Pith. sign in

REVIEW 4 major objections 6 minor 44 references

Breaking the Bias: Recalibrating the Attention of Industrial Anomaly Detection

T0 review · 4 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read This paper argues that quantizing an unsupervised anomaly detector and then fine-tuning it on normal images removes the attention bias that makes such models miss subtle defects, and reports state-of-the-art scores on three industrial…

desk verdict The attention-recalibration idea is worth discussing, but the SOTA claim rests on a baseline that is either misidentified or misrepresented. read the letter →

arxiv 2412.08189 v1 pith:JWWSCNBW submitted 2024-12-11 cs.CV

Xin Chen , Liujuan Cao , Shengchuan Zhang , Xiewu Zheng , Yan Zhang This is my paper · ORCID
classification cs.CV
keywords unsupervisedanomalydetectionindustrialinspectionmodelquantizationmixedprecisionattentionrecalibrationteacher-studentnetworklocalization
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 proposes a two-stage recipe for unsupervised industrial anomaly detection: first quantize a trained teacher–student model to lower numerical precision, then fine-tune it on the same normal-only images. The claim is that the precision drop suppresses an attention bias that makes these models focus on variable regions of normal samples while overlooking subtle defects in invariant areas, and that fine-tuning then redirects attention to defect-prone regions. The method, RAAD, uses a module called Hierarchical Quantization Scoring (HQS) to decide how many bits each layer should keep, based on how strongly its teacher and student outputs disagree. The authors report average detection AU-ROC of 98.90 on MVTec AD, 89.75 on MVTec LOCO, and 96.72 on VisA, calling this the best overall performance among the compared methods. If correct, the practical payoff is that industrial inspection can gain accuracy and parameter efficiency without new defect data or labels.

What carries the argument

The central mechanism is Hierarchical Quantization Scoring (HQS), a module placed between aligned teacher and student convolutional layers. For each layer it computes a scalar anomaly score from the layer-size-normalized mean squared teacher–student feature difference, and a piecewise function converts that score into one of four bit-widths: 2, 3, 4, or 8 bits. HQS determines which layers are compressed during post-training quantization and which keep higher precision, so that quantization suppresses diffuse, noisy attention in shallow layers while preserving sharp, defect-focused attention in deeper layers. Around this, the architecture is a frozen teacher and trainable student built from a four-layer fully convolutional patch description network, plus an autoencoder with a 64-dimensional bottleneck that supplies global and logical context. The work HQS does is to make the quantization step itself a principled attention-decomposition pass before the fine-tuning step recalibrates the remaining attention.

What would settle it

Run the official implementation of the baseline method from [1] on the MVTec AD benchmark under the same protocol; if its detection AU-ROC is at or above 96.98, the reproduced baseline in this paper, then the reported gains of RAAD over the published method shrink or disappear.

Watch

Extended reading notes

Core claim

RAAD's central discovery is that numerical perturbation of a trained anomaly detector can act as an attention corrector rather than a source of error. When the teacher–student feature extractor is quantized layer by layer, reducing parameter precision forces the student to keep only the most stable correspondences with the teacher, which spreads attention toward the object and away from background variation. Hierarchical Quantization Scoring makes the correction selective: it aligns each teacher and student layer, computes the normalized mean squared difference between their feature maps as an anomaly score, and maps that score through a piecewise function to a bit-width chosen from 2, 3, 4, or 8 bits, so layers with coarse noisy attention are compressed and layers with sharp defect-focused attention keep more precision. The quantized model is then fine-tuned with three distillation losses among the teacher, student, and a global autoencoder, and the local and global anomaly maps are averaged at inference. On 32 datasets the authors report detection AU-ROC of 98.90 on MVTec AD, 89.75 on MVTec LOCO, and 96.72 on VisA, and conclude that RAAD achieves the best overall anomaly detection performance among the methods compared.

Load-bearing premise

The load-bearing premise is that the authors' unofficial re-implementation of the baseline comparison method from [1] performs as well as the officially published version of that method; every reported improvement and the state-of-the-art claim is measured against that re-implementation.

Editorial extensions

If this is right

  • If RAAD works as reported, industrial inspection can improve detection and localization accuracy without collecting defect samples or changing deployment hardware, since quantization and normal-image fine-tuning are the only added steps.
  • The reported MVTec LOCO gain, from 84.09 to 89.75 mean detection AU-ROC, indicates the two-stage procedure helps logical and global anomalies, a category where pure patch-based teacher–student models are known to be weak.
  • HQS turns per-layer bit-width into a controllable attention dial: layers with diffuse attention can be heavily compressed without hurting detection, which lowers memory and latency.
  • Because the training and fine-tuning phases use only normal images, RAAD can be applied to existing unsupervised pipelines without altering data collection or annotation requirements.
  • The experiments span 32 datasets on a single 3090Ti, suggesting the added computational cost is within reach of a modest industrial setup.

Reading between the lines

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

  • Editorial inference: the underlying idea, that quantization regularizes where a representation learner looks, may transfer to other unsupervised methods such as memory-bank or reconstruction-based anomaly detectors; a direct test would apply the same quantize-then-finetune schedule to those families.
  • Editorial inference: HQS is essentially a cheap per-layer saliency measure, since the teacher–student disagreement it computes is already a localization signal; comparing it with gradient-based attribution maps could show whether the bit allocation tracks interpretable importance.
  • Editorial inference: the paper only demonstrates the effect on a convolutional patch description backbone, so whether the bias-breaking mechanism carries over to transformer-based anomaly detectors is an open experimental question.
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

4 major / 6 minor

Summary. The paper proposes RAAD, an unsupervised industrial anomaly detection method built on the EfficientAD teacher-student plus autoencoder architecture. RAAD operates in three stages: initial training on normal images, layer-wise mixed-precision post-training quantization guided by a proposed Hierarchical Quantization Scoring (HQS) module, and fine-tuning of the student and autoencoder to 'recalibrate' attention maps. The paper claims that quantization reduces the intrinsic bias of attention toward variable regions in normal samples, and that HQS improves this effect by allocating higher bit-widths to layers that contribute most to anomaly detection. Experiments on MVTec AD, MVTec LOCO, and VisA report detection AU-ROC scores of 98.90, 89.75, and 96.72, respectively, and the paper claims the best overall detection performance compared with PatchCore, GCAD, SimpleNet, and EfficientAD-S. The central claims are therefore that quantized mixed-precision models can improve both accuracy and parameter efficiency in unsupervised IAD.

Significance. If substantiated, the contribution would be practically valuable: it would show that lower precision can simultaneously improve accuracy and efficiency in a lightweight anomaly detector, an interesting and nontrivial result. The paper is well-motivated, evaluates on three standard benchmarks with per-class breakdowns, compares against multiple quantization baselines, and reports an ablation isolating the HQS module. The single-GPU evaluation on standard public datasets is also a strength for reproducibility of compute requirements. However, the validity of the headline claims hinges on the fidelity of an unofficial EfficientAD baseline and on the statistical reliability of small metric differences, both of which are currently unverified. The proposal is testable and the experimental framework is appropriate, so the result is potentially significant but not yet convincingly demonstrated.

major comments (4)
  1. [Section 4.2, 4.3] The paper states that EfficientAD has not publicly released its code and therefore uses an in-house reproduction as the baseline. This statement is factually incorrect: the official EfficientAD implementation has been publicly available since the release of the WACV 2024 paper. Because all headline comparisons in Tables 1, 2, and 3 and the 'best overall' claim in Section 4.3 are measured relative to this reproduction, the paper must provide evidence that the reproduction faithfully matches the official training protocol (teacher pre-training, 10% top-k masking, PDN architecture, loss weights) and must report official EfficientAD numbers as a sanity check. Without this, the reported gains of +1.92 AU-ROC on MVTec AD and +5.66 AU-ROC on MVTec LOCO cannot be attributed to RAAD rather than to a possibly weaker baseline.
  2. [Table 3, Section 4.4] The ablation in Table 3 shows that adding HQS over uniform 8-bit quantization raises detection AU-ROC from 98.77 to 98.90, an improvement of 0.13 percentage points, and segmentation AU-PRO from 92.17 to 92.92. Since no error bars, confidence intervals, or multiple-seed runs are reported for Tables 1–3, this improvement is within plausible run-to-run variation for stochastic training. The accompanying text claims that 'HQS significantly optimizes attention distribution'; this claim is not supported by the reported evidence.
  3. [Section 3.3, Eq. (4)] The bit-width mapping b(ℓ)=φ(anomaly score) is defined only as a 'piecewise function' with no stated thresholds or rule, and the allowed bit widths are fixed to {2,3,4,8}. The reviewer cannot determine whether φ is a simple step function, a linear mapping, or a tuned look-up table, nor how many additional free parameters it introduces. Because φ is selected after seeing benchmark results, the reported RAAD numbers incorporate an unspecified tuning dimension. In addition, the anomaly score used to allocate bits is the same teacher-student discrepancy that later produces the anomaly maps (Figure 4), so the allocation criterion is not an independent measure of layer-level 'anomaly detection contribution.'
  4. [Section 4.2] The implementation details state that the model is trained for 70k iterations and fine-tuned for up to 60k, but 'our experiments show that the model often achieves the best performance before reaching the full 60,000 iterations.' No early-stopping rule is specified. Selecting the best checkpoint across iterations is legitimate only if the selection criterion is predefined and does not use test labels; the paper does not say what that criterion is. This is especially important for the comparisons in Tables 1–3 and Figure 6, since unchecked checkpoint selection can be a source of inflated results.
minor comments (6)
  1. [Section 4.2] The text says 'We compared our quantization method with HQS [14] and OMPQ [25]', but reference [14] is Esser et al., 'Learned step size quantization' (LSQ), and Table 1 labels this column 'LSQ'; the text should say LSQ instead of HQS.
  2. [Section 4.3 vs. Table 1] Section 4.3 reports average detection AU-ROC scores of 98.8, 89.75, and 96.13, while Table 1 reports 98.90, 89.75, and 96.72; these numbers must be reconciled.
  3. [Abstract, Section 4.3] The paper claims validation on '32 datasets'; these are 32 object classes within three benchmark datasets (MVTec AD, MVTec LOCO, VisA), not 32 independent datasets. The wording should be corrected for accuracy.
  4. [Figure 6] Figure 6 contains garbled labels (including '图表标题', 'PQAD', and 'ft w/o increase') and the stacked bars and line chart are hard to parse. The figure should be redrawn with standard English labels and a clear description of what each bar segment represents.
  5. [Figure 7] The third column of Figure 7 is labeled 'After Quant(4bit)', while the RAAD pipeline in the main text uses mixed precision with widths up to 8 bits. The figure caption should clarify whether this is a separate 4-bit quantization experiment or part of the HQS allocation.
  6. [Section 3.4] The description of the loss masking is ambiguous: 'the value of 10% is used for backpropagation in each of the three dimensions of the mean squared error D.' Please specify whether the 10% largest errors are selected per spatial location, per channel, or globally, and how this interacts with the three losses in Eq. (6).

Circularity Check

1 steps flagged · score 3.0 of 10

HQS's bit-allocation metric is self-referential by construction (Eq. 4 uses the same teacher-student discrepancy that defines the anomaly map), though the final AU-ROC claim remains an independent empirical result; the EfficientAD-baseline concern is a reproducibility risk, not circularity.

  1. self definitional [Section 3.3, Eq. (4); Section 3.4, anomaly-map definition]
    "The bit width b for the ℓ-th layer (1 ≤ ℓ ≤ N) is defined as: b(ℓ) = ϕ(anomaly score) = ϕ((c(ℓ)w(ℓ)h(ℓ))−1 Σ_c ||T_c^(ℓ)(i) − S_c^(ℓ)(i)||_F^2) ... the 2D anomaly score map M ∈ R^(W×H) is given by Mw,h = C−1 Σ_c Dc,w,h ... where Dc,w,h = (T(I)c,w,h − S(I)c,w,h)^2."

    HQS computes each layer's 'anomaly detection contribution' as the per-layer teacher-student squared difference, which is exactly the same signal used at inference to produce the final anomaly map: the bit-width in Eq. (4) is a function of the layer-wise version of D, and the anomaly score map M is the cross-channel average of D. Thus the importance metric that sets precision is not an independent measure of a layer's causal contribution; it is the detection score itself. The statement that high-scoring layers 'contribute more to accurate anomaly localization' is therefore definitional: a layer contributes by construction because its output constitutes the anomaly signal.

full rationale

RAAD is presented as an empirical framework rather than a first-principles derivation, and most of its pipeline is non-circular: quantization, fine-tuning, and the teacher-student/autoencoder losses are standard components evaluated on held-out test sets. The only place where a quantity is both an input and an output is HQS: Eq. (4) defines the per-layer bit allocation as a function of the teacher-student squared difference, and Section 3.4 defines the anomaly map as the same squared difference averaged across channels. So 'anomaly detection contribution' is the anomaly score itself, not an independent causal measure of layer importance. This is a moderate self-definitional issue, but it does not force the reported AU-ROC gains, because whether this allocation helps is settled by held-out data and could have gone the other way. I did not count the EfficientAD baseline problem as circularity: the claim that the authors reproduced the method because no code was released is factually questionable and undermines the SOTA comparison, but it is a reproducibility and correctness risk, not a reduction of a result to its inputs by construction. Similarly, the undisclosed piecewise function φ and the absence of error bars weaken statistical evidence but are not circular. Overall score 3 reflects one self-referential importance metric with an otherwise independent empirical central claim.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The method rests on standard quantization machinery (BRECQ) and teacher-student IAD. The main hidden choices are the undisclosed HQS mapping, the early-stopping rule, and the unofficial baseline. No new entities are introduced.

free parameters (2)
  • HQS piecewise mapping phi = unspecified; bit-widths in {2,3,4,8}
    Eq. 4 defines bit-width as a piecewise function of anomaly score, but the thresholds are not given, and the bit-width set is manually chosen; this mapping could be tuned on the benchmark datasets.
  • Early-stopping iteration = best before 60k iterations
    Section 4.2 states the model often achieves best performance before 60k iterations without a validation protocol, enabling selection based on test performance.
assumptions (3)
  • domain assumption Teacher-student feature discrepancy is a valid proxy for layer-wise anomaly detection contribution.
    Eq. 4 uses this discrepancy to set bit-widths; the paper provides no evidence that larger discrepancy indicates better detection.
  • domain assumption Quantization redistributes attention away from variable normal regions toward defect-prone regions.
    Central mechanism asserted in Section 1 and Figure 2a without controlled experiments isolating the effect.
  • domain assumption The unofficial reproduction of EfficientAD faithfully represents the published method.
    All baseline comparisons depend on this reproduction (Section 4.2); no official numbers are provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Breaking the Bias: Recalibrating the Attention of Industrial Anomaly Detection." pith.science (2026). https://pith.science/paper/JWWSCNBW

@misc{pith2026241208189,
  author       = {Pith},
  title        = {Pith review of: Breaking the Bias: Recalibrating the Attention of Industrial Anomaly Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JWWSCNBW}},
  note         = {Machine review of arXiv:2412.08189}
}
read the original abstract

Due to the scarcity and unpredictable nature of defect samples, industrial anomaly detection (IAD) predominantly employs unsupervised learning. However, all unsupervised IAD methods face a common challenge: the inherent bias in normal samples, which causes models to focus on variable regions while overlooking potential defects in invariant areas. To effectively overcome this, it is essential to decompose and recalibrate attention, guiding the model to suppress irrelevant variations and concentrate on subtle, defect-susceptible areas. In this paper, we propose Recalibrating Attention of Industrial Anomaly Detection (RAAD), a framework that systematically decomposes and recalibrates attention maps. RAAD employs a two-stage process: first, it reduces attention bias through quantization, and second, it fine-tunes defect-prone regions for improved sensitivity. Central to this framework is Hierarchical Quantization Scoring (HQS), which dynamically allocates bit-widths across layers based on their anomaly detection contributions. HQS dynamically adjusts bit-widths based on the hierarchical nature of attention maps, compressing lower layers that produce coarse and noisy attention while preserving deeper layers with sharper, defect-focused attention. This approach optimizes both computational efficiency and the model' s sensitivity to anomalies. We validate the effectiveness of RAAD on 32 datasets using a single 3090ti. Experiments demonstrate that RAAD, balances the complexity and expressive power of the model, enhancing its anomaly detection capability.

Figures

Figures reproduced from arXiv: 2412.08189 by the authors.

Figure 1
Figure 1. Visualization of heatmaps. These samples are from the [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. (a) visualization of the attention maps at different stages [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Pipeline of RAAD. Our architecture consists of three [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (3 more)
Figure 5
Figure 5. Figure 5: The inference process of the models, the input is from the [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]
Figure 6
Figure 6. Figure 6: Quantizing and fine-tuning models using varying bit [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: Qualitative results of RAAD on the MVTec-AD dataset [ [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 40 canonical work pages

  1. [1]

    Efficien- tad: Accurate visual anomaly detection at millisecond-level latencies

    Kilian Batzner, Lars Heckler, and Rebecca K ¨onig. Efficien- tad: Accurate visual anomaly detection at millisecond-level latencies. In Proceedings of the IEEE/CVF Winter Confer- ence on Applications of Computer Vision , pages 128–138,

  2. [2]

    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. In 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2019. 1, 5, 8

    work page 2019

  3. [3]

    Improving unsupervised defect seg- mentation by applying structural similarity to autoencoders

    Paul Bergmann, Sindy L ¨owe, Michael Fauser, David Sattleg- ger, and Carsten Steger. Improving unsupervised defect seg- mentation by applying structural similarity to autoencoders. In Proceedings of the 14th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, 2019. 3

    work page 2019

  4. [4]

    Uninformed students: Student-teacher anomaly detection with discriminative latent embeddings

    Paul Bergmann, Michael Fauser, David Sattlegger, and Carsten Steger. Uninformed students: Student-teacher anomaly detection with discriminative latent embeddings. In 2020 IEEE/CVF Conference on Computer Vision and Pat- tern Recognition (CVPR), 2020. 2, 3

    work page 2020

  5. [5]

    The mvtec anomaly detection dataset: A comprehensive real-world dataset for unsuper- vised anomaly detection

    Paul Bergmann, Kilian Batzner, Michael Fauser, David Sat- tlegger, and Carsten Steger. The mvtec anomaly detection dataset: A comprehensive real-world dataset for unsuper- vised anomaly detection. International Journal of Computer Vision, page 1038–1059, 2021. 4, 5

    work page 2021

  6. [6]

    Beyond dents and scratches: Logical constraints in unsupervised anomaly detection and localization

    Paul Bergmann, Kilian Batzner, Michael Fauser, David Sat- tlegger, and Carsten Steger. Beyond dents and scratches: Logical constraints in unsupervised anomaly detection and localization. International Journal of Computer Vision, 130 (4):947–969, 2022. 1, 5, 7, 8

    work page 2022

  7. [7]

    Mahoney, and Kurt Keutzer

    Yaohui Cai, Zhewei Yao, Zhen Dong, Amir Gholami, MichaelW. Mahoney, and Kurt Keutzer. Zeroq: A novel zero shot quantization framework. Cornell University - arXiv,Cornell University - arXiv, 2020. 3

    work page 2020

  8. [8]

    Easynet: An easy net- work for 3d industrial anomaly detection

    Ruitao Chen, Guoyang Xie, Jiaqi Liu, Jinbao Wang, Ziqi Luo, Jinfan Wang, and Feng Zheng. Easynet: An easy net- work for 3d industrial anomaly detection. In Proceedings of the 31st ACM International Conference on Multimedia , pages 7038–7046, 2023. 3

    work page 2023

Show all 44 references
  1. [9]

    Cnn-based autoencoder and post-training quantization for on-device anomaly detection of cartesian coordinate robots

    Hyuntae Cho. Cnn-based autoencoder and post-training quantization for on-device anomaly detection of cartesian coordinate robots. In 2024 IEEE 14th Annual Comput- ing and Communication Workshop and Conference (CCWC), pages 0662–0666. IEEE, 2024. 3

    2024

  2. [10]

    Sub-image anomaly detection with deep pyramid correspondences

    Niv Cohen and Yedid Hoshen. Sub-image anomaly detection with deep pyramid correspondences. Cornell University - arXiv,Cornell University - arXiv, 2020. 2

    2020

  3. [11]

    Anomaly detection via reverse distillation from one-class embedding

    Hanqiu Deng and Xingyu Li. Anomaly detection via reverse distillation from one-class embedding. In 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022. 2, 3

    2022

  4. [12]

    Hawq-v2: Hessian aware trace-weighted quantization of neural networks

    Zhen Dong, Zhewei Yao, Daiyaan Arfeen, Amir Gholami, Michael W Mahoney, and Kurt Keutzer. Hawq-v2: Hessian aware trace-weighted quantization of neural networks. Ad- vances in neural information processing systems, 33:18518– 18529, 2020. 3

    2020

  5. [13]

    Few- shot defect image generation via defect-aware feature manip- ulation

    Yuxuan Duan, Yan Hong, Li Niu, and Liqing Zhang. Few- shot defect image generation via defect-aware feature manip- ulation. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 571–578, 2023. 3

    2023

  6. [14]

    Learned step size quantization

    Steven K Esser, Jeffrey L McKinstry, Deepika Bablani, Rathinakumar Appuswamy, and Dharmendra S Modha. Learned step size quantization. arXiv preprint arXiv:1902.08153, 2019. 4, 7

    1902 arXiv

  7. [15]

    Differ- entiable soft quantization: Bridging full-precision and low- bit neural networks

    Ruihao Gong, Xianglong Liu, Shenghu Jiang, Tianxiang Li, Peng Hu, Jiazhen Lin, Fengwei Yu, and Junjie Yan. Differ- entiable soft quantization: Bridging full-precision and low- bit neural networks. In Proceedings of the IEEE/CVF inter- national conference on computer vision , p...

  8. [16]

    Gruber, and Paul Tabatabai

    Matthias Haselmann, Dieter P. Gruber, and Paul Tabatabai. Anomaly detection using deep learning based image com- pletion. In 2018 17th IEEE International Conference on Ma- chine Learning and Applications (ICMLA), 2018. 3

    2018

  9. [17]

    Densely connected convolutional net- works

    Gao Huang, Zhuang Liu, Laurens Van Der Maaten, and Kil- ian Q Weinberger. Densely connected convolutional net- works. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 4700–4708, 2017. 4

    2017

  10. [18]

    Unified anomaly detection methods on edge device using knowledge distillation and quantization.arXiv preprint arXiv:2407.02968, 2024

    Sushovan Jena, Arya Pulkit, Kajal Singh, Anoushka Baner- jee, Sharad Joshi, Ananth Ganesh, Dinesh Singh, and Arnav Bhavsar. Unified anomaly detection methods on edge device using knowledge distillation and quantization.arXiv preprint arXiv:2407.02968, 2024. 3

    2024 arXiv

  11. [19]

    A survey of deep learning- based network anomaly detection

    Donghwoon Kwon, Hyunjoo Kim, Jinoh Kim, Sang C Suh, Ikkyun Kim, and Kuinam J Kim. A survey of deep learning- based network anomaly detection. Cluster Computing, 22: 949–961, 2019. 2

    2019

  12. [21]

    Brecq: Pushing the limit of post-training quantization by block reconstruction

    Yuhang Li, Ruihao Gong, Xu Tan, Yang Yang, Peng Hu, Qi Zhang, Fei Yu, Wei Wang, and Shi Gu. Brecq: Pushing the limit of post-training quantization by block reconstruction. Cornell University - arXiv,Cornell University - arXiv, 2021. 4

    2021

  13. [22]

    Radke, and Octavia Camps

    Wenqian Liu, Runze Li, Meng Zheng, Srikrishna Karanam, Ziyan Wu, Bir Bhanu, Richard J. Radke, and Octavia Camps. Towards visually explaining variational autoencoders. In 2020 IEEE/CVF Conference on Computer Vision and Pat- tern Recognition (CVPR), 2020. 3

    2020

  14. [23]

    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. In Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition , pages 20402–20411, 2023. 7

    2023

  15. [24]

    Remov- ing anomalies as noises for industrial defect localization

    Fanbin Lu, Xufeng Yao, Chi-Wing Fu, and Jiaya Jia. Remov- ing anomalies as noises for industrial defect localization. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 16166–16175, 2023. 3

    2023

  16. [25]

    Ompq: Orthogonal mixed precision quantization

    Yuexiao Ma, Taisong Jin, Xiawu Zheng, Yan Wang, Huixia Li, Yongjian Wu, Guannan Jiang, Wei Zhang, and Rongrong Ji. Ompq: Orthogonal mixed precision quantization. In Pro- ceedings of the AAAI conference on artificial intelligence , pages 9029–9037, 2023. 7

    2023

  17. [26]

    Data-free quantization through weight equal- ization and bias correction

    Markus Nagel, Mart Van Baalen, Tijmen Blankevoort, and Max Welling. Data-free quantization through weight equal- ization and bias correction. In 2019 IEEE/CVF International Conference on Computer Vision (ICCV), 2019. 3

    2019

  18. [27]

    Variational inference with normalizing flows

    DaniloJimenez Rezende and Shakir Mohamed. Variational inference with normalizing flows. International Conference on Machine Learning,International Conference on Machine Learning, 2015. 2

    2015

  19. [28]

    Moes- lund, and Mubarak Shah

    Nicolae-Catalin Ristea, Neelu Madan, Radu Tudor Ionescu, Kamal Nasrollahi, Fahad Shahbaz Khan, Thomas B. Moes- lund, and Mubarak Shah. Self-supervised predictive con- volutional attentive block for anomaly detection. In 2022 IEEE/CVF Conference on Computer Vision and Pattern R...

    2022

  20. [29]

    Towards to- tal recall in industrial anomaly detection

    Karsten Roth, Latha Pemula, Joaquin Zepeda, Bernhard Sch¨olkopf, Thomas Brox, and Peter Gehler. Towards to- tal recall in industrial anomaly detection. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 14318–14328, 2022. 2, 3, 7

    2022

  21. [30]

    Same same but differnet: Semi-supervised defect detection with normalizing flows

    Marco Rudolph, Bastian Wandt, and Bodo Rosenhahn. Same same but differnet: Semi-supervised defect detection with normalizing flows. In 2021 IEEE Winter Conference on Ap- plications of Computer Vision (WACV), 2021. 2

    2021

  22. [31]

    Kauffmann, Robert A

    Lukas Ruff, Jacob R. Kauffmann, Robert A. Vander- meulen, Gregoire Montavon, Wojciech Samek, Marius Kloft, Thomas G. Dietterich, and Klaus-Robert Muller. A unifying review of deep and shallow anomaly detection.Pro- ceedings of the IEEE, page 756–795, 2021. 2

    2021

  23. [32]

    Quantized autoen- coder (qae) intrusion detection system for anomaly detection in resource-constrained iot devices using rt-iot2022 dataset

    BS Sharmila and Rohini Nagapadma. Quantized autoen- coder (qae) intrusion detection system for anomaly detection in resource-constrained iot devices using rt-iot2022 dataset. Cybersecurity, 6(1):41, 2023. 3

    2023

  24. [33]

    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. Learning,Learning, 2020. 2

    2020

  25. [34]

    Deep learning for anomaly detection: A survey

    RaghavendraChalapathy(Universityof Sydney, Capital- MarketsCooperativeResearchCentre (CMCRC)), Sanjay- Chawla(QatarComputingResearchInstitute (QCRI), and HBKU) HBKU). Deep learning for anomaly detection: A survey. arXiv preprint arXiv:2308.06748, 2019. 2

    2019 arXiv

  26. [35]

    Deep learning for unsupervised anomaly lo- calization in industrial images: A survey

    Xian Tao, Xinyi Gong, Xin Zhang, Shaohua Yan, and Chan- dranath Adak. Deep learning for unsupervised anomaly lo- calization in industrial images: A survey. IEEE Transactions on Instrumentation and Measurement, 71:1–21, 2022

    2022

  27. [36]

    Im-iad: Indus- trial image anomaly detection benchmark in manufacturing

    Guoyang Xie, Jinbao Wang, Jiaqi Liu, Jiayi Lyu, Yong Liu, Chengjie Wang, Feng Zheng, and Yaochu Jin. Im-iad: Indus- trial image anomaly detection benchmark in manufacturing. IEEE Transactions on Cybernetics, 54(5):2720–2733, 2024. 2

    2024

  28. [37]

    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. Proceed- ings of the AAAI Conference on Artificial Intelligence , page 3110–3118, 2022. 3

    2022

  29. [38]

    Focus the discrepancy: Intra-and inter- correlation learning for image anomaly detection

    Xincheng Yao, Ruoqi Li, Zefeng Qian, Yan Luo, and Chongyang Zhang. Focus the discrepancy: Intra-and inter- correlation learning for image anomaly detection. In Pro- ceedings of the IEEE/CVF International Conference on Computer Vision, pages 6803–6813, 2023. 3

    2023

  30. [39]

    A unified model for multi-class anomaly detection

    Zhiyuan You, Lei Cui, Yujun Shen, Kai Yang, Xin Lu, Yu Zheng, and Xinyi Le. A unified model for multi-class anomaly detection. Advances in Neural Information Pro- cessing Systems, 35:4571–4584, 2022. 3

    2022

  31. [40]

    Wide residual net- works

    Sergey Zagoruyko and Nikos Komodakis. Wide residual net- works. In Procedings of the British Machine Vision Confer- ence 2016, 2016. 4, 5

    2016

  32. [41]

    Dr- AEm – a discriminatively trained reconstruction embedding for surface anomaly detection

    Vitjan Zavrtanik, Matej Kristan, and Danijel Sko ˇcaj. Dr- AEm – a discriminatively trained reconstruction embedding for surface anomaly detection. arXiv: Computer Vision and Pattern Recognition,arXiv: Computer Vision and Pat- tern Recognition, 2021. 3

    2021

  33. [42]

    Recon- struction by inpainting for visual anomaly detection

    Vitjan Zavrtanik, Matej Kristan, and Danijel Sko ˇcaj. Recon- struction by inpainting for visual anomaly detection. Pattern Recognition, page 107706, 2021. 3

    2021

  34. [43]

    Con- textual affinity distillation for image anomaly detection

    Jie Zhang, Masanori Suganuma, and Takayuki Okatani. Con- textual affinity distillation for image anomaly detection. In Proceedings of the IEEE/CVF Winter Conference on Appli- cations of Computer Vision, pages 149–158, 2024. 3

    2024

  35. [44]

    Unsupervised surface anomaly detection with diffusion probabilistic model

    Xinyi Zhang, Naiqi Li, Jiawei Li, Tao Dai, Yong Jiang, and Shu-Tao Xia. Unsupervised surface anomaly detection with diffusion probabilistic model. In Proceedings of the IEEE/CVF International Conference on Computer Vision , pages 6782–6791, 2023. 3

    2023

  36. [45]

    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. In Eu- ropean Conference on Computer Vision , pages 392–408. Springer, 2022. 5

    2022

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.