Pith. sign in

REVIEW 2 major objections 60 references

Leading 1/c² relativistic corrections to quantum-harmonic-oscillator wave-packet widths and variances admit closed forms, and for electrons at 1–10 keV they shift variances by 0.15–1.5 percent while leaving uncertainty-relation saturation i

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · grok-4.5

2026-07-15 12:55 UTC pith:AED5IQ4H

load-bearing objection Abstract promises clean O(1/c^{2}) closed forms for QHO wave-packet widths and a concrete 0.15–1.5 % electron claim, but the supplied full text is the unrelated VisualAD vision paper, so nothing can be checked. the 2 major comments →

arxiv 2603.07954 v2 pith:AED5IQ4H submitted 2026-03-09 quant-ph hep-th

Closed form perturbative relativistic modifications to wave-packet dynamics in the quantum harmonic oscillator

Jian Carlo Ramos , Sujoy K. Modak This is my paper
classification quant-ph hep-th MSC 81Q0581Q15 PACS 03.65.Ge03.65.Sq03.30.+p
keywords quantum harmonic oscillatorrelativistic correctionswave-packet dynamicsGaussian wave packetsuncertainty relationperturbative 1/c² expansionelectron confinement
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper claims that the ordinary non-relativistic dynamics of Gaussian wave packets in a quantum harmonic oscillator can be corrected, to first order in 1/c², by closed-form analytic expressions for the time-dependent widths, variances and uncertainty products. The authors work inside a perturbative expansion of the relativistic kinetic energy and specialise to unsqueezed, minimum-uncertainty Gaussians. When the same formulae are evaluated for an electron, the corrections become appreciable once the harmonic confinement energy reaches the keV scale, producing relative shifts of a few tenths to a little more than one percent in the variances. At the same order the familiar equality that saturates the uncertainty relation is left unchanged. A reader who cares about precision wave-packet evolution therefore obtains ready-to-use analytic formulae and a concrete energy window in which those formulae matter.

Core claim

Closed-form expressions exist for the leading O(1/c²) relativistic corrections to the time-dependent widths, variances and uncertainty relations of unsqueezed minimum-uncertainty Gaussian wave packets in the quantum harmonic oscillator; for electrons those corrections produce 0.15 %–1.5 % deviations in the variances at 1–10 keV confinement energies, while the ordinary saturation of the uncertainty relation remains unaffected.

What carries the argument

A first-order perturbative expansion in 1/c² of the relativistic kinetic energy (or an equivalent Foldy–Wouthuysen effective Hamiltonian) applied to the time-dependent Gaussian wave-packet parameters of the quantum harmonic oscillator.

Load-bearing premise

That a first-order expansion in 1/c² remains uniformly valid for the entire time-dependent wave-packet dynamics, including the widths, at the stated keV-scale electron confinement energies.

What would settle it

Compare the analytic O(1/c²) variance formulae against a fully relativistic (Dirac or equivalent) numerical evolution of an electron Gaussian wave packet in a harmonic trap at 1–10 keV; systematic disagreement outside the claimed 0.15–1.5 % window would falsify the result.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • At keV-scale electron confinement the non-relativistic Gaussian variances must be replaced by the derived O(1/c²) expressions if percent-level accuracy is required.
  • The ordinary equality that saturates the Heisenberg uncertainty relation continues to hold once leading relativistic corrections are included.
  • The same closed-form machinery can be evaluated for any mass and any confinement energy without re-solving the dynamics.
  • Higher-order or non-Gaussian packets can be treated by the same perturbative route once the leading terms are known.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Precision electron optics or high-frequency ion traps operating near 1–10 keV may need these corrections for percent-level metrology.
  • A direct numerical check against the exact Dirac oscillator would quantify the domain of validity of the 1/c² truncation.
  • The same expansion could be repeated for squeezed or coherent states to map how relativity modifies non-classical packet shapes.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

2 major / 0 minor

Summary. The manuscript claims to derive closed-form expressions for the leading-order O(1/c^{2}) relativistic corrections to the time-dependent widths, variances, and uncertainty products of wave packets in the quantum harmonic oscillator. The abstract states that the results are obtained within a perturbative framework, specialized to unsqueezed minimum-uncertainty Gaussian packets, and that for electrons the variance deviations reach 0.15 %–1.5 % at 1–10 keV confinement energies while the ordinary saturation of the uncertainty relation remains intact. No further technical content of that derivation is present in the supplied document.

Significance. If the claimed closed-form O(1/c^{2}) formulas and the associated numerical estimates are correct, the work would supply a compact, parameter-light reference for assessing when relativistic corrections become non-negligible in harmonic-trap wave-packet dynamics—an issue of practical interest for high-energy electron optics and precision cold-atom or ion-trap experiments. The assertion that uncertainty saturation is unaffected at this order would also be a clean, falsifiable statement. At present, however, none of these results can be verified because the body of the manuscript is missing.

major comments (2)
  1. The document supplied under the label of arXiv:2603.07954 is in fact the unrelated computer-vision paper “VisualAD” (arXiv:2603.07952). Consequently every load-bearing claim of the abstract—existence of closed-form O(1/c^{2}) expressions for time-dependent widths/variances/uncertainty products, the concrete 0.15 %–1.5 % electron figures, and the statement that uncertainty saturation is unaffected—rests solely on the abstract. No Hamiltonian, expansion, intermediate equation, or numerical procedure can be inspected. The manuscript as submitted is therefore incomplete and cannot be refereed on its scientific merits.
  2. Even if the correct quant-ph text were restored, the abstract’s weakest premise—that a first-order 1/c^{2} expansion of the relativistic kinetic energy remains uniformly valid for the full time-dependent wave-packet dynamics at keV-scale electron confinement—would still require explicit justification (domain of validity of the Foldy–Wouthuysen or effective Hamiltonian, comparison with exact Dirac–oscillator numerics, etc.). That justification is absent from the material provided.

Circularity Check

0 steps flagged

No circularity in the claimed QHO perturbative derivation; supplied full text is an unrelated CV paper, so equation-level chain cannot be inspected.

full rationale

The abstract of arXiv:2603.07954 describes a standard first-principles O(1/c²) perturbative expansion of relativistic corrections to Gaussian wave-packet dynamics in the quantum harmonic oscillator, yielding closed-form expressions for time-dependent widths, variances, and uncertainty products, then applying those formulas to electron keV-scale confinement to obtain 0.15%–1.5% variance deviations and the statement that uncertainty saturation is unaffected. None of these steps is self-definitional, a fit renamed as prediction, or load-bearing on an author-only uniqueness theorem: the percentages are direct numerical applications of the derived formulas, not parameters fitted to the same quantities. The CACHEABLE full-manuscript block, however, is the unrelated VisualAD computer-vision paper (arXiv:2603.07952), so no equation, Hamiltonian, or derivation step of the quant-ph claim can be quoted or reduced. On the abstract’s own description the work is self-contained and non-circular; residual uncertainty is only the absence of the correct full text, not evidence of circular construction. Score 0 with empty steps is therefore the honest finding.

Axiom & Free-Parameter Ledger

0 free parameters · 3 axioms · 0 invented entities

With only the abstract, the ledger is necessarily incomplete. The calculation rests on standard quantum mechanics plus a weak-relativistic expansion; no free parameters fitted to data are mentioned. The principal domain assumptions are the validity of the O(1/c²) truncation and the use of an effective non-relativistic Hamiltonian with relativistic kinetic corrections for a harmonic potential.

axioms (3)
  • domain assumption Leading-order relativistic corrections can be captured by a perturbative expansion of the kinetic energy (or equivalent Foldy–Wouthuysen Hamiltonian) truncated at O(1/c²).
    Required for all closed-form expressions and the quoted percentages; stated as the working framework in the abstract.
  • domain assumption The initial state is an unsqueezed, minimum-uncertainty Gaussian wave packet of the non-relativistic harmonic oscillator.
    Specific calculations are restricted to this class; general expressions are claimed but the concrete results use this ansatz.
  • standard math Standard quantum harmonic oscillator algebra and Gaussian wave-packet propagation formulas remain the unperturbed baseline.
    Background theory assumed throughout.

pith-pipeline@v1.1.0-grok45 · 21908 in / 2598 out tokens · 23220 ms · 2026-07-15T12:55:20.745476+00:00 · methodology

0 comments
read the original abstract

We derive closed form expressions of weak relativistic corrections to the wave-packet dynamics of the quantum harmonic oscillator within a perturbative framework. General expressions are derived for the leading-order relativistic contributions to wave-packet parameters, such as the time-dependent widths, variances, and uncertainty relations. Specific calculations are performed for unsqueezed, minimum-uncertainty Gaussian wave packets, valid up to leading order in ($1/c^{2}$). When applied to electron wave packets, the results indicate that relativistic effects become non-negligible for keV-scale harmonic confinement energies: the deviations in variances reach $0.15\% - 1.5\%$ for an electron wave packet confined within the $1-10 \mathrm{keV}$ energy range. We also show that the standard saturation of the uncertainty relation remains unaffected by the leading-order relativistic effect.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

60 extracted references · 8 linked inside Pith

  1. [1]

    Zero-shot versus many-shot: Unsupervised texture anomaly detection

    Toshimichi Aota, Lloyd Teh Tzer Tong, and Takayuki Okatani. Zero-shot versus many-shot: Unsupervised texture anomaly detection. InProceedings of the IEEE/CVF Win- ter Conference on Applications of Computer Vision, pages 5564–5572, 2023. 5, 6

    2023

  2. [2]

    The rsna-asnr-miccai brats 2021 benchmark on brain tumor segmentation and radiogenomic classification.ArXiv, abs/2107.02314, 2021

    Ujjwal Baid, Satyam Ghodasara, and Michel Bilello. The rsna-asnr-miccai brats 2021 benchmark on brain tumor segmentation and radiogenomic classification.ArXiv, abs/2107.02314, 2021. 5, 6

    Pith/arXiv arXiv 2021

  3. [3]

    Kirby, John B

    Spyridon Bakas, Hamed Akbari, Aristeidis Sotiras, Michel Bilello, Martin Rozycki, Justin S. Kirby, John B. Freymann, Keyvan Farahani, and Christos Davatzikos. Advancing the cancer genome atlas glioma mri collections with expert seg- mentation labels and radiomic features.Scientific Data, 4,

  4. [4]

    Diagnostic assessment of deep learning algo- rithms for detection of lymph node metastases in women with breast cancer.JAMA, 318:2199–2210, 2017

    Babak Ehteshami Bejnordi, Mitko Veta, and Paul Johannes van Diest. Diagnostic assessment of deep learning algo- rithms for detection of lymph node metastases in women with breast cancer.JAMA, 318:2199–2210, 2017. 5, 6

    2017

  5. [5]

    Yao, Guillaume Alain, and Pascal Vin- cent

    Yoshua Bengio, L. Yao, Guillaume Alain, and Pascal Vin- cent. Generalized denoising auto-encoders as generative models.ArXiv, abs/1305.6663, 2013. 3

    Pith/arXiv arXiv 2013

  6. [6]

    Mvtec ad — a comprehensive real-world dataset for unsupervised anomaly detection

    Paul Bergmann, Michael Fauser, David Sattlegger, and Carsten Steger. Mvtec ad — a comprehensive real-world dataset for unsupervised anomaly detection. In2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2019. 5, 6

    2019

  7. [7]

    Wm-dova maps for accurate polyp high- lighting in colonoscopy: Validation vs

    Jorge Bernal, Francisco Javier S ´anchez, Gl `oria Fern ´andez- Esparrach, Debora Gil, Cristina Rodr ´ıguez de Miguel, and Fernando Vilari˜no. Wm-dova maps for accurate polyp high- lighting in colonoscopy: Validation vs. saliency maps from physicians.Computerized medical imaging and graphics : the official journal of the Computerized Medical Imaging So- c...

    2015

  8. [8]

    Segment any anomaly without training via hybrid prompt regularization.CoRR, abs/2305.10724, 2023

    Yunkang Cao, Xiaohao Xu, Chen Sun, Yuqi Cheng, Zongwei Du, Liang Gao, and Weiming Shen. Segment any anomaly without training via hybrid prompt regularization.CoRR, abs/2305.10724, 2023. 3

    Pith/arXiv arXiv 2023

  9. [9]

    Adaclip: Adapting clip with hybrid learnable prompts for zero-shot anomaly de- tection, 2024

    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, 2024. 2, 3, 5, 6

    2024

  10. [10]

    Xuhai Chen and et al. A zero-/few-shot anomaly classifica- tion and segmentation method for CVPR 2023 V AND work- shop challenge tracks 1&2: 1st place on zero-shot AD and 4th place on few-shot AD.CoRR, abs/2305.17382, 2023. 2, 5, 6

    Pith/arXiv arXiv 2023

  11. [11]

    CLIP-AD: A language-guided staged dual-path model for zero-shot anomaly detection.CoRR, abs/2311.00453, 2023

    Xuhai Chen, Jiangning Zhang, Guanzhong Tian, Haoyang He, Wuhao Zhang, Yabiao Wang, Chengjie Wang, Yunsheng Wu, and Yong Liu. CLIP-AD: A language-guided staged dual-path model for zero-shot anomaly detection.CoRR, abs/2311.00453, 2023. 2, 3, 5, 6

    Pith/arXiv arXiv 2023

  12. [12]

    Padim: a patch distribution modeling framework for anomaly detection and localization

    Thomas Defard, Aleksandr Setkov, Ang ´elique Loesch, and Romaric Audigier. Padim: a patch distribution modeling framework for anomaly detection and localization. InICPR Workshops, 2020. 3

    2020

  13. [13]

    An image is worth 16x16 words: Trans- formers for image recognition at scale

    Alexey Dosovitskiy, Lucas Beyer, and Alexander Kolesnikov. An image is worth 16x16 words: Trans- formers for image recognition at scale. In9th International Conference on Learning Representations, ICLR, 2021. 3

    2021

  14. [14]

    Metauas: Universal anomaly segmentation with one-prompt meta-learning.ArXiv, abs/2505.09265,

    Bin-Bin Gao. Metauas: Universal anomaly segmentation with one-prompt meta-learning.ArXiv, abs/2505.09265,

    Pith/arXiv arXiv

  15. [15]

    Learning to detect multi-class anomalies with just one normal image prompt

    Bin-Bin Gao. Learning to detect multi-class anomalies with just one normal image prompt. InEuropean Conference on Computer Vision, 2025. 3

    2025

  16. [16]

    Adaptclip: Adapting clip for universal vi- sual anomaly detection.ArXiv, abs/2505.09926, 2025

    Bin-Bin Gao, Yue Zhou, Jiangtao Yan, Yuezhi Cai, Weixi Zhang, Meng Wang, Jun Liu, Yong Liu, Lei Wang, and Chengjie Wang. Adaptclip: Adapting clip for universal vi- sual anomaly detection.ArXiv, abs/2505.09926, 2025. 2, 3

    Pith/arXiv arXiv 2025

  17. [17]

    Dong Gong, Lingqiao Liu, Vuong Le, Budhaditya Saha, Moussa Reda Mansour, Svetha Venkatesh, and Anton van den Hengel. Memorizing normality to detect anomaly: Memory-augmented deep autoencoder for unsupervised anomaly detection.2019 IEEE/CVF International Confer- ence on Computer Vision (ICCV), pages 1705–1714, 2019. 3

    2019

  18. [18]

    Memorizing normality to detect anomaly: Memory-augmented deep autoencoder for unsupervised anomaly detection

    Dong Gong, Lingqiao Liu, Vuong Le, Budhaditya Saha, Moussa Reda Mansour, Svetha Venkatesh, and Anton van den Hengel. Memorizing normality to detect anomaly: Memory-augmented deep autoencoder for unsupervised anomaly detection. InIEEE/CVF International Conference on Computer Vision, ICCV, pages 1705–1714. IEEE, 2019. 2

    2019

  19. [19]

    Halvorsen, Hugo Lewi Hammer, and M

    Steven Hicks, Debesh Jha, Vajira Lasantha Thambawita, P. Halvorsen, Hugo Lewi Hammer, and M. Riegler. The endo- tect 2020 challenge: Evaluation and comparison of classifi- cation, segmentation and inference time for endoscopy. In ICPR Workshops, 2020. 5, 6

    2020

  20. [20]

    Enhancing zero-shot anomaly detection: Clip-sam col- laboration with cascaded prompts

    Yanning Hou, Ke Xu, Junfa Li, Yanran Ruan, and Jianfeng Qiu. Enhancing zero-shot anomaly detection: Clip-sam col- laboration with cascaded prompts. InChinese Conference on Pattern Recognition and Computer Vision, 2024. 3

    2024

  21. [21]

    Stackclip: Clustering-driven stacked prompt in zero-shot industrial anomaly detection.ArXiv, abs/2506.23577, 2025

    Yanning Hou, Yanran Ruan, Junfa Li, Shanshan Wang, Jian- feng Qiu, and Ke Xu. Stackclip: Clustering-driven stacked prompt in zero-shot industrial anomaly detection.ArXiv, abs/2506.23577, 2025. 2

    Pith/arXiv arXiv 2025

  22. [22]

    Soft reasoning paths for knowledge graph comple- tion

    Yanning Hou, Sihang Zhou, Ke Liang, Lingyuan Meng, Xi- aoshu Chen, Ke Xu, Siwei Wang, Xinwang Liu, and Jian Huang. Soft reasoning paths for knowledge graph comple- tion. InInternational Joint Conference on Artificial Intelli- gence, 2025. 3

    2025

  23. [23]

    Winclip: Zero- /few-shot anomaly classification and segmentation

    Jongheon Jeong, Yang Zou, Taewan Kim, Dongqing Zhang, Avinash Ravichandran, and Onkar Dabeer. Winclip: Zero- /few-shot anomaly classification and segmentation. In IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR, pages 19606–19616. IEEE, 2023. 2, 3, 5, 6

    2023

  24. [24]

    Riegler, P

    Debesh Jha, Pia Helen Smedsrud, M. Riegler, P. Halvorsen, Thomas de Lange, Dag Johansen, and Haavard D. Johansen. Kvasir-seg: A segmented polyp dataset. InConference on Multimedia Modeling, 2019. 5, 6

    2019

  25. [25]

    Kermany, Michael H

    Daniel S. Kermany, Michael H. Goldbaum, andet al. Iden- tifying medical diagnoses and treatable diseases by image- based deep learning.Cell, 172:1122–1131.e9, 2018. 5, 6

    2018

  26. [26]

    Self-regulating prompts: Foundational model adaptation without forgetting

    Muhammad Uzair Khattak, Syed Talal Wasim, Muzam- mal Naseer, Salman Khan, Ming-Hsuan Yang, and Fa- had Shahbaz Khan. Self-regulating prompts: Foundational model adaptation without forgetting. InIEEE/CVF Interna- tional Conference on Computer Vision, ICCV, pages 15144– 15154, 2023. 3

    2023

  27. [27]

    Cutpaste: Self-supervised learning for anomaly detec- tion and localization.2021 IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR), pages 9659– 9669, 2021

    Chun-Liang Li, Kihyuk Sohn, Jinsung Yoon, and Tomas Pfis- ter. Cutpaste: Self-supervised learning for anomaly detec- tion and localization.2021 IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR), pages 9659– 9669, 2021. 3

    2021

  28. [28]

    Junnan Li, Dongxu Li, Caiming Xiong, and Steven C. H. Hoi. BLIP: bootstrapping language-image pre-training for unified vision-language understanding and generation. In International Conference on Machine Learning, ICML 2022, 17-23 July 2022, Baltimore, Maryland, USA, pages 12888– 12900, 2022. 2, 3

    2022

  29. [29]

    Junnan Li, Dongxu Li, Silvio Savarese, and Steven C. H. Hoi. BLIP-2: bootstrapping language-image pre-training with frozen image encoders and large language models. In International Conference on Machine Learning, ICML 2023, 23-29 July 2023, Honolulu, Hawaii, USA, pages 19730– 19742, 2023. 2

    2023

  30. [30]

    Promptad: Learn- ing prompts with only normal samples for few-shot anomaly detection

    Xiaofan Li, Zhizhong Zhang, Xin Tan, Chengwei Chen, Yanyun Qu, Yuan Xie, and Lizhuang Ma. Promptad: Learn- ing prompts with only normal samples for few-shot anomaly detection. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 16838–16848, 2024. 2

    2024

  31. [31]

    Grounding DINO: marrying DINO with grounded pre-training for open-set object detection.CoRR,

    Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, and Lei Zhang. Grounding DINO: marrying DINO with grounded pre-training for open-set object detection.CoRR,

  32. [32]

    Aa- clip: Enhancing zero-shot anomaly detection via anomaly- aware clip.2025 IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition (CVPR), pages 4744–4754,

    Wenxin Ma, Xu Zhang, Qingsong Yao, Fenghe Tang, Chenxu Wu, Yingtai Li, Rui Yan, Zihang Jiang, and S.Kevin Zhou. Aa- clip: Enhancing zero-shot anomaly detection via anomaly- aware clip.2025 IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition (CVPR), pages 4744–4754,

    2025

  33. [33]

    MOCCA: multilayer one-class classification for anomaly de- tection.IEEE Trans

    Fabio Valerio Massoli, Fabrizio Falchi, Alperen Kantarci, Seymanur Akti, Hazim Kemal Ekenel, and Giuseppe Amato. MOCCA: multilayer one-class classification for anomaly de- tection.IEEE Trans. Neural Networks Learn. Syst., 33(6): 2313–2323, 2022. 2

    2022

  34. [34]

    The multimodal brain tumor image segmentation benchmark (brats).IEEE Transactions on Medical Imaging, 34:1993– 2024, 2015

    Bjoern H Menze, Andr ´as Jakab, Stefan Bauer, and et al. The multimodal brain tumor image segmentation benchmark (brats).IEEE Transactions on Medical Imaging, 34:1993– 2024, 2015. 5

    1993

  35. [35]

    VT-ADL: A vision trans- former network for image anomaly detection and localiza- tion

    Pankaj Mishra, Riccardo Verk, Daniele Fornasier, Claudio Piciarelli, and Gian Luca Foresti. VT-ADL: A vision trans- former network for image anomaly detection and localiza- tion. In30th IEEE International Symposium on Industrial Electronics, ISIE 2021, Kyoto, Japan, June 20-23, 2021, pages 1–6, 2021. 5, 6

    2021

  36. [36]

    RGI: robust gan- inversion for mask-free image inpainting and unsupervised pixel-wise anomaly detection

    Shancong Mou, Xiaoyi Gu, Meng Cao, Haoping Bai, Ping Huang, Jiulong Shan, and Jianjun Shi. RGI: robust gan- inversion for mask-free image inpainting and unsupervised pixel-wise anomaly detection. InThe Eleventh International Conference on Learning Representations, ICLR 2023, Ki- gali, Rwanda, May 1-5, 2023, 2023. 1

    2023

  37. [37]

    Rabbat, Vasu Sharma, Gabriel Synnaeve, Hu Xu, Herv´e J´egou, Julien Mairal, Patrick Labatut, Armand Joulin, and Piotr Bojanowski

    Maxime Oquab, Timoth ´ee Darcet, Th ´eo Moutakanni, Huy Vo, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, Mah- moud Assran, Nicolas Ballas, Wojciech Galuba, Rus- sell Howes, Po-Yao Huang, Shang-Wen Li, Ishan Misra, Michael G. Rabbat, Vasu Sharma, Gabriel Synnaeve, Hu Xu, Herv´e J´egou, Julien Mairal...

    2023

  38. [38]

    Pramuditha Perera, Ramesh Nallapati, and Bing Xiang. Ocgan: One-class novelty detection using gans with con- strained latent representations.2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 2893–2901, 2019. 3

    2019

  39. [39]

    Vcp-clip: A visual context prompting model for zero-shot anomaly segmenta- tion

    Zhen Qu, Xian Tao, Mukesh Prasad, Fei Shen, Zhengtao Zhang, Xinyi Gong, and Guiguang Ding. Vcp-clip: A visual context prompting model for zero-shot anomaly segmenta- tion. InEuropean Conference on Computer Vision, 2024. 3

    2024

  40. [40]

    Bayesian prompt flow learning for zero-shot anomaly detec- tion.2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 30398–30408, 2025

    Zhen Qu, Xian Tao, Xinyi Gong, Shichen Qu, Qiyu Chen, Zhengtao Zhang, Xingang Wang, and Guiguang Ding. Bayesian prompt flow learning for zero-shot anomaly detec- tion.2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 30398–30408, 2025. 3

    2025

  41. [41]

    Learning transferable visual models from natural language supervision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, and Ilya Sutskever. Learning transferable visual models from natural language supervision. InProceedings of the 38th International Conference on Machine Learning, ICML, pages 8748–8763, 2021. 2, 3, 6

    2021

  42. [42]

    Towards to- tal recall in industrial anomaly detection.2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 14298–14308, 2021

    Karsten Roth, Latha Pemula, Joaquin Zepeda, Bernhard Scholkopf, Thomas Brox, and Peter Gehler. Towards to- tal recall in industrial anomaly detection.2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 14298–14308, 2021. 3

    2022

  43. [43]

    Karsten Roth, Latha Pemula, Joaquin Zepeda, Bernhard Sch¨olkopf, Thomas Brox, and Peter V . Gehler. Towards total recall in industrial anomaly detection. InIEEE/CVF Con- ference on Computer Vision and Pattern Recognition, CVPR 2022, New Orleans, LA, USA, June 18-24, 2022, 2022. 1

    2022

  44. [44]

    Fully convolutional cross-scale-flows for image- based defect detection.2022 IEEE/CVF Winter Conference on Applications of Computer Vision (W ACV), pages 1829– 1838, 2021

    Marco Rudolph, Tom Wehrbein, Bodo Rosenhahn, and Bas- tian Wandt. Fully convolutional cross-scale-flows for image- based defect detection.2022 IEEE/CVF Winter Conference on Applications of Computer Vision (W ACV), pages 1829– 1838, 2021. 3

    2022

  45. [45]

    Mohammadreza Salehi, Niousha Sadjadi, Soroosh Baselizadeh, Mohammad Hossein Rohban, and Hamid R. Rabiee. Multiresolution knowledge distillation for anomaly detection.2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 14897–14907,

    2021

  46. [46]

    Learning and evaluating representations for deep one-class classification

    Kihyuk Sohn, Chun-Liang Li, Jinsung Yoon, Minho Jin, and Tomas Pfister. Learning and evaluating representations for deep one-class classification. InInternational Conference on Learning Representations, ICLR, 2021. 2

    2021

  47. [47]

    Segmentation-based deep-learning approach for surface-defect detection.J

    Domen Tabernik, Samo Sela, Jure Skvarc, and Danijel Skocaj. Segmentation-based deep-learning approach for surface-defect detection.J. Intell. Manuf., 31, 2020. 5, 6

    2020

  48. [48]

    Multimodal industrial anomaly detection via hybrid fusion

    Yue Wang, Jinlong Peng, Jiangning Zhang, Ran Yi, Yabiao Wang, and Chengjie Wang. Multimodal industrial anomaly detection via hybrid fusion. InIEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2023, Van- couver, BC, Canada, June 17-24, 2023, 2023. 1

    2023

  49. [49]

    Weakly supervised learn- ing for industrial optical inspection

    Matthias Wieler and Tobias Hahn. Weakly supervised learn- ing for industrial optical inspection. InDAGM symposium in, page 11, 2007. 5, 6

    2007

  50. [50]

    Pushing the limits of fewshot anomaly detection in industry vision: Graphcore

    Guoyang Xie, Jinbao Wang, Jiaqi Liu, Yaochu Jin, and Feng Zheng. Pushing the limits of fewshot anomaly detection in industry vision: Graphcore. InThe Eleventh International Conference on Learning Representations, ICLR 2023, Ki- gali, Rwanda, May 1-5, 2023, 2023. 1

    2023

  51. [51]

    Learning semantic context from nor- mal samples for unsupervised anomaly detection

    Xudong Yan, Huaidong Zhang, Xuemiao Xu, Xiaowei Hu, and Pheng-Ann Heng. Learning semantic context from nor- mal samples for unsupervised anomaly detection. InAAAI Conference on Artificial Intelligence, 2021. 3

    2021

  52. [52]

    Prior normality prompt transformer for multiclass industrial image anomaly de- tection.IEEE Transactions on Industrial Informatics, 20: 11866–11876, 2024

    Haiming Yao, Yunkang Cao, Wei Luo, Weihang Zhang, Wenyong Yu, and Weiming Shen. Prior normality prompt transformer for multiclass industrial image anomaly de- tection.IEEE Transactions on Industrial Informatics, 20: 11866–11876, 2024. 3

    2024

  53. [53]

    Focus the discrepancy: Intra- and inter- correlation learning for image anomaly detection.2023 IEEE/CVF International Conference on Computer Vision (ICCV), pages 6780–6790, 2023

    Xincheng Yao, Ruoqing Li, Zefeng Qian, Yan Luo, and Chongyang Zhang. Focus the discrepancy: Intra- and inter- correlation learning for image anomaly detection.2023 IEEE/CVF International Conference on Computer Vision (ICCV), pages 6780–6790, 2023. 3

    2023

  54. [54]

    Patch SVDD: patch-level SVDD for anomaly detection and segmentation

    Jihun Yi and Sungroh Yoon. Patch SVDD: patch-level SVDD for anomaly detection and segmentation. In15th Asian Con- ference on Computer Vision ACCV, pages 375–390, 2020. 2

    2020

  55. [55]

    Zaheer, Jin ha Lee, M

    M. Zaheer, Jin ha Lee, M. Astrid, and Seung-Ik Lee. Old is gold: Redefining the adversarially learned one-class clas- sifier training paradigm.2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 14171–14181, 2020. 3

    2020

  56. [56]

    Vitjan Zavrtanik, Matej Kristan, and Danijel Skovcaj. Draem – a discriminatively trained reconstruction embed- ding for surface anomaly detection.2021 IEEE/CVF In- ternational Conference on Computer Vision (ICCV), pages 8310–8319, 2021. 3

    2021

  57. [57]

    Ni, and Heung-Yeung Shum

    Hao Zhang, Feng Li, Shilong Liu, Lei Zhang, Hang Su, Jun Zhu, Lionel M. Ni, and Heung-Yeung Shum. DINO: DETR with improved denoising anchor boxes for end-to-end ob- ject detection. InThe Eleventh International Conference on Learning Representations, ICLR, 2023. 2, 3

    2023

  58. [58]

    Anomalyclip: Object-agnostic prompt learning for zero-shot anomaly detection

    Qihang Zhou, Guansong Pang, Yu Tian, Shibo He, and Jim- ing Chen. Anomalyclip: Object-agnostic prompt learning for zero-shot anomaly detection. InThe Twelfth International Conference on Learning Representations (ICLR), pages 1– 33, 2024. 2, 3, 5, 6

    2024

  59. [59]

    Jiawen Zhu and Guansong Pang. Toward generalist anomaly detection via in-context residual learning with few- shot sample prompts.2024 IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR), pages 17826– 17836, 2024. 2

    2024

  60. [60]

    Spot-the-difference self-supervised pre- training for anomaly detection and segmentation

    Yang Zou, Jongheon Jeong, Latha Pemula, Dongqing Zhang, and Onkar Dabeer. Spot-the-difference self-supervised pre- training for anomaly detection and segmentation. InCom- puter Vision - ECCV 2022 - 17th European Conference, Tel Aviv, pages 392–408, 2022. 5, 6

    2022