Pith. sign in

REVIEW 5 major objections 5 minor 65 references

Track Any Anomalous Object: A Granular Video Anomaly Detection Pipeline

T0 review · 5 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read TAO is the first pipeline to turn object-centric anomaly scores into pixel-level tracks of multiple anomalies, reporting state-of-the-art Pixel-F1 of 64.12 on UCSD Ped2 and RBDC/TBDC of 83.6/93.2.

desk verdict A useful system integration and evaluation protocol, but the SOTA claims rest on test-set threshold tuning and an apples-to-oranges baseline comparison. read the letter →

arxiv 2506.05175 v1 pith:SBE3QMUR submitted 2025-06-05 cs.CV

Yuzhi Huang , Chenxin Li , Haitao Zhang , Zixu Lin , Yunlong Lin , Hengyu Liu , Wuyang Li , Xinyu Liu
show 4 more authors
Jiechao Gao Yue Huang Xinghao Ding Yixuan Yuan
This is my paper
classification cs.CV
keywords videoanomalydetectionpixel-levellocalizationobject-centricSAM2promptingtemporalconsistencyfilteringsegmentationRBDC/TBDCmetrics
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 video anomaly detection should be treated as a pixel-level tracking problem, not just a frame-level or object-level scoring problem. It proposes TAO, a pipeline that takes bounding boxes from an object-centric anomaly scorer, filters them for temporal consistency, and feeds the surviving boxes and their center points into SAM2, a prompt-based video segmentation model that propagates masks across the whole clip. On UCSD Ped2, TAO reports Pixel-F1 of 64.12 versus 10.89 for the best image-anomaly baseline, with RBDC/TBDC of 83.6/93.2; on ShanghaiTech Campus it reports RBDC/TBDC of 62.1/85.4. If these numbers hold, TAO is the first framework to localize multiple anomalous objects at pixel level in long videos while still tracking each anomaly as a coherent object.

What carries the argument

The load-bearing mechanism is the coupling of three components: an object-centric anomaly scorer that assigns anomaly scores to detected bounding boxes, the Boxes Robustness Filtering algorithm that enforces temporal consistency by requiring each candidate box to overlap boxes in at least m of the previous or next k frames to inherit or receive a label, and SAM2, a prompt-based video segmentation model that takes the surviving boxes and their center points, saved every l frames, and propagates pixel masks across the entire clip. The filter converts noisy box-level scores into reliable prompts; in the ablation, enabling video tracking and filtering together raises Pixel-AUROC from 41.72 to 75.11 on UCSD Ped2.

What would settle it

Take a test clip in which a true anomalous object is occluded or moves so quickly that its bounding boxes overlap less than the h=0.2 threshold for at least two frames within any five-frame window, run TAO with the paper's fixed parameters, and measure Pixel-F1 against the ground-truth masks; if the score collapses relative to supplying ground-truth boxes directly to SAM2, the filter, not the segmenter, is the failure point.

Watch

Extended reading notes

Core claim

The central claim is that pixel-level anomaly segmentation can be obtained without dense per-pixel training by converting an object-centric anomaly detector into a prompt source for a video segmentation foundation model. The paper introduces a Boxes Robustness Filtering step that keeps only anomaly boxes reappearing with sufficient overlap in at least three of five neighboring frames, preventing redundant boxes from corrupting SAM2's temporal propagation; with this filter, TAO achieves the reported state-of-the-art results on UCSD Ped2 and ShanghaiTech under both pixel-level and object-level metrics. The authors also propose a dual-level benchmark that evaluates pixel metrics (AUROC, AP, AUPRO, F1) and object metrics (RBDC, TBDC) together, arguing that this combined view is better for complex scenes with overlapping or multiple anomalies.

Load-bearing premise

The whole tracking stage assumes that a genuine anomaly reappears in at least three of the neighboring five frames with enough spatial overlap, while false alarms appear only sporadically; if an anomaly moves too fast, is occluded, or overlaps a normal object, the filter can drop it and SAM2 then propagates the wrong mask.

Editorial extensions

If this is right

  • Any object-centric video anomaly detector can be upgraded to a pixel-level multi-object tracker by appending the filtering and prompt-propagation stages, with no retraining of the segmentation model.
  • Future video anomaly detection models will be judged on two dimensions at once, spatial precision from pixel metrics and temporal coherence from RBDC and TBDC, so a method that is strong on only one axis is no longer sufficient.
  • Sparse prompting works: saving filtered boxes every l frames, 5 on Ped2 and 15 on ShanghaiTech, still propagates masks to all frames, reducing storage and compute for long surveillance clips.
  • The framework stays stable across SAM2 backbone sizes from Tiny to Large and across anomaly thresholds in the range 1.0 to 1.9, indicating the gains are not tied to a particular prompt threshold or model capacity.

Reading between the lines

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

  • A natural extension is to make the filter's window parameters scene-adaptive; fast-moving or intermittently occluded anomalies would likely need a smaller overlap threshold or a shorter window than the fixed k=5, m=3, h=0.2 used here.
  • The main reported comparisons are against frame-by-frame image anomaly detectors; a video-native baseline with its own temporal smoothing could be a stronger test of TAO's tracking contribution.
  • Because the pipeline still relies on a score threshold to nominate candidate boxes, the paper's 'threshold-free' claim is better read as 'less sensitive to the threshold' rather than 'no threshold at all'.
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

5 major / 5 minor

Summary. The paper proposes Track Any Anomalous Object (TAO), a video anomaly detection pipeline that combines object-centric bounding-box anomaly scoring with a temporal consistency filter (Algorithm 1) and SAM2-based promptable segmentation. The claimed contributions are a new dual-level (pixel and object) evaluation benchmark and state-of-the-art results on UCSD Ped2 and ShanghaiTech Campus, including Pixel-F1 of 64.12 and RBDC/TBDC of 83.6/93.2 on Ped2. The pipeline is described in Sections 3.1-3.3 and evaluated in Section 4 with ablations, sensitivity analyses, and qualitative comparisons.

Significance. If validated, TAO would be a useful integration of object-level anomaly scoring with a frozen video segmentation foundation model, and the proposed dual-level benchmark would address a real gap in VAD evaluation: existing frame-level and object-level metrics do not measure pixel-accurate localization of multiple anomalous objects. The paper is clearly written and Algorithm 1 is well specified, and the idea of linking anomaly scores to SAM2 prompts is simple and falsifiable. However, the headline SOTA claim is not currently established because the evaluation has load-bearing weaknesses: per-dataset threshold selection on the test sets, an underspecified baseline protocol, no error bars, and an asserted removal of threshold tuning that is contradicted by the method itself.

major comments (5)
  1. [Sec. 4.2, Eq. (3), Fig. 5] The abstract's claim that TAO 'removes the need for threshold tuning' is contradicted by the method: Eq. (3) retains boxes only when s_i > tau, and Sec. 4.2 assigns dataset-specific thresholds (tau=1.5 for Ped2, tau=1.6 for ShanghaiTech) and save intervals (l=5 vs. l=15). Figure 5 is a sensitivity sweep on the ShanghaiTech test set whose optimum (tau=1.6) is then used in the reported results. Since tau and l affect Pixel-F1, RBDC, and TBDC, the reported SOTA numbers are partly fitted to the test data. Please select all hyperparameters on held-out validation data, fix them across datasets, or report full operating curves; otherwise the central empirical claim is not an independent prediction.
  2. [Sec. 4.3, Table 1] The comparison with image anomaly detectors is not controlled. The baselines are applied frame-by-frame and 'pixel-level segmentation is achieved by applying a threshold to these heatmaps,' but the threshold is never specified, and the baselines are not given temporal filtering. The large Pixel-F1 margin (64.12 vs. 10.89 for DRAEM) could therefore reflect an unfavorable binarization threshold for the baselines rather than a property of the TAO pipeline. Report the heatmap thresholds used for each baseline (or best thresholds under the same protocol), and ideally evaluate baselines with a simple temporal smoothing step to isolate the contribution of temporal consistency.
  3. [Sec. 4.1, Benchmark Datasets] The proposed 'new comprehensive VAD benchmark' is not reproducible as described. The paper defines pixel-level metrics on UCSD Ped2 and ShanghaiTech Campus but never states how the pixel-level ground-truth masks were obtained, whether they are part of the original datasets, or how any box-level annotations were converted to masks. Without this annotation protocol, the Pixel-AUROC, Pixel-AP, Pixel-AUPRO, and Pixel-F1 numbers cannot be independently verified. Specify the source of pixel labels and any preprocessing, and release the annotation code.
  4. [Sec. 3.2, Algorithm 1] The Boxes Robustness Filtering relies on the temporal-consistency premise that true anomalous boxes reappear with IoU > h in at least m of k neighboring frames, while false positives appear sporadically. This premise is asserted without evidence, and no experiments cover fast-moving, occluded, or mutually overlapping anomalies, exactly the cases where the premise can fail. The paper should include failure-case analysis or synthetic stress tests; otherwise the claimed robustness of the pipeline is unsupported for object classes beyond those in the two test sets.
  5. [Tables 1-3] No variance or repeated-run statistics are reported. On ShanghaiTech, the TBDC improvement over STPT is only 0.8 points (85.4 vs. 84.6), and with the threshold-fitted protocol described above it is unclear whether this difference is significant. Report means and standard deviations over at least three runs, or justify why the pipeline is deterministic and the single-run numbers are exact.
minor comments (5)
  1. [Sec. 4.4 heading] The heading 'Comparsion to Conventional Video Anomaly Detection Models' contains a typo; it should be 'Comparison'.
  2. [Sec. 4.2] The sentence 'segmentation probability maps are uniformly binarized' omits the binarization threshold; since Pixel-F1 depends directly on this threshold, its value should be stated.
  3. [Eq. (4)] The set notation writes I={(c_j, b_j, f_i) | ...}, which mixes frame indices and frame contents; clarify that the prompts contain the actual frame tensors for the saved indices.
  4. [Fig. 6] The caption says 'Quantitative Results' but does not identify which metric is plotted on the vertical axis; please specify.
  5. [References] Several references lack venues or identifiers (e.g., [7], [8], [36], [45]), and some cited papers appear unrelated to video anomaly detection; please verify the relevance and completeness of the reference list.

Circularity Check

1 steps flagged · score 6.0 of 10

The headline SOTA numbers are partially fitted: the anomaly threshold is selected on the same test set via a sensitivity sweep, and the abstract's 'no threshold tuning' claim is contradicted by Eq. (3).

  1. fitted input called prediction [Section 4.2, Section 4.5 (Figure 5), and Tables 1 and 3; with Equation (3)]
    "In the bounding box threshold filtering stage, we assign anomaly scores based on pose and depth features, with thresholds set to τ = 1.5 for UCSD Ped2 and τ = 1.6 for ShanghaiTech Campus. ... As τ ranges from 1.0 to 1.9, the framework performs stably, with optimal results at τ = 1.6. TBDC is most sensitive, peaking at 85% at τ = 1.6."

    The reported ShanghaiTech TBDC of 85.4 (Table 3) is produced with τ = 1.6, the same value that Section 4.5 identifies as the optimum of the sensitivity sweep on the same benchmark ('peaking at 85% at τ = 1.6'). The headline result is therefore the evaluation metric evaluated at a parameter chosen to maximize that metric on the test set, rather than an independent prediction at a fixed setting. The circularity is compounded by the abstract's claim that the method 'removes the need for threshold tuning,' which is contradicted by Eq. (3), where boxes are retained only when s_i > τ, and by the per-dataset choices of τ and save interval l in Section 4.2. The SOTA claim is thus partially a fitted result, not a validated out-of-sample prediction.

full rationale

The method itself is a composition of external, independently available components: an object-centric anomaly scorer (Reiss and Hoshen) and SAM2, with no load-bearing self-citation chain or imported uniqueness theorem. The pixel-level tracking idea has independent content beyond any single fit. However, the empirical SOTA claim is not fully independent: the anomaly threshold τ is selected per dataset after inspecting the same evaluation benchmarks, and the sensitivity analysis explicitly chooses τ = 1.6 on ShanghaiTech because TBDC peaks there, while Table 3 reports the ShanghaiTech TBDC as 85.4 at that same τ. Thus the central evaluation number is partly the output of a fit to the test set, which matches the 'fitted input called prediction' circularity pattern. The abstract's assertion that the method removes threshold tuning is directly contradicted by the method's own Eq. (3) and the implementation details in Sec. 4.2; this is a correctness and reporting risk, though it is not itself a circular derivation. No other circular steps were found: the baseline protocol and metric choices raise fairness concerns, but those are correctness risks rather than circularity. Overall, the partial reduction of the headline SOTA metric to threshold tuning on the same benchmark warrants a score of 6.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The central claim rests on two external models (an object-centric anomaly scorer and SAM2) whose reliability is assumed, and on a heuristic filtering algorithm with several hand-chosen parameters. The only new empirical content is the integration and the reported numbers, which depend on hyperparameters fitted to the test sets.

free parameters (6)
  • anomaly score threshold tau = 1.5 (UCSD Ped2), 1.6 (ShanghaiTech)
    Filters candidate anomalous boxes in Eq. (3); chosen per dataset and optimized via sensitivity analysis on the test set (Fig. 5).
  • save interval l = 5 (Ped2), 15 (ShanghaiTech)
    Determines how often prompts are saved for SAM2; set per dataset without detailed justification, affecting temporal propagation quality.
  • tracking window size k = 5
    Controls how many past/future frames are considered in Algorithm 1; chosen ad hoc.
  • frame match threshold m = 3
    Number of frames that must satisfy IoU overlap for label inheritance/assignment; chosen ad hoc.
  • overlap threshold h = 0.2
    IoU threshold for matching boxes in Algorithm 1; chosen ad hoc.
  • segmentation binarization threshold = not specified
    Segmentation probability maps are 'uniformly binarized' but the threshold value is not stated; required to compute pixel-F1 and pixel-AP.
assumptions (5)
  • domain assumption True anomalous boxes maintain temporal consistency (IoU > h for at least m of k frames), while redundant boxes appear sporadically.
    Core premise of the Boxes Robustness Filtering (Algorithm 1, Section 3.2). If anomalies move fast, get occluded, or overlap, this assumption breaks and the filter may discard true anomalies or keep noise.
  • domain assumption The object-centric anomaly scoring algorithm [44] provides reliable anomaly scores for detected bounding boxes.
    The pipeline inherits the accuracy of an unspecified object-centric VAD algorithm; the paper only cites [44] and does not detail it, so its behavior is taken as given.
  • domain assumption SAM2 can segment and propagate anomalous objects from box/point prompts without fine-tuning on anomaly data.
    The method depends on SAM2's zero-shot generalization to anomalies; no evidence beyond the reported results supports this assumption.
  • domain assumption The object detector D detects all anomalous objects; anomalies without detected boxes are missed by construction.
    Bounding boxes are generated by a pre-trained detector (Eq. 1); frame-level anomalies like smoke or tiny objects may be absent, so recall is bounded by the detector.
  • domain assumption Ground-truth pixel annotations of UCSD Ped2 and ShanghaiTech plus the RBDC/TBDC protocol are applied consistently to all compared methods.
    Comparisons assume uniform evaluation; the paper does not provide the exact evaluation code or annotation preprocessing.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Track Any Anomalous Object: A Granular Video Anomaly Detection Pipeline." pith.science (2026). https://pith.science/paper/SBE3QMUR

@misc{pith2026250605175,
  author       = {Pith},
  title        = {Pith review of: Track Any Anomalous Object: A Granular Video Anomaly Detection Pipeline},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SBE3QMUR}},
  note         = {Machine review of arXiv:2506.05175}
}
read the original abstract

Video anomaly detection (VAD) is crucial in scenarios such as surveillance and autonomous driving, where timely detection of unexpected activities is essential. Although existing methods have primarily focused on detecting anomalous objects in videos -- either by identifying anomalous frames or objects -- they often neglect finer-grained analysis, such as anomalous pixels, which limits their ability to capture a broader range of anomalies. To address this challenge, we propose a new framework called Track Any Anomalous Object (TAO), which introduces a granular video anomaly detection pipeline that, for the first time, integrates the detection of multiple fine-grained anomalous objects into a unified framework. Unlike methods that assign anomaly scores to every pixel, our approach transforms the problem into pixel-level tracking of anomalous objects. By linking anomaly scores to downstream tasks such as segmentation and tracking, our method removes the need for threshold tuning and achieves more precise anomaly localization in long and complex video sequences. Experiments demonstrate that TAO sets new benchmarks in accuracy and robustness. Project page available online.

Figures

Figures reproduced from arXiv: 2506.05175 by the authors.

Figure 1
Figure 1. Analyzing the Limitations of Existing VAD Models. Video anomaly detection (VAD) models are predominantly frame￾centric or object-centric. Frame-centric methods detect anomalies in frames without localizing them, while object-centric methods identify anomalous objects but lack pixel-level accuracy. Pixel￾centric models address these gaps by providing pixel-level local￾ization, delivering fine-grained segmentation and… view at source ↗
Figure 2
Figure 2. Pipeline of our proposed TAO. We first generate bounding boxes to identify objects in each frame. Next, we score these boxes using object-centric video anomaly detection algorithms to extract potential anomalous boxes. To ensure robustness, we apply filtering to eliminate redundant boxes. Finally, the filtered boxes and original frames are input into a prompt-based segmentation model to produce pixel-level anomaly s… view at source ↗
Figure 3
Figure 3. Impact of Redundant Segmentation on Performance. Potential anomalous boxes are used as prompts for SAM2 segmentation (red masks: true anomalies, green masks: redundant results). Initially, SAM2 tracks true anomalies accurately, but tracking errors accumu￾late, leading to catastrophic forgetting and a collapse in Pixel-F1 performance. Algorithm 1 Boxes Robustness Filtering 1: Input: A set of anomalous bounding boxes … view at source ↗
Figures from the paper (2 more)
Figure 5
Figure 5. Figure 5: Sensitivity Analysis of Anomaly Threshold (τ ). We evaluate performance under varying threshold on ShanghaiTech Campus dataset. Quantitative Results(%) Tiny Small Base+ Large [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Robust Testing of SAM2 Backbones. We compare the performance across different SAM2 model variants. 4.5. Further Empirical Study Ablation Study. We conduct ablation experiments on the UCSD Ped2 dataset to evaluate two components: video tracking mode and box robust filte…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

65 extracted references · 50 canonical work pages

  1. [1]

    VideoPatchCore: An Effective Method to Memorize Normality for Video Anomaly Detection

    Sunghyun Ahn, Youngwan Jo, Kijung Lee, and Sanghyun Park. Videopatchcore: An effective method to memo- rize normality for video anomaly detection.arXiv preprint arXiv:2409.16225, 2024. 1, 2

    work page Pith review arXiv 2024

  2. [2]

    Integrat- ing view conditions for image synthesis.arXiv preprint arXiv:2310.16002, 2023

    Jinbin Bai, Zhen Dong, Aosong Feng, Xiao Zhang, Tian Ye, Kaicheng Zhou, and Mike Zheng Shou. Integrat- ing view conditions for image synthesis.arXiv preprint arXiv:2310.16002, 2023. 1

    arXiv 2023

  3. [3]

    Humanedit: A high- quality human-rewarded dataset for instruction-based image editing.arXiv preprint arXiv:2412.04280, 2024

    Jinbin Bai, Wei Chow, Ling Yang, Xiangtai Li, Juncheng Li, Hanwang Zhang, and Shuicheng Yan. Humanedit: A high- quality human-rewarded dataset for instruction-based image editing.arXiv preprint arXiv:2412.04280, 2024. 1

    arXiv 2024

  4. [4]

    Meissonic: Revitalizing Masked Generative Trans- formers for Efficient High-Resolution Text-to-Image Synthe- sis.arXiv preprint arXiv:2410.08261, 2024

    Jinbin Bai, Tian Ye, Wei Chow, Enxin Song, Qing-Guo Chen, Xiangtai Li, Zhen Dong, Lei Zhu, and Shuicheng Yan. Meissonic: Revitalizing Masked Generative Trans- formers for Efficient High-Resolution Text-to-Image Synthe- sis.arXiv preprint arXiv:2410.08261, 2024. 2

    arXiv

  5. [5]

    Adaclip: Adapting clip with hybrid learnable prompts for zero-shot anomaly de- tection

    Yunkang Cao, Jiangning Zhang, Luca Frittoli, Yuqi Cheng, Weiming Shen, and Giacomo Boracchi. Adaclip: Adapting clip with hybrid learnable prompts for zero-shot anomaly de- tection. InEuropean Conference on Computer Vision, pages 55–72. Springer, 2025. 6, 7

    work page 2025

  6. [6]

    Unsupervised anomaly segmentation for brain lesions using dual semantic-manifold reconstruc- tion

    Zhiyuan Ding, Qi Dong, Haote Xu, Chenxin Li, Xinghao Ding, and Yue Huang. Unsupervised anomaly segmentation for brain lesions using dual semantic-manifold reconstruc- tion. InInternational Conference on Neural Information Processing, pages 133–144. Springer International Publish- ing Cham, 2022. 2

    work page 2022

  7. [7]

    Any-shot sequential anomaly detection in surveillance videos

    Keval Doshi and Yasin Yilmaz. Any-shot sequential anomaly detection in surveillance videos. 2020. 2

    work page 2020

  8. [8]

    Continual learning for anomaly detection in surveillance videos

    Keval Doshi and Yasin Yilmaz. Continual learning for anomaly detection in surveillance videos. 2020. 1, 2

    work page 2020

Show all 65 references
  1. [9]

    Instantsplat: Sparse-view gaussian splatting in sec- onds, 2024

    Zhiwen Fan, Kairun Wen, Wenyan Cong, Kevin Wang, Jian Zhang, Xinghao Ding, Danfei Xu, Boris Ivanovic, Marco Pavone, Georgios Pavlakos, Zhangyang Wang, and Yue Wang. Instantsplat: Sparse-view gaussian splatting in sec- onds, 2024. 2

    2024

  2. [10]

    Mariana-Iuliana Georgescu, RaduTudor Ionescu, Fahad- Shahbaz Khan, Marius Popescu, and Mubarak Shah. A background-agnostic framework with adversarial training for abnormal event detection in video.IEEE Transactions on Pattern Analysis and Machine Intelligence,IEEE Transac- tio...

  3. [11]

    Anomaly detection in video via self- supervised and multi-task learning

    Mariana-Iuliana Georgescu, Antonio Barbalau, Radu Tu- dor Ionescu, Fahad Shahbaz Khan, Marius Popescu, and Mubarak Shah. Anomaly detection in video via self- supervised and multi-task learning. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition...

    2021

  4. [12]

    Roy-Chowdhury, and Larry S

    Mahmudul Hasan, Jonghyun Choi, Jan Neumann, Amit K. Roy-Chowdhury, and Larry S. Davis. Learning temporal regularity in video sequences. In2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016. 1, 2

    2016

  5. [13]

    Degradation-resistant unfolding network for heterogeneous image fusion

    Chunming He, Kai Li, Guoxia Xu, Yulun Zhang, Runze Hu, Zhenhua Guo, and Xiu Li. Degradation-resistant unfolding network for heterogeneous image fusion. InICCV, pages 12611–12621, 2023. 2

    2023

  6. [14]

    Weakly- supervised concealed object segmentation with sam-based pseudo labeling and multi-scale feature grouping.NeurIPS, 36, 2024

    Chunming He, Kai Li, Yachao Zhang, Guoxia Xu, Longxi- ang Tang, Yulun Zhang, Zhenhua Guo, and Xiu Li. Weakly- supervised concealed object segmentation with sam-based pseudo labeling and multi-scale feature grouping.NeurIPS, 36, 2024. 2

    2024

  7. [15]

    Strategic preys make acute predators: Enhancing camouflaged object detectors by generating camouflaged objects.ICLR, 2024

    Chunming He, Kai Li, Yachao Zhang, Yulun Zhang, Zhen- hua Guo, Xiu Li, Martin Danelljan, and Fisher Yu. Strategic preys make acute predators: Enhancing camouflaged object detectors by generating camouflaged objects.ICLR, 2024. 2

    2024

  8. [16]

    Reti-diff: Illumination degradation image restoration with retinex-based latent diffusion model.ICLR, 2025

    Chunming He, Chengyu Fang, Yulun Zhang, Kai Li, Longx- iang Tang, Chenyu You, Fengyang Xiao, Zhenhua Guo, and Xiu Li. Reti-diff: Illumination degradation image restoration with retinex-based latent diffusion model.ICLR, 2025. 2

    2025

  9. [17]

    Diffusion models in low-level vision: A survey.TPAMI, 2025

    Chunming He, Yuqi Shen, Chengyu Fang, Fengyang Xiao, Longxiang Tang, Yulun Zhang, Wangmeng Zuo, Zhenhua Guo, and Xiu Li. Diffusion models in low-level vision: A survey.TPAMI, 2025. 2

    2025

  10. [18]

    Run: Reversible unfolding network for concealed object segmentation.arXiv preprint arXiv:2501.18783, 2025

    Chunming He, Rihan Zhang, Fengyang Xiao, Chenyu Fang, Longxiang Tang, Yulun Zhang, Linghe Kong, Deng-Ping Fan, Kai Li, and Sina Farsiu. Run: Reversible unfolding network for concealed object segmentation.arXiv preprint arXiv:2501.18783, 2025. 2

    2025 arXiv

  11. [19]

    Joint detection and recounting of abnormal events by learning deep generic knowledge

    Ryota Hinami, Tao Mei, and Shin’ichi Satoh. Joint detection and recounting of abnormal events by learning deep generic knowledge. In2017 IEEE International Conference on Com- puter Vision (ICCV), 2017. 1

    2017

  12. [20]

    Object-centric auto-encoders and dummy anomalies for abnormal event detection in video

    Radu Tudor Ionescu, Fahad Shahbaz Khan, Mariana-Iuliana Georgescu, and Ling Shao. Object-centric auto-encoders and dummy anomalies for abnormal event detection in video. In 2019 IEEE/CVF Conference on Computer Vision and Pat- tern Recognition (CVPR), 2019. 1, 2, 6

    2019

  13. [21]

    Real-time weakly supervised video anomaly detection

    Hamza Karim, Keval Doshi, and Yasin Yilmaz. Real-time weakly supervised video anomaly detection. InProceedings of the IEEE/CVF winter conference on applications of com- puter vision, pages 6848–6856, 2024. 2

    2024

  14. [22]

    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. InProceedings of the IEEE/CVF International Con- ference on Computer Vision, pages 4015–4026, 2023. 3

    2023

  15. [23]

    Unsupervised anomaly segmentation using image-semantic cycle translation.arXiv preprint arXiv:2103.09094, 2021

    Chenxin Li, Yunlong Zhang, Jiongcheng Li, Yue Huang, and Xinghao Ding. Unsupervised anomaly segmentation using image-semantic cycle translation.arXiv preprint arXiv:2103.09094, 2021. 2

    2021 arXiv

  16. [24]

    Consistent posterior distributions under vessel-mixing: a regularization for cross-domain reti- nal artery/vein classification

    Chenxin Li, Yunlong Zhang, Zhehan Liang, Wenao Ma, Yue Huang, and Xinghao Ding. Consistent posterior distributions under vessel-mixing: a regularization for cross-domain reti- nal artery/vein classification. In2021 IEEE International Conference on Image Processing (ICIP), page...

  17. [25]

    Hierarchical deep network with uncertainty-aware semi-supervised learning 9 for vessel segmentation.Neural Computing and Applica- tions, pages 1–14, 2022

    Chenxin Li, Wenao Ma, Liyan Sun, Xinghao Ding, Yue Huang, Guisheng Wang, and Yizhou Yu. Hierarchical deep network with uncertainty-aware semi-supervised learning 9 for vessel segmentation.Neural Computing and Applica- tions, pages 1–14, 2022. 2

    2022

  18. [26]

    U-kan makes strong backbone for medical image segmentation and gener- ation.arXiv preprint arXiv:2406.02918, 2024

    Chenxin Li, Xinyu Liu, Wuyang Li, Cheng Wang, Hengyu Liu, Yifan Liu, Zhen Chen, and Yixuan Yuan. U-kan makes strong backbone for medical image segmentation and gener- ation.arXiv preprint arXiv:2406.02918, 2024. 5

    2024 arXiv

  19. [27]

    Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models

    Junnan Li, Dongxu Li, Silvio Savarese, and Steven Hoi. Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models. InIn- ternational conference on machine learning, pages 19730– 19742. PMLR, 2023. 2

    2023

  20. [28]

    Fusion2void: Unsupervised multi-focus image fusion based on image in- painting.IEEE Transactions on Circuits and Systems for Video Technology, 2024

    Huangxing Lin, Yunlong Lin, Jingyuan Xia, Linyu Fan, Feifei Li, Yingying Wang, and Xinghao Ding. Fusion2void: Unsupervised multi-focus image fusion based on image in- painting.IEEE Transactions on Circuits and Systems for Video Technology, 2024. 2

    2024

  21. [29]

    Aglldiff: Guiding diffusion models towards unsuper- vised training-free real-world low-light image enhancement

    Yunlong Lin, Tian Ye, Sixiang Chen, Zhenqi Fu, Yingying Wang, Wenhao Chai, Zhaohu Xing, Lei Zhu, and Xinghao Ding. Aglldiff: Guiding diffusion models towards unsuper- vised training-free real-world low-light image enhancement. arXiv preprint arXiv:2407.14900, 2024. 2

    2024 arXiv

  22. [30]

    Visual instruction tuning.Advances in neural information processing systems, 36, 2024

    Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. Visual instruction tuning.Advances in neural information processing systems, 36, 2024. 2

    2024

  23. [31]

    Fu- ture frame prediction for anomaly detection–a new baseline

    Wen Liu, Weixin Luo, Dongze Lian, and Shenghua Gao. Fu- ture frame prediction for anomaly detection–a new baseline. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 6536–6545, 2018. 5

    2018

  24. [32]

    A hybrid video anomaly detection framework via memory-augmented flow reconstruction and flow-guided frame prediction

    Zhian Liu, Yongwei Nie, Chengjiang Long, Qing Zhang, and Guiqing Li. A hybrid video anomaly detection framework via memory-augmented flow reconstruction and flow-guided frame prediction. InProceedings of the IEEE/CVF interna- tional conference on computer vision, pages 13588–13597,

  25. [33]

    A hybrid video anomaly detection framework via memory-augmented flow reconstruction and flow-guided frame prediction

    Zhian Liu, Yongwei Nie, Chengjiang Long, Qing Zhang, and Guiqing Li. A hybrid video anomaly detection framework via memory-augmented flow reconstruction and flow-guided frame prediction. In2021 IEEE/CVF International Confer- ence on Computer Vision (ICCV), 2021. 1, 2

    2021

  26. [34]

    Simplenet: A simple network for image anomaly detection and localization

    Zhikang Liu, Yiming Zhou, Yuansheng Xu, and Zilei Wang. Simplenet: A simple network for image anomaly detection and localization. InProceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition, pages 20402–20411, 2023. 6, 7

    2023

  27. [35]

    Shao-Yuan Lo, Poojan Oza, and Vishal M. Patel. Adversar- ially robust one-class novelty detection.IEEE Transactions on Pattern Analysis and Machine Intelligence, page 1–12,

  28. [36]

    Video-chatgpt: Towards detailed video understanding via large vision and language models.arXiv preprint arXiv:2306.05424, 2023

    Muhammad Maaz, Hanoona Rasheed, Salman Khan, and Fa- had Shahbaz Khan. Video-chatgpt: Towards detailed video understanding via large vision and language models.arXiv preprint arXiv:2306.05424, 2023. 2

    2023 arXiv

  29. [37]

    MULDE: Multiscale Log- Density Estimation via Denoising Score Matching for Video Anomaly Detection

    Jakub Micorek, Horst Possegger, Dominik Narnhofer, Horst Bischof, and Mateusz Kozi ´nski. MULDE: Multiscale Log- Density Estimation via Denoising Score Matching for Video Anomaly Detection. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CV...

    2024

  30. [38]

    Anomaly detection with conditioned denoising diffusion models.arXiv preprint arXiv:2305.15956, 2023

    Arian Mousakhan, Thomas Brox, and Jawad Tayyub. Anomaly detection with conditioned denoising diffusion models.arXiv preprint arXiv:2305.15956, 2023. 6, 7

    2023 arXiv

  31. [39]

    Spatio-temporal predictive tasks for abnormal event detection in videos

    Yassine Naji, Aleksandr Setkov, Angélique Loesch, Michèle Gouiffès, and Romaric Audigier. Spatio-temporal predictive tasks for abnormal event detection in videos. 2022. 1, 6, 7

    2022

  32. [40]

    Learn- ing memory-guided normality for anomaly detection

    Hyunjong Park, Jongyoun Noh, and Bumsub Ham. Learn- ing memory-guided normality for anomaly detection. In 2020 IEEE/CVF Conference on Computer Vision and Pat- tern Recognition (CVPR), 2020. 1, 2

    2020

  33. [41]

    Street scene: A new dataset and evaluation protocol for video anomaly detection

    Bharathkumar Ramachandra and Michael Jones. Street scene: A new dataset and evaluation protocol for video anomaly detection. InProceedings of the IEEE/CVF Win- ter Conference on Applications of Computer Vision, pages 2569–2578, 2020. 1, 6, 7

    2020

  34. [42]

    Jones, and Ranga Raju Vatsavai

    Bharathkumar Ramachandra, Michael J. Jones, and Ranga Raju Vatsavai. Learning a distance function with a siamese network to localize anomalies in videos. In2020 IEEE Win- ter Conference on Applications of Computer Vision (WACV),

  35. [43]

    Sam 2: Segment anything in images and videos.arXiv preprint arXiv:2408.00714, 2024

    Nikhila Ravi, Valentin Gabeur, Yuan-Ting Hu, Ronghang Hu, Chaitanya Ryali, Tengyu Ma, Haitham Khedr, Roman Rädle, Chloe Rolland, Laura Gustafson, et al. Sam 2: Segment anything in images and videos.arXiv preprint arXiv:2408.00714, 2024. 2, 3, 6

    2024 arXiv

  36. [44]

    Attribute-based representa- tions for accurate and interpretable video anomaly detection

    Tal Reiss and Yedid Hoshen. Attribute-based representa- tions for accurate and interpretable video anomaly detection

  37. [45]

    Video anomaly detection via sequentially learning multiple pretext tasks

    Chenrui Shi, Che Sun, Yuwei Wu, and Yunde Jia. Video anomaly detection via sequentially learning multiple pretext tasks. 1, 2

  38. [46]

    Few- shot medical image segmentation using a global correlation network with discriminative embedding.Computers in biol- ogy and medicine, 140:105067, 2022

    Liyan Sun, Chenxin Li, Xinghao Ding, Yue Huang, Zhong Chen, Guisheng Wang, Yizhou Yu, and John Paisley. Few- shot medical image segmentation using a global correlation network with discriminative embedding.Computers in biol- ogy and medicine, 140:105067, 2022. 2

    2022

  39. [47]

    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. InProceedings of the IEEE/CVF international conference on computer vision, pages 4975–498...

    2021

  40. [48]

    Anomaly detection in crowd scene

    Shu Wang and Zhenjiang Miao. Anomaly detection in crowd scene. InIEEE 10th International Conference on Signal Pro- cessing Proceedings, pages 1220–1223. IEEE, 2010. 5

    2010

  41. [49]

    Learning high-frequency feature enhancement and alignment for pan-sharpening

    Yingying Wang, Yunlong Lin, Ge Meng, Zhenqi Fu, Yuhang Dong, Linyu Fan, Hedeng Yu, Xinghao Ding, and Yue Huang. Learning high-frequency feature enhancement and alignment for pan-sharpening. InProceedings of the 31st ACM International Conference on Multimedia, pages 358– 367, 2023. 1

    2023

  42. [50]

    Learn- ing diffusion high-quality priors for pan-sharpening: A two- stage approach with time-aware adapter fine-tuning.IEEE Transactions on Geoscience and Remote Sensing, 2025

    Yingying Wang, Yunlong Lin, Xuanhua He, Hui Zheng, Keyu Yan, Linyu Fan, Yue Huang, and Xinghao Ding. Learn- ing diffusion high-quality priors for pan-sharpening: A two- stage approach with time-aware adapter fine-tuning.IEEE Transactions on Geoscience and Remote Sensing, 2025. 1

    2025

  43. [51]

    Self-supervised sparse representa- 10 tion for video anomaly detection

    Jhih-Ciang Wu, He-Yen Hsieh, Ding-Jie Chen, Chiou-Shann Fuh, and Tyng-Luh Liu. Self-supervised sparse representa- 10 tion for video anomaly detection. InEuropean Conference on Computer Vision, pages 729–745. Springer, 2022. 2

    2022

  44. [52]

    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. InProceedings of the AAAI Con- ference on Artificial Intelligence, pages 6074–6082, 2024. 2

    2024

  45. [53]

    A survey of camouflaged object detection and be- yond.CAAI AIR, 2024

    Fengyang Xiao, Sujie Hu, Yuqi Shen, Chengyu Fang, Jinfa Huang, Chunming He, Longxiang Tang, Ziyun Yang, and Xiu Li. A survey of camouflaged object detection and be- yond.CAAI AIR, 2024. 2

    2024

  46. [54]

    Nestedformer: Nested modality-aware transformer for brain tumor segmentation

    Zhaohu Xing, Lequan Yu, Liang Wan, Tong Han, and Lei Zhu. Nestedformer: Nested modality-aware transformer for brain tumor segmentation. InInternational Conference on Medical Image Computing and Computer-Assisted Interven- tion, pages 140–150. Springer, 2022. 2

    2022

  47. [55]

    Diff-unet: A diffusion embedded network for volumet- ric segmentation.arXiv preprint arXiv:2303.10326, 2023

    Zhaohu Xing, Liang Wan, Huazhu Fu, Guang Yang, and Lei Zhu. Diff-unet: A diffusion embedded network for volumet- ric segmentation.arXiv preprint arXiv:2303.10326, 2023. 2

    2023 arXiv

  48. [56]

    Cross-conditioned diffu- sion model for medical image to image translation

    Zhaohu Xing, Sicheng Yang, Sixiang Chen, Tian Ye, Yi- jun Yang, Jing Qin, and Lei Zhu. Cross-conditioned diffu- sion model for medical image to image translation. InIn- ternational Conference on Medical Image Computing and Computer-Assisted Intervention, pages 201–211. Springer,

  49. [57]

    Segmamba: Long-range sequential modeling mamba for 3d medical image segmentation

    Zhaohu Xing, Tian Ye, Yijun Yang, Guang Liu, and Lei Zhu. Segmamba: Long-range sequential modeling mamba for 3d medical image segmentation. InInternational Conference on Medical Image Computing and Computer-Assisted Interven- tion, pages 578–588. Springer, 2024

    2024

  50. [58]

    Follow the rules: Reasoning for video anomaly detection with large language models

    Yuchen Yang, Kwonjoon Lee, Behzad Dariush, Yinzhi Cao, and Shao-Yuan Lo. Follow the rules: Reasoning for video anomaly detection with large language models. InPro- ceedings of the European Conference on Computer Vision (ECCV), 2024. 1, 2

    2024

  51. [59]

    Cloze test helps: Effec- tive video anomaly detection via learning to complete video events

    Guang Yu, Siqi Wang, Zhiping Cai, En Zhu, Chuanfu Xu, Jianping Yin, and Marius Kloft. Cloze test helps: Effec- tive video anomaly detection via learning to complete video events. InProceedings of the 28th ACM International Con- ference on Multimedia, 2020. 2

    2020

  52. [60]

    Harnessing large language mod- els for training-free video anomaly detection

    Luca Zanella, Willi Menapace, Massimiliano Mancini, Yim- ing Wang, and Elisa Ricci. Harnessing large language mod- els for training-free video anomaly detection. InProceedings of the IEEE/CVF Conference on Computer Vision and Pat- tern Recognition, pages 18527–18536, 2024. 2

    2024

  53. [61]

    Draem- a discriminatively trained reconstruction embedding for sur- face anomaly detection

    Vitjan Zavrtanik, Matej Kristan, and Danijel Skoˇcaj. Draem- a discriminatively trained reconstruction embedding for sur- face anomaly detection. InProceedings of the IEEE/CVF international conference on computer vision, pages 8330– 8339, 2021. 6, 7

    2021

  54. [62]

    Video-llama: An instruction-tuned audio-visual language model for video un- derstanding.arXiv preprint arXiv:2306.02858, 2023

    Hang Zhang, Xin Li, and Lidong Bing. Video-llama: An instruction-tuned audio-visual language model for video un- derstanding.arXiv preprint arXiv:2306.02858, 2023. 2

    2023 arXiv

  55. [63]

    Generator versus segmentor: Pseudo-healthy synthesis

    Yunlong Zhang, Chenxin Li, Xin Lin, Liyan Sun, Yihong Zhuang, Yue Huang, Xinghao Ding, Xiaoqing Liu, and Yizhou Yu. Generator versus segmentor: Pseudo-healthy synthesis. InMedical Image Computing and Computer As- sisted Intervention–MICCAI 2021: 24th International Con- ference...

    2021

  56. [64]

    Anomalyclip: Object-agnostic prompt learn- ing for zero-shot anomaly detection.arXiv preprint arXiv:2310.18961, 2023

    Qihang Zhou, Guansong Pang, Yu Tian, Shibo He, and Jiming Chen. Anomalyclip: Object-agnostic prompt learn- ing for zero-shot anomaly detection.arXiv preprint arXiv:2310.18961, 2023. 6, 7

    2023

  57. [65]

    Segment everything everywhere all at once.Advances in Neural Information Processing Systems, 36, 2024

    Xueyan Zou, Jianwei Yang, Hao Zhang, Feng Li, Linjie Li, Jianfeng Wang, Lijuan Wang, Jianfeng Gao, and Yong Jae Lee. Segment everything everywhere all at once.Advances in Neural Information Processing Systems, 36, 2024. 3 11

    2024

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.