomaly detection. In European Conference on Computer Vision, pages 333–350. Springer, 2022. 6

    2022

  6. [14]

    Self-training multi- sequence learning with transformer for weakly supervised video anomaly detection

    Shuo Li, Fang Liu, and Licheng Jiao. Self-training multi- sequence learning with transformer for weakly supervised video anomaly detection. In Proceedings of the AAAI Con- ference on Artificial Intelligence, pages 1395–1403, 2022. 3, 6

    2022

  7. [15]

    Scaling (down) clip: A comprehensive analysis of data, architecture, and training strategies

    Zichao Li, Cihang Xie, and Ekin Dogus Cubuk. Scaling (down) clip: A comprehensive analysis of data, architecture, and training strategies. arXiv preprint arXiv:2404.08197 ,

    arXiv

  8. [16]

    Social mil: Interaction-aware for crowd anomaly de- tection

    Shuheng Lin, Hua Yang, Xianchao Tang, Tianqi Shi, and Lin Chen. Social mil: Interaction-aware for crowd anomaly de- tection. In 2019 16th IEEE International Conference on Ad- vanced Video and Signal Based Surveillance (AVSS) , pages 1–8. IEEE, 2019. 2

    2019

  9. [17]

    Localizing anomalies from weakly-labeled videos

    Hui Lv, Chuanwei Zhou, Zhen Cui, Chunyan Xu, Yong Li, and Jian Yang. Localizing anomalies from weakly-labeled videos. IEEE transactions on image processing , 30:4505– 4515, 2021. 6

    2021

  10. [18]

    Unbiased multiple instance learning for weakly supervised video anomaly detection

    Hui Lv, Zhongqi Yue, Qianru Sun, Bin Luo, Zhen Cui, and Hanwang Zhang. Unbiased multiple instance learning for weakly supervised video anomaly detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pat- tern Recognition, pages 8022–8031, 2023. 2, 6

    2023

  11. [19]

    Weakly-supervised joint anomaly detection and classification

    Snehashis Majhi, Srijan Das, Franc ¸ois Br ´emond, Ratnakar Dash, and Pankaj Kumar Sa. Weakly-supervised joint anomaly detection and classification. In 2021 16th IEEE International Conference on Automatic Face and Gesture Recognition (FG 2021), pages 1–7. IEEE, 2021. 2

    2021

  12. [20]

    Dam: Dissimilarity attention module for weakly-supervised video anomaly detection

    Snehashis Majhi, Srijan Das, and Franc ¸ois Br´emond. Dam: Dissimilarity attention module for weakly-supervised video anomaly detection. In 2021 17th IEEE International Con- ference on Advanced Video and Signal Based Surveillance (AVSS), pages 1–8, 2021. 2

    2021

  13. [21]

    Human- scene network: A novel baseline with self-rectifying loss for weakly supervised video anomaly detection

    Snehashis Majhi, Rui Dai, Quan Kong, Lorenzo Garat- toni, Gianpiero Francesca, and Francois Bremond. Human- scene network: A novel baseline with self-rectifying loss for weakly supervised video anomaly detection. Computer Vi- sion and Image Understanding, 241:103955, 2024. 6

    2024

  14. [22]

    Repre- sentation learning with contrastive predictive coding

    Aaron van den Oord, Yazhe Li, and Oriol Vinyals. Repre- sentation learning with contrastive predictive coding. arXiv preprint arXiv:1807.03748, 2018. 5

    2018 arXiv

  15. [23]

    Dance with self-attention: A new look of conditional ran- dom fields on anomaly detection in videos

    Didik Purwanto, Yie-Tarng Chen, and Wen-Hsien Fang. Dance with self-attention: A new look of conditional ran- dom fields on anomaly detection in videos. In Proceedings of the IEEE/CVF International Conference on Computer Vi- sion, pages 173–183, 2021. 2

    2021

  16. [24]

    Learning transferable visual models from natural language supervi- sion

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervi- sion. In International conference on machine learning, ...

    2021

  17. [25]

    Fine-tuned clip models are efficient video learners

    Hanoona Rasheed, Muhammad Uzair Khattak, Muhammad Maaz, Salman Khan, and Fahad Shahbaz Khan. Fine-tuned clip models are efficient video learners. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6545–6554, 2023. 3, 4

    2023

  18. [26]

    Just add π! pose induced video transformers for understanding activities of daily liv- ing

    Dominick Reilly and Srijan Das. Just add π! pose induced video transformers for understanding activities of daily liv- ing. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR), pages 18340– 18350, 2024. 3

    2024

  19. [27]

    Real-world anomaly detection in surveillance videos

    Waqas Sultani, Chen Chen, and Mubarak Shah. Real-world anomaly detection in surveillance videos. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 6479–6488, 2018. 2, 5, 6

    2018

  20. [28]

    Raft: Recurrent all-pairs field transforms for optical flow

    Zachary Teed and Jia Deng. Raft: Recurrent all-pairs field transforms for optical flow. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23– 28, 2020, Proceedings, Part II 16, pages 402–419. Springer,

    2020

  21. [29]

    Weakly-supervised video anomaly detection with robust temporal feature magni- tude learning

    Yu Tian, Guansong Pang, Yuanhong Chen, Rajvinder Singh, Johan W Verjans, and Gustavo Carneiro. Weakly-supervised video anomaly detection with robust temporal feature magni- tude learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 4975–49...

    2021

  22. [30]

    Weakly supervised video anomaly detection via center- guided discriminative learning

    Boyang Wan, Yuming Fang, Xue Xia, and Jiajie Mei. Weakly supervised video anomaly detection via center- guided discriminative learning. In 2020 IEEE International Conference on Multimedia and Expo (ICME) , pages 1–6. IEEE, 2020. 2

    2020

  23. [31]

    Yolov7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors

    Chien-Yao Wang, Alexey Bochkovskiy, and Hong- Yuan Mark Liao. Yolov7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors. In Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 7464–7475, 2023. 4

    2023

  24. [32]

    Not only look, but also listen: Learning multimodal violence detection under weak supervision

    Peng Wu, Jing Liu, Yujia Shi, Yujia Sun, Fangtao Shao, Zhaoyang Wu, and Zhiwei Yang. Not only look, but also listen: Learning multimodal violence detection under weak supervision. In European Conference on Computer Vision , pages 322–339. Springer, 2020. 2, 5, 6

    2020

  25. [33]

    Vadclip: Adapting vision-language models for weakly supervised video anomaly detection

    Peng Wu, Xuerong Zhou, Guansong Pang, Lingru Zhou, Qingsen Yan, Peng Wang, and Yanning Zhang. Vadclip: Adapting vision-language models for weakly supervised video anomaly detection. In Proceedings of the AAAI Con- ference on Artificial Intelligence, pages 6074–6082, 2024. 6

    2024

  26. [34]

    Depth any- thing v2

    Lihe Yang, Bingyi Kang, Zilong Huang, Zhen Zhao, Xiao- gang Xu, Jiashi Feng, and Hengshuang Zhao. Depth any- thing v2. arXiv preprint arXiv:2406.09414, 2024. 4

    2024 arXiv

  27. [35]

    Text prompt with nor- mality guidance for weakly supervised video anomaly detec- tion

    Zhiwei Yang, Jing Liu, and Peng Wu. Text prompt with nor- mality guidance for weakly supervised video anomaly detec- tion. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR), pages 18899– 18908, 2024. 6

    2024

  28. [36]

    Modality-aware contrastive instance learning with self-distillation for weakly-supervised audio-visual violence detection

    Jiashuo Yu, Jinyu Liu, Ying Cheng, Rui Feng, and Yuejie Zhang. Modality-aware contrastive instance learning with self-distillation for weakly-supervised audio-visual violence detection. In Proceedings of the 30th ACM international con- ference on multimedia, pages 6278–6287, 2022. 6

    2022

  29. [37]

    A self-reasoning framework for anomaly detection using video-level labels

    Muhammad Zaigham Zaheer, Arif Mahmood, Hochul Shin, and Seung-Ik Lee. A self-reasoning framework for anomaly detection using video-level labels. IEEE Signal Processing Letters, 27:1705–1709, 2020. 2

    2020

  30. [38]

    Exploiting completeness and uncertainty of pseudo labels for weakly supervised video anomaly detection

    Chen Zhang, Guorong Li, Yuankai Qi, Shuhui Wang, Laiyun Qing, Qingming Huang, and Ming-Hsuan 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 R...

    2023

  31. [39]

    Temporal con- volutional network with complementary inner bag loss for weakly supervised anomaly detection

    Jiangong Zhang, Laiyun Qing, and Jun Miao. Temporal con- volutional network with complementary inner bag loss for weakly supervised anomaly detection. In 2019 IEEE In- ternational Conference on Image Processing (ICIP) , pages 4030–4034. IEEE, 2019. 2

    2019

  32. [40]

    Li, and Ge Li

    Jia-Xing Zhong, Nannan Li, Weijie Kong, Shan Liu, Thomas H. Li, and Ge Li. Graph convolutional label noise cleaner: Train a plug-and-play action classifier for anomaly detection. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2019. 2

    2019

  33. [41]

    Dual memory units with uncertainty regulation for weakly supervised video anomaly detection

    Hang Zhou, Junqing Yu, and Wei Yang. Dual memory units with uncertainty regulation for weakly supervised video anomaly detection. arXiv preprint arXiv:2302.05160, 2023. 5, 6, 7

    2023 arXiv

  34. [42]

    Advancing video anomaly detection: A concise re- view and a new dataset

    Liyun Zhu, Lei Wang, Arjun Raj, Tom Gedeon, and Chen Chen. Advancing video anomaly detection: A concise re- view and a new dataset. In The Thirty-eight Conference on Neural Information Processing Systems Datasets and Bench- marks Track, 2024. 2, 6

    2024

  35. [43]

    Motion-aware feature for improved video anomaly detection

    Yi Zhu and Shawn Newsam. Motion-aware feature for improved video anomaly detection. arXiv preprint arXiv:1907.10211, 2019. 2

    1907 arXiv

Pith tools

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