Exploring Large Vision-Language Models for Robust and Efficient Industrial Anomaly Detection

Kun Qian; Tianyu Sun; Wenhong Wang

REVIEW 5 major objections 6 minor 29 references

Exploring Large Vision-Language Models for Robust and Efficient Industrial Anomaly Detection

T0 review · 5 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read CLAD, a contrastive cross-modal training method, claims to outperform existing industrial anomaly detection methods on image-level and pixel-level benchmarks by aligning visual and textual embeddings.

desk verdict The paper's central claim is internally falsified by its own Table 1 on VisA pixel AUC, and the contrastive loss is incoherent under the stated normal-only setting; this is a desk reject. read the letter →

arxiv 2412.00890 v1 pith:GJNEEYHG submitted 2024-12-01 cs.CV

Kun Qian , Tianyu Sun , Wenhong Wang This is my paper · ORCID

classification cs.CV

keywords LargeVision-LanguageModelsIndustrialAnomalyDetectionContrastiveLearningCross-ModalAlignmentLocalizationMVTec-ADVisAInterpretability

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

The paper introduces CLAD, a method that trains a vision encoder and a text encoder together so that an industrial product image is embedded close to its text description and normal instances are grouped together while anomalies are pushed apart. The claim is that this cross-modal contrastive alignment, followed by task-driven fine-tuning on industrial data, gives better anomaly detection and localization than existing unsupervised methods. On the MVTec-AD and VisA benchmarks, CLAD reports the highest Image-AUC and Pixel-AUC among the compared methods, and a human evaluation rates its localization heatmaps as more accurate than the baselines. The authors argue that this makes vision-language models a practical route to interpretable industrial quality control.

What carries the argument

The key mechanism is the contrastive cross-modal loss in Eq. (3), which pulls matched image-text pairs within a margin $\alpha$ and pushes every non-matching pair beyond a margin $\beta$, together with a reconstruction loss that preserves visual and textual details during fine-tuning. The visual encoder (ViT or CNN) and textual encoder (a pretrained transformer) map inputs to a shared embedding space, and the anomaly score $S(I,T)=\exp(-\|z_v-z_t\|^2/\sigma)$ turns the embedding distance into a similarity measure. For localization, the paper uses Grad-CAM to produce a heatmap of the image regions that most contribute to the mismatch.

What would settle it

Train CLAD on the official MVTec-AD normal-only training split and compute the negative term in Eq. (3); if normal-instance embeddings spread apart under the loss while the paper's stated goal is to group them, then the reported 94.1 Image-AUC cannot be reproduced without adding anomalous pairs to training.

Watch

Extended reading notes

Core claim

The central discovery is that aligning image and text embeddings with a contrastive loss is enough to separate normal from anomalous industrial images without relying on reconstruction error or memory banks. CLAD computes an anomaly score from the distance between the test image embedding and the text description embedding, so an image that does not match its expected textual prototype is flagged as anomalous, and Grad-CAM highlights the regions responsible. Across the 15 categories of MVTec-AD and 12 categories of VisA, the paper reports Image-AUC of 94.1 and 86.1 respectively, both above the strongest baseline WinCLIP, with Pixel-AUC of 95.3 and 96.2. The paper interprets these results as evidence that cross-modal alignment generalizes to unseen anomaly types and yields interpretable localizations.

Load-bearing premise

The training signal assumes the dataset provides both normal and anomalous instances to pull apart, but the benchmark is used in its standard unsupervised form with normal-only training images, so the paper never states where the negative pairs come from.

Editorial extensions

If this is right

If the reported numbers hold, a single contrastive alignment objective can outperform dedicated reconstruction and memory-bank pipelines for industrial defect detection and localization.
The text-guided anomaly score means new product types could be supported by writing a description of the normal object, without collecting defect samples for training.
The high human rating of the heatmaps suggests the same embeddings can serve as a visual explanation channel in factory quality-control interfaces.
Because the method scores by mismatch to the text prototype, it could in principle flag anomaly types never seen during training, which is the paper's stated goal.

Reading between the lines

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

The loss as written in Eq. (3) pushes all non-matching instances apart, so under the standard unsupervised setup with normal-only training images, the negative term would separate normal instances from each other; a natural follow-up experiment is to test whether the reported results require anomalous training pairs to be present.
A testable extension is to compare CLAD's text prototype score against a simpler zero-shot CLIP distance baseline with no fine-tuning, which would isolate the contribution of the contrastive training signal.
The same contrastive objective could be adapted to few-shot defect detection by adding a small set of textual descriptions of known defect types, turning the anomaly score into a multi-class defect classifier.
If the reconstruction loss term is removed, one can measure whether the remaining contrastive alignment alone preserves the reported localization quality, clarifying how much of the gain comes from reconstruction versus alignment.

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.

Desk Editor's Note

read the letter

Colleague, you should know two things about this one. First, the headline claim — that CLAD consistently achieves the highest image and pixel AUC on both MVTec-AD and VisA — is contradicted by the paper's own Table 1. On VisA pixel-level AUC, WinCLIP reports 96.4 and CLAD reports 96.2. That is not an external reproducibility quibble; it's an internal falsification of the central contribution. Second, the training loss in Eq. 3 is not coherent with the stated unsupervised setting, where only normal images are used. The negative pair term pushes different instances apart, which would separate normal samples from each other. If the authors used labeled anomalies at training time, they are not solving the problem they claim to solve. They never resolve this.

What's actually new here is not much. The method is a direct application of CLIP-style contrastive cross-modal alignment plus Grad-CAM, and the authors cite WinCLIP and AnomalyGPT, which already apply vision-language models to IAD. There is no new architecture, no new loss, and no new theoretical insight. The paper is organized and readable, and the direction — leveraging textual descriptions for anomaly localization — is sensible. But the machinery is standard, and the experiments do not back the claims.

The soft spots are severe and mechanical. Eq. 7 uses inverse encoders f_v^{-1} and f_t^{-1} that are never defined, and there is no indication how reconstruction is computed. Baseline numbers in Table 1 do not match published values; for instance, WinCLIP's MVTec image AUC is typically around 91.8, not 93.1. The human evaluation protocol is unspecified — how many experts, what instructions, how the scale was used. The IoU results in Table 4 have no per-class breakdown and no statistical testing. The ablation study is only on MVTec, has no error bars, and the "simple CNN" row is an uninterpretable confound. There is no code, no training details, no hyperparameter settings.

In short, this paper is not ready for serious review. It might serve as a rough student project template, but the internal contradiction alone is enough to desk reject. I would not send it to referees; the flaws are load-bearing, not fixable with minor revision. Recommendation: reject without further consideration.

Referee Report

5 major / 6 minor

Summary. This manuscript proposes CLAD, a method that uses a pretrained vision-language model to embed images and text, aligns them with a contrastive loss, and scores anomalies by cross-modal embedding distance, with Grad-CAM-style localization. Experiments on MVTec-AD and VisA report image-level and pixel-level AUC compared with SPADE, PaDiM, PatchCore, and WinCLIP, plus an ablation study, human evaluation, and IoU analysis. The paper claims CLAD achieves the highest scores in both detection and localization on both datasets.

Significance. The topic is timely, and a simple cross-modal alignment approach for industrial anomaly detection would be a useful contribution if the claimed results were supported. The manuscript, however, does not provide a working, reproducible specification of the method: the central experimental claim is contradicted by its own Table 1, the contrastive loss is ill-posed under the stated normal-only training setting, and the reconstruction loss depends on undefined inverse encoder functions. No code, no text prompts, no hyperparameters, and no evaluation protocols are given. The paper's strengths are its clear identification of the problem and the presence of ablations; but the evidence is insufficient to support the stated claims.

major comments (5)

[Section 4.2, Table 1] The central claim that CLAD "consistently outperforms all other methods on both datasets, achieving the highest scores in both anomaly detection and localization tasks" is directly contradicted by the numbers in Table 1: on VisA pixel-level AUC, CLAD is reported as 96.2±0.1 while WinCLIP is reported as 96.4±0.4. With the numbers as printed, CLAD does not achieve the highest pixel-level AUC on VisA. Since the entire contribution is defined by this reported superiority, the table must be corrected or the claim withdrawn.
[Section 3.2, Eq. (3)] The contrastive loss treats pairs (i,j), j≠i, as negative pairs. Section 4.1 describes the standard unsupervised setting with normal-only training images. Under normal-only training, those negative pairs are normal-normal pairs, so the loss pushes normal instances apart, which is the opposite of the stated goal of grouping normal instances. If labeled anomalous pairs are used instead, the method is supervised and no longer matches the claimed unsupervised IAD setting. The paper never resolves this ambiguity.
[Section 3.4, Eq. (7)] The reconstruction loss uses inverse encoder functions f_v^{-1} and f_t^{-1}, which are never defined anywhere in the architecture or experimental setup. Pretrained CLIP-style encoders do not provide such inverses, and no decoder or inversion procedure is described. Consequently the total loss in Eq. (6) is not computable as stated, and the training procedure is incomplete.
[Sections 4.4 and 4.5, Tables 3 and 4] The human evaluation and IoU results are reported without any protocol. There is no information about the number of expert raters, the number of images rated, the exact rating instructions, whether the raters were blinded to method, or whether differences are statistically significant. Table 4 does not state on which dataset or split the IoU was computed, nor the threshold used to binarize the predicted anomaly maps. These results are load-bearing for the interpretability claim and cannot be evaluated as presented.
[Section 4.1 and Table 1] The baseline comparison protocol is underspecified. The paper does not state whether the SPADE, PaDiM, PatchCore, and WinCLIP numbers are taken from the original publications or from a reimplementation, nor does it report the backbone, input resolution, prompt templates, or threshold choices used for the baselines. Without this information, the comparison in Table 1 cannot be independently verified, even setting aside the internal contradiction noted above.

minor comments (6)

[Section 1 and Section 4.1] The introduction names AnomalyGPT, PaDiM, and PatchCore as prior approaches, but Table 1 does not include AnomalyGPT. Since AnomalyGPT is an LVLM-based anomaly detection method and CLAD is also LVLM-based, its omission weakens the comparison and should be justified.
[Section 3.3, Eq. (5)] The symbols α_k and A_k in the Grad-CAM equation are not defined in the text; as written they are free symbols rather than the final-layer weights and activation maps of the visual encoder.
[Section 3.2, Eq. (3)] The notation for the positive-part function is inconsistent: the first term uses [·]_+ while the second term uses max(0, ·). Use a single notation throughout.
[Section 3.5 and Eq. (4)] The anomaly threshold and the scaling factor σ are never specified, and no sensitivity analysis is provided. This makes the classification rule in Section 3.5 difficult to reproduce.
[Section 2.2, Ref. [25]] Reference [25] is cited in support of Vision Transformers for industrial image anomaly detection, but the cited paper is about exemplar convolutional neural networks, not vision transformers; the citation appears mismatched.
[Entire manuscript] The paper contains no limitations section. Given the claims of real-world applicability, the authors should discuss known limitations, such as reliance on textual descriptions and sensitivity to prompt design.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported AUC results are empirical measurements against external benchmarks, not derivations from the model's own definitions.

full rationale

The paper makes an empirical claim that CLAD outperforms baselines on MVTec-AD and VisA. There is no derivation chain in which a fitted parameter or self-cited theorem is renamed as a prediction: the contrastive loss in Eq. 3, the anomaly score in Eq. 4, and Grad-CAM localization in Eq. 5 are the method itself, and the reported Image-AUC and Pixel-AUC values are measurements on held-out benchmarks. The references to prior vision-language models are background context, not load-bearing evidence for the central claim. The identified weaknesses, such as the internally inconsistent VisA Pixel-AUC comparison in Table 1, the undefined inverse encoders in Eq. 7, and the underspecified human-evaluation protocol, are correctness or reproducibility concerns rather than circularity. Because the central claim is externally falsifiable and the paper does not reduce any prediction to its inputs by construction, the appropriate circularity score is 0.

Assumptions & free parameters 5 free parameters · 3 assumptions · 2 invented entities

The method depends on several hand-set hyperparameters, undefined inverse encoders, and the assumption that anomalies are separable from normal text embeddings. None of these are given values or independent evidence.

free parameters (5)

alpha = not reported
Positive-pair margin in contrastive loss Eq. 3.
beta = not reported
Negative-pair margin in Eq. 3.
sigma = not reported
Scaling factor in anomaly score Eq. 4.
lambda = not reported
Weight balancing contrastive and reconstruction losses in Eq. 6.
anomaly threshold = not reported
Threshold below which S(I,T) is classified as anomalous in Section 3.5.

assumptions (3)

domain assumption Anomalous images are far from normal-object text embeddings in the shared CLIP-like space.
Used in Eq. 4 to define anomaly score S(I,T) as a function of embedding distance; no evidence is provided for this separation.
ad hoc to paper Negative pairs in Eq. 3 can be formed from different instances even though training sets are normal-only.
Equation 3 pushes all pairs j not equal to i apart; with normal-only training this contradicts the goal of grouping normal examples. The paper does not state whether anomaly labels are used.
ad hoc to paper Inverse encoder functions f_v^{-1} and f_t^{-1} exist and can reconstruct images and text from embeddings.
Equation 7 uses these inverses for the reconstruction loss, but no decoder architecture, pretrained weights, or training procedure is provided.

invented entities (2)

Inverse visual encoder f_v^{-1} and inverse text encoder f_t^{-1}
purpose: Reconstruct original image and text from embeddings to compute reconstruction loss in Eq. 7.
No architecture, parameters, or pretrained weights are specified; they are introduced ad hoc for the loss.
Contextualized reasoning module
purpose: Generate textual explanations for detected anomalies, claimed as a contribution in Section 1.
Never defined in the method section and never evaluated in the experiments; it exists only in the contributions list.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Exploring Large Vision-Language Models for Robust and Efficient Industrial Anomaly Detection." pith.science (2026). https://pith.science/paper/GJNEEYHG

@misc{pith2026241200890,
  author       = {Pith},
  title        = {Pith review of: Exploring Large Vision-Language Models for Robust and Efficient Industrial Anomaly Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GJNEEYHG}},
  note         = {Machine review of arXiv:2412.00890}
}

read the original abstract

Industrial anomaly detection (IAD) plays a crucial role in the maintenance and quality control of manufacturing processes. In this paper, we propose a novel approach, Vision-Language Anomaly Detection via Contrastive Cross-Modal Training (CLAD), which leverages large vision-language models (LVLMs) to improve both anomaly detection and localization in industrial settings. CLAD aligns visual and textual features into a shared embedding space using contrastive learning, ensuring that normal instances are grouped together while anomalies are pushed apart. Through extensive experiments on two benchmark industrial datasets, MVTec-AD and VisA, we demonstrate that CLAD outperforms state-of-the-art methods in both image-level anomaly detection and pixel-level anomaly localization. Additionally, we provide ablation studies and human evaluation to validate the importance of key components in our method. Our approach not only achieves superior performance but also enhances interpretability by accurately localizing anomalies, making it a promising solution for real-world industrial applications.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 17 canonical work pages

[1]

In: Findings of the Association for Comput ational Linguistics, ACL 2024, Bangkok, Thailand and virtual meeting, August 11- 16, 2024

Zhou, Y., Li, X., Wang, Q., Shen, J.: Visual in-context lea rning for large vision- language models. In: Findings of the Association for Comput ational Linguistics, ACL 2024, Bangkok, Thailand and virtual meeting, August 11- 16, 2024. pp. 15890– 15902. Association for Computational Linguistics (2024)

2024
[2]

arXiv preprin t arXiv:2410.19732 (2024)

Zhou, Y., Rao, Z., Wan, J., Shen, J.: Rethinking visual dep endency in long-context reasoning for large vision-language models. arXiv preprin t arXiv:2410.19732 (2024)

arXiv 2024
[3]

In: Proceedings of the IEEE/CVF conference on computer vision and pattern recogni tion

Bergmann, P., Fauser, M., Sattlegger, D., Steger, C.: Mvt ec ad–a comprehen- sive real-world dataset for unsupervised anomaly detectio n. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recogni tion. pp. 9592–9600 (2019)

work page 2019
[4]

In: European Conference on Computer Vision

Zou, Y., Jeong, J., Pemula, L., Zhang, D., Dabeer, O.: Spot -the-diﬀerence self- supervised pre-training for anomaly detection and segment ation. In: European Conference on Computer Vision. pp. 392–408. Springer (2022 )

work page 2022
[5]

In: Wooldridge, M.J., 12 K

Gu, Z., Zhu, B., Zhu, G., Chen, Y., Tang, M., Wang, J.: Anoma lygpt: Detect- ing industrial anomalies using large vision-language mode ls. In: Wooldridge, M.J., 12 K. Qian et al. Dy, J.G., Natarajan, S. (eds.) Thirty-Eighth AAAI Conferen ce on Artiﬁcial Intelli- gence, AAAI 2024, Thirty-Sixth Conference on Innovative Ap plications of Artiﬁcial Intelligenc...

2024
[7]

In: Proceedi ngs of the IEEE/CVF conference on computer vision and pattern recognition

Roth, K., Pemula, L., Zepeda, J., Schölkopf, B., Brox, T., Gehler, P.: Towards total recall in industrial anomaly detection. In: Proceedi ngs of the IEEE/CVF conference on computer vision and pattern recognition. pp. 14318–14328 (2022)

work page 2022
[8]

In: Meila, M., Zhang, T

Radford, A., Kim, J.W., Hallacy, C., Ramesh, A., Goh, G., A garwal, S., Sas- try, G., Askell, A., Mishkin, P., Clark, J., Krueger, G., Sut skever, I.: Learn- ing transferable visual models from natural language super vision. In: Meila, M., Zhang, T. (eds.) Proceedings of the 38th International C onference on Machine Learning, ICML 2021, 18-24 July 2021, V...

work page 2021
[9]

In: Proceedings of the 17th Conference of the Euro pean Chapter of the Association for Computational Linguistics

Zhou, Y., Long, G.: Improving cross-modal alignment for t ext-guided image in- painting. In: Proceedings of the 17th Conference of the Euro pean Chapter of the Association for Computational Linguistics. pp. 3445–3456 (2023)

2023

Show all 29 references

[10]

In: Proceedings of the 17th Conference of the Eu ropean Chapter of the Association for Computational Linguistics

Zhou, Y., Long, G.: Multimodal event transformer for ima ge-guided story ending generation. In: Proceedings of the 17th Conference of the Eu ropean Chapter of the Association for Computational Linguistics. pp. 3434–3444 (2023)

2023
[11]

In: Findings of the Association for Computational Ling uistics: EACL 2023

Zhou, Y., Long, G.: Style-aware contrastive learning fo r multi-style image caption- ing. In: Findings of the Association for Computational Ling uistics: EACL 2023. pp. 2257–2267 (2023)

2023
[12]

UGC Care Group I Journal 8(14), 71–75 (2021)

Reddy, M.D.M., Basha, M.S.M., Hari, M.M.C., Penchalaia h, M.N.: Dall-e: Creating images from text. UGC Care Group I Journal 8(14), 71–75 (2021)

2021
[13]

CoRR abs/1908.03557 (2019), http://arxiv.org/abs/1908.03557

Li, L.H., Yatskar, M., Yin, D., Hsieh, C., Chang, K.: Visu albert: A simple and performant baseline for vision and language. CoRR abs/1908.03557 (2019), http://arxiv.org/abs/1908.03557

2019 arXiv
[14]

In: Proceedings of the 60th Annual Meeting o f the Association for Computational Linguistics (Volume 1: Long Papers)

Zhou, Y., Shen, T., Geng, X., Long, G., Jiang, D.: Claret: Pre-training a correlation-aware context-to-event transformer for even t-centric generation and classiﬁcation. In: Proceedings of the 60th Annual Meeting o f the Association for Computational Linguistics (Volume 1: Lon...

2022
[15]

In: Proceedings of the ACM Web Conference 2022

Zhou, Y., Geng, X., Shen, T., Long, G., Jiang, D.: Eventbe rt: A pre-trained model for event correlation reasoning. In: Proceedings of the ACM Web Conference 2022. pp. 850–859 (2022)

2022
[16]

C oRR abs/1909.11740 (2019), http://arxiv.org/abs/1909.11740

Chen, Y., Li, L., Yu, L., Kholy, A.E., Ahmed, F., Gan, Z., C heng, Y., Liu, J.: UNITER: learning universal image-text representations. C oRR abs/1909.11740 (2019), http://arxiv.org/abs/1909.11740

2019 arXiv
[17]

In: 8th Int ernational Conference on L VLM for Industrial Anomaly Detection 13 Learning Representations, ICLR 2020, Addis Ababa, Ethiopi a, April 26-30, 2020

Su, W., Zhu, X., Cao, Y., Li, B., Lu, L., Wei, F., Dai, J.: VL -BERT: pre-training of generic visual-linguistic representations. In: 8th Int ernational Conference on L VLM for Industrial Anomaly Detection 13 Learning Representations, ICLR 2020, Addis Ababa, Ethiopi a, April 26...

2020
[18]

In: ICASSP 2021-2021 IEEE Inter national Conference on Acoustics, Speech and Signal Processing (ICASSP)

Zhou, Y., Tao, W., Zhang, W.: Triple sequence generative adversarial nets for unsu- pervised image captioning. In: ICASSP 2021-2021 IEEE Inter national Conference on Acoustics, Speech and Signal Processing (ICASSP). pp. 75 98–7602. IEEE (2021)

2021
[19]

In: ICASSP 2022-2022 IEE E International Confer- ence on Acoustics, Speech and Signal Processing (ICASSP)

Zhou, Y.: Sketch storytelling. In: ICASSP 2022-2022 IEE E International Confer- ence on Acoustics, Speech and Signal Processing (ICASSP). p p. 4748–4752. IEEE (2022)

2022
[20]

Advances in neural information processing systems 34, 9694–9705 (2021)

Li, J., Selvaraju, R., Gotmare, A., Joty, S., Xiong, C., H oi, S.C.H.: Align before fuse: Vision and language representation learning with momentum distillation. Advances in neural information processing systems 34, 9694–9705 (2021)

2021
[21]

, Hu, H., Huang, X., Li, B., Li, C., Liu, C., Liu, M., Liu, Z., Lu, Y., Shi, Y., Wa ng, L., Wang, J., Xiao, B., Xiao, Z., Yang, J., Zeng, M., Zhou, L., Zhang, P

Yuan, L., Chen, D., Chen, Y., Codella, N., Dai, X., Gao, J. , Hu, H., Huang, X., Li, B., Li, C., Liu, C., Liu, M., Liu, Z., Lu, Y., Shi, Y., Wa ng, L., Wang, J., Xiao, B., Xiao, Z., Yang, J., Zeng, M., Zhou, L., Zhang, P. : Florence: A new foundation model for computer vision....

2021 arXiv
[22]

Procedia CIRP 93, 1281–1285 (2020)

Heger, J., Desai, G., El Abdine, M.Z.: Anomaly detection in formed sheet metals using convolutional autoencoders. Procedia CIRP 93, 1281–1285 (2020)

2020
[23]

In: Tetko, I.V., Kurková, V., Karpov, P., Theis, F.J

Li, D., Chen, D., Jin, B., Shi, L., Goh, J., Ng, S.: MAD-GAN : multivariate anomaly detection for time series data with generative adve rsarial networks. In: Tetko, I.V., Kurková, V., Karpov, P., Theis, F.J. (eds.) Art iﬁcial Neural Networks and Machine Learning - ICANN 2019: ...

2019
[24]

In: Practi- cal Applications of Data Processing, Algorithms, and Model ing, pp

Palakurti, N.R.: Challenges and future directions in an omaly detection. In: Practi- cal Applications of Data Processing, Algorithms, and Model ing, pp. 269–284. IGI Global (2024)

2024
[25]

IEEE Trans

Dosovitskiy, A., Fischer, P., Springenberg, J.T., Ried miller, M.A., Brox, T.: Discriminative unsupervised feature learning with exe mplar con- volutional neural networks. IEEE Trans. Pattern Anal. Mach . Intell. 38(9), 1734–1747 (2016). https://doi.org/10.1109/TPAMI.2015.2496...

2016
[26]

arXiv preprint arXiv:2405.0 3673 (2024)

Wang, Q., Hu, H., Zhou, Y.: Memorymamba: Memory-augment ed state space model for defect recognition. arXiv preprint arXiv:2405.0 3673 (2024)

2024
[27]

In: 2020 International Conference on Electronics and Sustainable C ommunication Systems (ICESC)

Parsai, S., Mahajan, S.: Anomaly detection using long sh ort-term memory. In: 2020 International Conference on Electronics and Sustainable C ommunication Systems (ICESC). pp. 333–337. IEEE (2020)

2020
[28]

ISPRS Inter national Journal of Geo-Information 10(3), 177 (2021)

Zou, H., Cao, K., Jiang, C.: Spatio-temporal visual anal ysis for urban traﬃc char- acters based on video surveillance camera data. ISPRS Inter national Journal of Geo-Information 10(3), 177 (2021)

2021
[29]

In: Bimbo, A.D., Cucchiara, R., Sclaroﬀ, S., Farinella, G.M., M ei, T., Bertini, M., Escalante, H.J., Vezzani, R

Defard, T., Setkov, A., Loesch, A., Audigier, R.: Padim: A patch dis- tribution modeling framework for anomaly detection and loc alization. In: Bimbo, A.D., Cucchiara, R., Sclaroﬀ, S., Farinella, G.M., M ei, T., Bertini, M., Escalante, H.J., Vezzani, R. (eds.) Pattern Recognit...

2020 doi
[30]

I n: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni tion

Jeong, J., Zou, Y., Kim, T., Zhang, D., Ravichandran, A., Dabeer, O.: Winclip: Zero-/few-shot anomaly classiﬁcation and segmentation. I n: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni tion. pp. 19606– 19616 (2023)

2023

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

In: Findings of the Association for Comput ational Linguistics, ACL 2024, Bangkok, Thailand and virtual meeting, August 11- 16, 2024

Zhou, Y., Li, X., Wang, Q., Shen, J.: Visual in-context lea rning for large vision- language models. In: Findings of the Association for Comput ational Linguistics, ACL 2024, Bangkok, Thailand and virtual meeting, August 11- 16, 2024. pp. 15890– 15902. Association for Computational Linguistics (2024)

2024

[2] [2]

arXiv preprin t arXiv:2410.19732 (2024)

Zhou, Y., Rao, Z., Wan, J., Shen, J.: Rethinking visual dep endency in long-context reasoning for large vision-language models. arXiv preprin t arXiv:2410.19732 (2024)

arXiv 2024

[3] [3]

In: Proceedings of the IEEE/CVF conference on computer vision and pattern recogni tion

Bergmann, P., Fauser, M., Sattlegger, D., Steger, C.: Mvt ec ad–a comprehen- sive real-world dataset for unsupervised anomaly detectio n. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recogni tion. pp. 9592–9600 (2019)

work page 2019

[4] [4]

In: European Conference on Computer Vision

Zou, Y., Jeong, J., Pemula, L., Zhang, D., Dabeer, O.: Spot -the-diﬀerence self- supervised pre-training for anomaly detection and segment ation. In: European Conference on Computer Vision. pp. 392–408. Springer (2022 )

work page 2022

[5] [5]

In: Wooldridge, M.J., 12 K

Gu, Z., Zhu, B., Zhu, G., Chen, Y., Tang, M., Wang, J.: Anoma lygpt: Detect- ing industrial anomalies using large vision-language mode ls. In: Wooldridge, M.J., 12 K. Qian et al. Dy, J.G., Natarajan, S. (eds.) Thirty-Eighth AAAI Conferen ce on Artiﬁcial Intelli- gence, AAAI 2024, Thirty-Sixth Conference on Innovative Ap plications of Artiﬁcial Intelligenc...

2024

[6] [7]

In: Proceedi ngs of the IEEE/CVF conference on computer vision and pattern recognition

Roth, K., Pemula, L., Zepeda, J., Schölkopf, B., Brox, T., Gehler, P.: Towards total recall in industrial anomaly detection. In: Proceedi ngs of the IEEE/CVF conference on computer vision and pattern recognition. pp. 14318–14328 (2022)

work page 2022

[7] [8]

In: Meila, M., Zhang, T

Radford, A., Kim, J.W., Hallacy, C., Ramesh, A., Goh, G., A garwal, S., Sas- try, G., Askell, A., Mishkin, P., Clark, J., Krueger, G., Sut skever, I.: Learn- ing transferable visual models from natural language super vision. In: Meila, M., Zhang, T. (eds.) Proceedings of the 38th International C onference on Machine Learning, ICML 2021, 18-24 July 2021, V...

work page 2021

[8] [9]

In: Proceedings of the 17th Conference of the Euro pean Chapter of the Association for Computational Linguistics

Zhou, Y., Long, G.: Improving cross-modal alignment for t ext-guided image in- painting. In: Proceedings of the 17th Conference of the Euro pean Chapter of the Association for Computational Linguistics. pp. 3445–3456 (2023)

2023