The DOI matches its Crossref record; 35 of the 38 references it cites were found in a registry or an index; 3 of the 13 sources its claims cite were checked for support.
VT-ADL: A Vision Transformer Network for Image Anomaly Detection and Localization
Open accessAbstract
We present a transformer-based image anomaly detection and localization network. Our proposed model is a combination of a reconstruction-based approach and patch embedding. The use of transformer networks helps preserving the spatial information of the embedded patches, which is later processed by a Gaussian mixture density network to localize the anomalous areas. In addition, we also publish BTAD, a real-world industrial anomaly dataset. Our results are compared with other state-of-the-art algorithms using publicly available datasets like MNIST and MVTec.
Overview
CoChat Summary
TL;DR
VT-ADL combines a Vision Transformer patch encoder, a reconstruction decoder, and a Gaussian mixture density network for semi-supervised image anomaly detection and localization. It also releases BTAD, a real-world industrial anomaly dataset, and reports competitive results on MNIST, MVTec, and BTAD.
Key findings
- On MNIST, VT-ADL achieves a mean AUC of 0.984, compared with 0.95 for OC SVM, 0.97 for VAE, 0.97 for LSA, and 0.97 for Pyr. AE.
- On MVTec, VT-ADL reaches a mean PRO score of 0.807, performing at par with recent state-of-the-art methods and outperforming them in 7 product categories; the best comparator, Uni. Stud., has a mean of 0.857.
- On the newly published BTAD dataset, VT-ADL reports a mean PRO score of 0.89 and mean PR AUC of 0.90, compared with convolutional autoencoders using MSE loss (0.78) and MSE+SSIM loss (0.79).
- Adding Gaussian noise to transformer-encoded features improves generalization: with 150 Gaussians, PRO score is 0.897 with noise versus 0.807 without noise.
Methods
- The model is trained semi-supervised using only normal data. Images are split into patches, embedded with positional information, encoded by a Vision Transformer, decoded for reconstruction, and modeled by a Gaussian mixture density network for anomaly localization.
- Training combines MSE, SSIM, and negative conditional log-likelihood losses, with Gaussian noise N(0, 0.2) added to transformer-embedded features for regularization. Hyperparameters include Adam learning rate 0.0001, weight decay 0.0001, batch size 8, 400 epochs, 150 Gaussians, and patch dimension P = 64.
- Experiments use MNIST, MVTec (5,354 high-resolution images, scaled to 550×550 then center cropped to 512×512), and BTAD (2,830 real-world images of 3 industrial products, with 400, 1000, and 399 training images respectively). Evaluation uses AUC for MNIST and PRO for MVTec and BTAD.
Limitations
- The paper does not include a dedicated limitations section. It states that dropout was not used because it causes instability in the Gaussian approximation network.
- Evident from the reported results: on MVTec, VT-ADL’s mean PRO (0.807) is below Uni. Stud. (0.857), and it underperforms that method in several categories, such as Capsule (0.672 vs. 0.916), Pill (0.705 vs. 0.935), and Metal Nut (0.726 vs. 0.895).
- Evident from the text: BTAD results are described as the first results on the new dataset, and the comparison is limited to basic convolutional autoencoders with MSE and MSE+SSIM losses.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures

![Fig. 2 Left image: model overview. Image is split into patches, which are augmented with positional embedding. The resulting sequence is fed to the Transformer encoder. Then encoded features are summed into a reconstruction vector which is fed to decoder. The transformer encoded features are also fed into an Gaussian approximation network [10], which is later used to localize the anomaly. Right image: detailed structure of the tranformer encoder (image from [9]).](https://figures.cochat.ai/paper-figures/373c6fb2a8524c9a86329165d4328f84/vt-adl-a-vision-transformer-network-for-image-anomaly-detection-and-localization-figure-2.png)


Verification
Claims the paper makes
2 Valid · 2 Not valid · 6 Not checked · 14 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“In our work we show how an adapted transformer network can be used for anomaly localization using Gaussian Approximation [10], [11] of the latent features and also how different configurations can be tweaked to win some of the shortcomings of the vision transformer network.”
What the sources say
- Partly supportedcites [10] [10] C. M. Bishop, Mixture density networks . Aston University, 1994. -Experimental — accuracy not yet measured
The paper’s “Conventional Least Squares” section describes representing the conditional distribution of target data with a Gaussian function. It does not describe Gaussian approximation of latent features for anomaly localization. The adapted transformer is presented as the citing authors’ own work, not as a result of this source. (LLM, conf=0.82)
- Not checked — no full text of the sourcecites [11] [11] N. Ueda, R. Nakano, Z. Ghahramani, and G. E. Hinton, 'Split and merge em algorithm for improving gaussian mixture density estimates,' in Neural Networks for Signal Processing VIII. Proceedings of the 1998 IEEE Signal Processing Society Workshop (Cat. No. 98TH8378) . IEEE, 1998, pp. 274-283. - source
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“The use of transformer networks helps preserving the spatial information of the embedded patches, which is later processed by a Gaussian mixture density network to localize the anomalous areas.”
- Also stated in: § I. INTRODUCTION · p. 1 · plain
“The vision transformer network model, recently proposed by Dosovitskiy et al. [9], is a network designed to work on image patches trying to preserve their positional information.”
What the sources say
- Supportedcites Dosovitskiy et al. [9] [9] A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Minderer, G. Heigold, S. Gelly, J. Uszkoreit, and N. Houlsby, 'An image is worth 16x16 words: Transformers for image recognition at scale,' 2020. - source
Section 3.1 (p. 3) says position embeddings are added to patch embeddings to retain positional information before the sequence enters the Transformer encoder. Section 4.5 (p. 8) reports that the learned embeddings represent the image’s 2D topology, supporting the attributed point about spatial information. (LLM, conf=0.91)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Also stated in: § I. INTRODUCTION · p. 1 · plain
The cited sources that could be checked couldn’t confirm or dispute this claim.
In the paper
“Adding noise during training is seen as a form of data augmentation and regularization that biases towards smooth functions [10], [31].”
What the sources say
- Could not verifycites [10] [10] C. M. Bishop, Mixture density networks . Aston University, 1994. -Experimental — accuracy not yet measured
Section 2, “Conventional Least Squares” (page unavailable), mentions regularization as a way to control model complexity, but does not connect it to adding noise. Section 5, “A Simple Inverse Problem” (page unavailable), uses noise to generate example data; it does not describe noise added during training as augmentation or as a bias toward smooth functions. Searches for noise injection and smooth functions found no passage addressing that claim. (LLM, conf=0.76)
- Not checked — no full text of the sourcecites [31] [31] G. An, 'The effects of adding noise during backpropagation training on a generalization performance,' Neural Computation , vol. 8, no. 3, pp. 643-674, 1996. source
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Could not verify
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The dataset contains a total of 2830 real-world images of 3 industrial products showcasing body and surface defects.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“In addition, we also publish BTAD, a real-world industrial anomaly dataset.”
- Also stated in: § I. INTRODUCTION · p. 1 · plain
“In addition to this, we are also publishing a real-world industrial dataset (the beanTech Anomaly Detection dataset - BTAD) for the anomaly detection task.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Also stated in: § I. INTRODUCTION · p. 1 · plain
8 more claims
Create a free account to see all 24 claims and to verify the 3 references still open.
- It may state two separate claims as one. 10
- The quote only partly states this claim. 1
- Probably not a substantial claim. 2
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“The BeanTech AD dataset, designed for unsupervised anomaly detection in industrial settings”
From An Unsupervised Method for Industrial Image Anomaly Detection with Vision Transformer-Based Autoencoder (2024) · cites this paper as [13]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“These datasets are extensively utilized to evaluate anomaly detection algorithms, offering diverse testing scenarios and data for this research.”
From An Unsupervised Method for Industrial Image Anomaly Detection with Vision Transformer-Based Autoencoder (2024) · cites this paper as [13]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“However, these methods directly utilize vanilla transformer, and do not figure out why transformer brings improvement.”
From A Unified Model for Multi-Class Anomaly Detection (2022) · cites this paper as [29]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“It comprises 2830 RGB images across three different industrial products.”
From An Unsupervised Method for Industrial Image Anomaly Detection with Vision Transformer-Based Autoencoder (2024) · cites this paper as [13]
Citations
Cites
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017) · arXiv · doi:10.48550/arxiv.1706.03762 · Cited by 120,404
An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., et al. (2020) · arXiv · doi:10.48550/arxiv.2010.11929 · Cited by 47,472 · cited 5×
“[9], is a network designed to work on image patches”
BackgroundMethodInfluentialMVTec AD — A Comprehensive Real-World Dataset for Unsupervised Anomaly Detection
Bergmann, P., Fauser, M., Sattlegger, D., & Steger, C. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00982 · Cited by 2,203 · cited 2×
“In this case the system tries to learn a “normality” model from the training data and thus subsequently classify new samples as anomaly if they don’t fit the model [1], [16], [12], [24].”
BackgroundMethod(2011) · Cited by 1,258
(1994) · Aston Publications Explorer (Aston University) · Cited by 1,245
High-dimensional and large-scale anomaly detection using a linear one-class SVM with deep learning
Erfani, S. M., Rajasegarar, S., Karunasekera, S., & Leckie, C. (2016) · Pattern Recognition · doi:10.1016/j.patcog.2016.03.028 · Cited by 1,162 · cited 1×
“Some of the works handled these two steps independently [25], [26], while others achieved better results by doing the two steps collectively, in order to extract the best features for subsequent anomaly detection [27], [28].”
Uninformed Students: Student-Teacher Anomaly Detection With Discriminative Latent Embeddings
Bergmann, P., Fauser, M., Sattlegger, D., & Steger, C. (2019) · Computer Vision and Pattern Recognition · doi:10.48550/arxiv.1911.02357 · Cited by 1,052
Uninformed Students: Student-Teacher Anomaly Detection With Discriminative Latent Embeddings
Bergmann, P., Fauser, M., Sattlegger, D., & Steger, C. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00424 · Cited by 990 · cited 5×
“Then we employed the PRO (Per Region Overlap) [8], [24] as the evaluation metric for the MVTec and BTAD datasets.”
BackgroundMethodInfluentialAdversarially Learned One-Class Classifier for Novelty Detection
Sabokrou, M., Khalooei, M., Fathy, M., & Adeli, E. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.48550/arxiv.1802.09088 · Cited by 768
Adversarially Learned One-Class Classifier for Novelty Detection
Sabokrou, M., Khalooei, M., Fathy, M., & Adeli, E. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00356 · Cited by 765 · cited 1×
“Network architectures mostly consist of various configuration of autoencoders [12], [13], [14], [15], [16] or Generative Adversarial Network (GAN) [17], [18].”
Background
Research works printed but not linked: 1
- goodfellow, 2016, Deep Learning
Cited by
WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentation
Jeong, J., Zou, Y., Kim, T., Zhang, D., Ravichandran, A., & Dabeer, O. (2023) · 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52729.2023.01878 · Cited by 585
“Additionally, Vision Transformer (ViT) have recently shown its potential in visual inspection [9, 25].”
BackgroundDeep Industrial Image Anomaly Detection: A Survey
Liu, J., Xie, G., Wang, J., Li, S., Wang, C., Zheng, F., & Jin, Y. (2024) · Machine Intelligence Research · doi:10.1007/s11633-023-1459-z · Cited by 463
“(3) Transformer Model VT-ADL [107] L2, SSIM, Log-Likelihood The paper proposes a transformer-based framework to reconstruct images and detects anomalies.”
InfluentialEfficientAD: Accurate Visual Anomaly Detection at Millisecond-Level Latencies
Batzner, K., Heckler, L., & König, R. (2024) · 2024 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) · doi:10.1109/wacv57701.2024.00020 · Cited by 386
“Recently, several new industrial anomaly detection datasets have been introduced [8, 11, 25, 27, 37, 69].”
BackgroundMethodRevisiting Reverse Distillation for Anomaly Detection
Tien, T. D., Nguyen, A. T., Tran, N. H., Huy, T. D., Duong, S. T. M., Nguyen, C. D. T., & Truong, S. Q. H. (2023) · 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52729.2023.02348 · Cited by 268
“Anomaly detection and localization results in terms of AUROC at image-level/pixel-level/PRO on the BTAD dataset [22].”
BackgroundMethodInfluentialYang, M., Wu, P., & Feng, H. (2023) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2023.105835 · Cited by 257
Metadata
- Authors
- Mishra, P., Verk, R., Fornasier, D., Piciarelli, C., & Foresti, G. L.
- Affiliations
- University of Udine (IT)
- Venue
- 2021 IEEE 30th International Symposium on Industrial Electronics (ISIE)
- Publisher
- IEEE
- Published
- 2021-06-20
- Pages
- 01-06
- OpenAlex
- W3154719617
- Cited by
- 462
- References
- 38
- Citations by year
- 2026: 91 · 2025: 139 · 2024: 124 · 2023: 81 · 2022: 26 · 2021: 1
- FWCI
- 26.74
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- Locations
- IEEE International Symposium on Industrial Electronics (ISIE) (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- Institutional Research Information System (University of Udine) (submittedVersion)
- arXiv (Cornell University)
- Funders
- ONRG (N62909-20-1-2075)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Anomaly Detection Techniques and Applications
- Physical Sciences › Engineering › Industrial and Manufacturing Engineering › Industrial Vision Systems and Defect Detection
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Keywords
- image anomaly detection, anomaly localization, vision transformer, patch embedding, reconstruction-based anomaly detection
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: VT-ADL: A Vision Transformer Network for Image Anomaly Detection and Localization. https://app.cochat.ai/papers/10.1109/isie45552.2021.9576231
- The paper
- Mishra, P., Verk, R., Fornasier, D., Piciarelli, C., & Foresti, G. L. (2021). VT-ADL: A Vision Transformer Network for Image Anomaly Detection and Localization. 2021 IEEE 30th International Symposium on Industrial Electronics (ISIE). https://doi.org/10.1109/isie45552.2021.9576231
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.