Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Towards Accurate Unified Anomaly Segmentation

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read UniAS claims that multi-level hybrid reconstruction with gated multi-granularity CNNs segments industrial anomalies at pixel level, reaching pAP/DSC of 65.12/59.33 on MVTec-AD and 40.06/32.50 on VisA, surpassing previous unified models.

desk verdict Solid one-for-all anomaly segmentation with a credible pAP win; the DSC headline is oracle-thresholded and overstated, but the architecture and analysis deserve referee time. read the letter →

arxiv 2501.12295 v1 pith:WEUC3KKI submitted 2025-01-21 cs.CV

Wenxin Ma , Qingsong Yao , Xiang Zhang , Zhelong Huang , Zihang Jiang , S. Kevin Zhou This is my paper · ORCID
classification cs.CV
keywords anomalysegmentationunsuperviseddetectionone-for-allmulti-levelreconstructionMGG-CNNpixel-wiseaverageprecisionDicesimilaritycoefficientindustrialdefect
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

Unsupervised anomaly segmentation is usually scored by AUROC, but this paper argues that AUROC is inflated when anomalies occupy only a few percent of pixels, as in the MVTec and VisA datasets. The authors propose UniAS, a single one-for-all model that reconstructs normal features at multiple resolutions and combines the reconstruction errors into a pixel-level anomaly map. The central claim is that multi-level hybrid reconstruction, adding a multi-granularity gated CNN to transformer layers, preserves the low-level detail needed to segment defect boundaries while suppressing false-positive regions. On the two benchmarks the model reports pAP/DSC of 65.12/59.33 and 40.06/32.50, surpassing prior unified and per-class methods on these segmentation metrics. A reader should care because accurate pixel-level localization is what is actually needed for monitoring and removing industrial defects, and a single model would remove the cost of per-class training.

What carries the argument

The load-bearing mechanism is the multi-level hybrid decoder. At each level, features from a frozen CNN backbone are patchified and fed as keys and values to a transformer layer, whose cross-attention with a learnable query reconstructs the normal pattern globally; the query then passes through the MGG-CNN, four parallel convolutional branches gated by GELU with receptive fields of 1, 3, 5, and 7, which adds local detail at multiple granularities. A transposed convolution unpatchifies the result back to the original feature resolution, and training minimizes cosine distance plus mean squared error against the extracted features. At inference the per-level cosine distances are upsampled to image size and multiplied pixelwise, an operation that suppresses false positives because a pixel must be anomalous at every level to survive. A sample-aware reweighting module, using channel and spatial attention computed from the highest-level feature, reweights the random initial query so that one-for-all reconstruction adapts to the input sample.

What would settle it

Recompute UniAS and its baselines on MVTec-AD and VisA with a single threshold derived from normal training images, such as the 95th percentile of anomaly scores on normal validation data, instead of per-class oracle thresholds, and compare pAP and DSC; if the margin over HVQ-Trans and UniAD largely disappears, the headline DSC numbers depend on oracle thresholding rather than on the model's segmentation ability.

Watch

Extended reading notes

Core claim

The paper's discovery is that prior one-for-all anomaly detectors, by downsampling features and reconstructing at a coarse scale, trade away exactly the low-level information that pixel-level segmentation needs. UniAS instead reconstructs features at each of four levels in a coarse-to-fine decoder: a transformer layer supplies global context and a multi-granularity gated CNN (MGG-CNN) refines local details at receptive fields of 1, 3, 5, and 7 pixels. During inference, the cosine distance between reconstructed and extracted features yields an anomaly map per level, and the maps are upsampled and multiplied together; the authors show these maps play complementary roles, with high levels finding the anomaly semantically and low levels tracing its contour. With this design, plus a sample-aware reweighted query that makes the shared memory query more specific to each input, UniAS reports state-of-the-art pAP/DSC of 65.12/59.33 on MVTec-AD and 40.06/32.50 on VisA, and its ablations attribute most of the gain to the multi-level pipeline and the gated multi-granularity branch.

Load-bearing premise

The load-bearing premise is that comparing models on pAP and DSC with per-category thresholds tuned on the test set's ground-truth masks is the right way to measure anomaly segmentation; if those masks are not available at deployment, the reported Dice advantage may not survive.

Editorial extensions

If this is right

  • Pixel-level AUROC should no longer be the headline metric for anomaly segmentation; pAP and DSC expose false positives that AUROC hides.
  • A single unified model trained only on normal images can segment defects across many industrial object categories at pixel level, removing per-class fine-tuning.
  • Multi-level reconstruction with multiplicative fusion is a transferable design choice for other reconstruction-based detection models.
  • Fine anomaly boundaries, such as the thin defects in MVTec, become segmentable when low-level features are reconstructed rather than discarded.
  • The SAR query suggests that sample-adaptive initialization of shared memory queries can improve cross-category reconstruction without extra per-category parameters.

Reading between the lines

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

  • Editorial extension: the reported DSC values are computed with per-category thresholds selected on the test set by maximizing precision and recall; a deployment-oriented evaluation would set thresholds from normal training data only, and the comparative advantage may be smaller under that protocol.
  • Editorial extension: the same coarse-to-fine multiplicative fusion could be tested on medical anomaly segmentation, where lesions are often small and boundary accuracy is clinically important.
  • Editorial extension: because the feature-combination ablation shows the lowest level helps most when added to coarser levels, the method's gain may come less from the transformer than from preserving the shallowest features; an experiment that holds the decoder fixed and varies only the set of levels would separate these factors.
  • Editorial extension: the Gaussian-filter preprocessing with residual concatenation is presented as a detail, but the supplementary ablation shows it changes DSC by several points; varying filter sigma could be a cheap hyperparameter to tune further.
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.

Referee Report

3 major / 5 minor

Summary. The paper proposes UniAS, a unified one-for-all unsupervised anomaly segmentation model that reconstructs multi-level features through a hybrid Transformer and multi-granularity gated CNN (MGG-CNN) pipeline, with a sample-aware reweighted query. The authors argue that AUROC is an inflated metric under imbalanced anomaly settings and advocate pAP and DSC as segmentation-oriented metrics. On MVTec-AD and VisA they report 65.12/59.33 and 40.06/32.50 pAP/DSC, respectively, claiming state-of-the-art anomaly segmentation with significant improvements over prior methods.

Significance. If the results hold, UniAS offers a practical single-model, per-class-fine-tuning-free anomaly segmentation approach and contributes a useful critique of AUROC for imbalanced pixel-level evaluation. The threshold-free pAP ranking is credible, the multi-level design is supported by the ablation study in Sec. 4.7, and the code is shared, which strengthens reproducibility. The principal caveat is that the DSC half of the headline result depends on an oracle-threshold evaluation protocol that is not deployment-realistic; this affects the strength of the central SOTA claim as stated.

major comments (3)
  1. [Sec. 4.1 / Supp. A.2] The reported DSC values are obtained by choosing, for each category, the binarization threshold that maximizes precision plus recall on the test-set PR curve, using ground-truth masks. This is an oracle-threshold quantity, not a label-free prediction, and the large margins quoted in the Abstract and Sec. 4.6 (e.g., +26.70 DSC over HVQ-Trans on MVTec) are properties of that protocol rather than of a deployable anomaly-segmentation pipeline. The pAP metric is threshold-free and is not affected, but the DSC half of the headline claim should be re-reported under a threshold-free or fixed-threshold protocol, or clearly labeled as oracle-threshold DSC throughout.
  2. [Supp. A.2] The empty-ground-truth DSC rule for normal images (DSC = 1 if the predicted mask is empty, 0 if any pixel is predicted) is non-standard and, because a large fraction of test images are normal, it can dominate category-mean DSC and can either inflate or depress scores depending on where the oracle threshold lands. Please report DSC on anomalous test images only, and/or with a standard convention for empty ground truth, and quantify how this empty-GT rule affects the reported means.
  3. [Abstract / Sec. 4.6] The claim that UniAS 'surpasses previous methods significantly' is strongly supported for MVTec pAP (65.12 vs 59.20 for DeSTSeg), but on VisA the pAP margin over the next-best method is 1.83 points (40.06 vs 38.23 for RD4AD), and no variance or repeated-run information is provided. The strength of the 'significant' claim should be qualified, or supported with standard errors or confidence intervals.
minor comments (5)
  1. [Supp. A.1 / A.2 / Supp. Tables] The supplementary material contains unresolved citation placeholders ('[?]') and no corresponding reference list; these must be completed before publication.
  2. [Tables 2 and 3] The label '25’W ACV' for the proposed method should read 'WACV 2025' (or a consistent citation string for the venue).
  3. [Supp. B.3.2] There is a typo: 'datatset' should be 'dataset'.
  4. [Sec. 4.1] pAP and DSC are described but not defined by equations; since the paper argues for their use as primary metrics, formal definitions or explicit citations would improve clarity.
  5. [Supp. Tables 1 and 2] Several per-class DSC values are near zero (e.g., VisA Macaroni2 0.07, Pcb2 0.69); a brief comment on threshold degeneracy for low-anomaly-rate classes would help readers interpret the stability of the DSC protocol.

Circularity Check

1 steps flagged · score 5.0 of 10

No circularity in the model derivation; however, the DSC half of the SOTA claim is partly circular because per-class segmentation thresholds are fitted on test-set ground-truth masks, making the reported Dice values oracle-threshold numbers.

  1. fitted input called prediction [Sec. 4.1 (Metrics); Supplementary A.2 (Hyper-parameter Settings)]
    "In our implementation, we determine the segmentation threshold by maximizing the sum of precision and recall on the Precision-Recall (PR) curve for each category. ... During evaluation, the threshold for segmentation is chosen based on the PR curve of each class to maximize the sum of precision and recall, and DSC is calculated sample-wisely."

    The headline DSC numbers (59.33/32.50) are not label-free predictions of the model: the threshold that converts the anomaly map into a binary mask is chosen per category by maximizing precision and recall on the test-set PR curve, i.e. using the same ground-truth masks that are then used to compute DSC. The reported value is therefore an oracle-threshold score, fitted on the evaluation labels, not a deployment-realistic segmentation metric. The empty-GT rule additionally makes DSC reward predicting 'nothing' on normal images. Since the abstract and Sec. 4.6 present these DSC values as part of 'surpassing previous methods significantly', one of the two headline metrics is partly constructed from the evaluation target. The pAP metric and the model's anomaly-map generation (Eq.

full rationale

Walking the derivation chain, UniAS's reconstruction objective is trained on normal features only (cosine distance + MSE), and the anomaly map is defined as the product of per-level cosine dissimilarities (Eq. 6); no step defines the model in terms of the anomaly ground truth. There is no load-bearing self-citation: the cited prior works (UniAD, HVQ-Trans, Mask2Former, CBAM) are external baselines or standard components, not uniqueness theorems from the same authors, and the architecture choices are ablated rather than assumed. The only reduction-by-construction I can exhibit is in the evaluation protocol: the DSC numbers are obtained after fitting per-category thresholds on the test-set PR curve using ground-truth masks, and the supplementary empty-GT rule further conditions DSC on an all-or-nothing prediction for normal images. This makes the DSC half of the SOTA claim partially circular, although the pAP result and the relative ordering under a shared protocol retain independent content. The supplementary also contains unresolved citation placeholders (e.g., A.1 and A.2), which are completeness issues rather than circularity evidence. Overall score 5 reflects moderate, partial circularity confined to the DSC metric, not to the model derivation.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The central empirical claim rests on several assumptions that are standard in reconstruction-based UAD but unverified here: reconstruction error marks anomalies, frozen ImageNet features are class-agnostic enough for one-for-all reconstruction, and the chosen evaluation protocol is valid. The main extra fitted quantity is the per-category test-set threshold used for DSC.

free parameters (5)
  • Segmentation binarization threshold (per category) = Not reported; chosen to maximize precision and recall on test PR curve
    Used to compute DSC; selected with ground-truth masks in Sec. 4.1 and Supp. A.2, making DSC a label-dependent number.
  • Patch sizes per feature level = 8, 4, 2, 1 for levels 4 to 1
    Chosen by hand in Supp. A.2 to make patch counts equal across levels; directly defines the multi-level reconstruction granularity.
  • Number of reconstructed feature levels = 4
    Selected by feature combination ablation on the MVTec-AD test set (Table 5); model selection uses the same benchmark as the final report.
  • MGG-CNN branch structure = 1x1 conv plus 3x3 convs stacked 1, 2, and 3 times (receptive fields 1, 3, 5, 7)
    Ablation in Table 4 shows it helps; chosen by hand over simpler Conv3 and ungated variants.
  • Loss balance between cosine distance and MSE = Implicitly equal weights (sum)
    No weighting or sensitivity analysis is provided; the two losses are simply summed following prior work, and the balance affects reconstruction quality.
assumptions (4)
  • domain assumption Reconstruction error is a valid anomaly score: normal features reconstruct well, anomaly-corrupted features reconstruct poorly.
    Core assumption of reconstruction-based UAD, stated in Sec. 2.1 and used throughout; the paper acknowledges the known 'identical shortcut' failure mode.
  • domain assumption A frozen ImageNet-pretrained EfficientNet provides sufficiently rich multi-scale features for all classes in the one-for-all setting.
    Sec. 3.2 and 4.3 fix the backbone and do not fine-tune it; if the features do not transfer, the decoder cannot reconstruct normal data well.
  • domain assumption Anomalies are rare and small, making AUROC insensitive to false positives, so pAP and DSC are the correct evaluation metrics.
    Used in Sec. 3.1.2 to motivate the whole paper; depends on dataset statistics and relies on cited arguments from prior work.
  • ad hoc to paper Per-category test-set threshold selection by maximizing precision and recall is an acceptable evaluation protocol.
    Introduced in Sec. 4.1 and Supp. A.2; not a standard unsupervised protocol and biases DSC upward.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Towards Accurate Unified Anomaly Segmentation." pith.science (2026). https://pith.science/paper/WEUC3KKI

@misc{pith2026250112295,
  author       = {Pith},
  title        = {Pith review of: Towards Accurate Unified Anomaly Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WEUC3KKI}},
  note         = {Machine review of arXiv:2501.12295}
}
read the original abstract

Unsupervised anomaly detection (UAD) from images strives to model normal data distributions, creating discriminative representations to distinguish and precisely localize anomalies. Despite recent advancements in the efficient and unified one-for-all scheme, challenges persist in accurately segmenting anomalies for further monitoring. Moreover, this problem is obscured by the widely-used AUROC metric under imbalanced UAD settings. This motivates us to emphasize the significance of precise segmentation of anomaly pixels using pAP and DSC as metrics. To address the unsolved segmentation task, we introduce the Unified Anomaly Segmentation (UniAS). UniAS presents a multi-level hybrid pipeline that progressively enhances normal information from coarse to fine, incorporating a novel multi-granularity gated CNN (MGG-CNN) into Transformer layers to explicitly aggregate local details from different granularities. UniAS achieves state-of-the-art anomaly segmentation performance, attaining 65.12/59.33 and 40.06/32.50 in pAP/DSC on the MVTec-AD and VisA datasets, respectively, surpassing previous methods significantly. The codes are shared at https://github.com/Mwxinnn/UniAS.

Figures

Figures reproduced from arXiv: 2501.12295 by the authors.

Figure 1
Figure 1. Prediction examples (Left) of the MVTec dataset [ [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. (Left) An example of MVTec-AD with a red line delin [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. (Left) The overview of UniAS. After extracting features, our multi-level hybrid decoder, composed of Transformer and MGG￾CNN hybrid blocks, hierarchically reconstructs normal features from coarse to fine. Transformers and MGG-CNNs play complementary roles in global and local modeling. The SAR Query is incorporated to facilitate One-for-All segmentation. (Right) Illustration of the training loss and anomaly map gener… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Qualitative results on MVTec-AD(left) and VisA(right). We visualize the anomaly maps and the generated masks. UniAS localizes anomaly precisely, exhibiting meaningful segmentation results. Image Level4 Level3 Level2 Level1 Anomaly Map GT [PITH_FULL_IMAGE:figures/full_…
Figure 5
Figure 5. Figure 5: (Left) Examples of input images and corresponding anomaly maps from feature level 4 to level 1. (Right) Aggregated anomaly [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 1
Figure 1. Figure 1: The illustration of the Sample-Aware Reweighting [PITH_FULL_IMAGE:figures/full_fig_p012_1.png]
Figure 2
Figure 2. Figure 2: Examples of failed cases of our UniAS. Our model can [PITH_FULL_IMAGE:figures/full_fig_p014_2.png]
Figure 3
Figure 3. Figure 3: Additional visualization of the anomaly maps in each level and final predictions. UniAS leverages the benefits of multi-level [PITH_FULL_IMAGE:figures/full_fig_p015_3.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Pre-Trained LLM is a Semantic-Aware and Generalizable Segmentation Booster

    cs.CV 2025-06 conditional novelty 6.0 of 10

    A frozen pre-trained LLM layer inserted between a CNN encoder and decoder improves medical image segmentation across ultrasound, dermoscopy, polyp, and CT benchmarks with few added trainable parameters.

Reference graph

Works this paper leans on

54 extracted references · 37 canonical work pages · cited by 1 Pith paper

  1. [1]

    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 Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 9592–9600, 2019. 1, 3, 5

    work page 2019

  2. [2]

    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. arXiv preprint arXiv:1807.02011, 2018. 2

    arXiv 2018

  3. [3]

    AUPIMO: Redefining Visual Anomaly Detection Benchmarks with High Speed and Low Tolerance

    Joao PC Bertoldo, Dick Ameln, Ashwin Vaidya, and Samet Akc ¸ay. Aupimo: Redefining visual anomaly detection benchmarks with high speed and low tolerance. arXiv preprint arXiv:2401.01984, 2024. 2, 3

    work page Pith review arXiv 2024

  4. [4]

    Swin-unet: Unet-like pure transformer for medical image segmentation

    Hu Cao, Yueyue Wang, Joy Chen, Dongsheng Jiang, Xi- aopeng Zhang, Qi Tian, and Manning Wang. Swin-unet: Unet-like pure transformer for medical image segmentation. In European conference on computer vision, pages 205–218. Springer, 2022. 4

    work page 2022

  5. [5]

    End-to- end object detection with transformers

    Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey Zagoruyko. End-to- end object detection with transformers. In European confer- ence on computer vision, pages 213–229. Springer, 2020. 3, 4

    work page 2020

  6. [6]

    Transunet: Transformers make strong encoders for medi- cal image segmentation

    Jieneng Chen, Yongyi Lu, Qihang Yu, Xiangde Luo, Ehsan Adeli, Yan Wang, Le Lu, Alan L Yuille, and Yuyin Zhou. Transunet: Transformers make strong encoders for medi- cal image segmentation. arXiv preprint arXiv:2102.04306,

    arXiv

  7. [7]

    Utrad: Anomaly detection and localization with u-transformer

    Liyang Chen, Zhiyuan You, Nian Zhang, Juntong Xi, and Xinyi Le. Utrad: Anomaly detection and localization with u-transformer. Neural Networks, 147:53–62, 2022. 2, 3

    work page 2022

  8. [8]

    Masked-attention mask transformer for universal image segmentation

    Bowen Cheng, Ishan Misra, Alexander G Schwing, Alexan- der Kirillov, and Rohit Girdhar. Masked-attention mask transformer for universal image segmentation. In Proceed- ings of the IEEE/CVF conference on computer vision and pattern recognition, pages 1290–1299, 2022. 3, 4

    work page 2022

Show all 54 references
  1. [9]

    Per- pixel classification is not all you need for semantic segmen- tation

    Bowen Cheng, Alex Schwing, and Alexander Kirillov. Per- pixel classification is not all you need for semantic segmen- tation. Advances in Neural Information Processing Systems, 34:17864–17875, 2021. 3, 4

    2021

  2. [10]

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

    Thomas Defard, Aleksandr Setkov, Angelique Loesch, and Romaric Audigier. Padim: a patch distribution modeling framework for anomaly detection and localization. In Inter- national Conference on Pattern Recognition, pages 475–489. Springer, 2021. 2, 3

    2021

  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 Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 9737–9746, 2022. 2, 3, 6

    2022

  4. [12]

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

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, et al. An image is worth 16x16 words: Trans- formers for image recognition at scale. arXiv preprint a...

    2010 arXiv

  5. [13]

    Deep learning for medical anomaly detection–a survey

    Tharindu Fernando, Harshala Gammulle, Simon Denman, Sridha Sridharan, and Clinton Fookes. Deep learning for medical anomaly detection–a survey. ACM Computing Sur- veys (CSUR), 54(7):1–37, 2021. 1

    2021

  6. [14]

    Cflow-ad: Real-time unsupervised anomaly detection with localization via conditional normalizing flows

    Denis Gudovskiy, Shun Ishizaka, and Kazuki Kozuka. Cflow-ad: Real-time unsupervised anomaly detection with localization via conditional normalizing flows. In Proceed- ings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 98–107, 2022. 2, 3

    2022

  7. [15]

    Diad: A diffusion-based framework for multi-class anomaly detection

    Haoyang He, Jiangning Zhang, Hongxu Chen, Xuhai Chen, Zhishan Li, Xu Chen, Yabiao Wang, Chengjie Wang, and Lei Xie. Diad: A diffusion-based framework for multi-class anomaly detection. arXiv preprint arXiv:2312.06607, 2023. 1, 2, 3, 5, 6

    2023 arXiv

  8. [16]

    Gaussian error linear units (gelus)

    Dan Hendrycks and Kevin Gimpel. Gaussian error linear units (gelus). arXiv preprint arXiv:1606.08415, 2016. 4

    2016 arXiv

  9. [17]

    Label-free liver tumor segmentation

    Qixin Hu, Yixiong Chen, Junfei Xiao, Shuwen Sun, Jieneng Chen, Alan L Yuille, and Zongwei Zhou. Label-free liver tumor segmentation. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 7422–7432, 2023. 5

    2023

  10. [18]

    Oneformer: One transformer to rule universal image segmentation

    Jitesh Jain, Jiachen Li, Mang Tik Chiu, Ali Hassani, Nikita Orlov, and Humphrey Shi. Oneformer: One transformer to rule universal image segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2989–2998, 2023. 3

    2023

  11. [19]

    Variational inference with normalizing flows

    Danilo Jimenez Rezende and Shakir Mohamed. Variational inference with normalizing flows. arXiv e-prints , pages arXiv–1505, 2015. 2

    2015

  12. [20]

    Sanflow: Semantic-aware normalizing flow for anomaly detection

    Daehyun Kim, Sungyong Baik, and Tae Hyun Kim. Sanflow: Semantic-aware normalizing flow for anomaly detection. In Thirty-seventh Conference on Neural Information Process- ing Systems, 2023. 2

    2023

  13. [21]

    Panoptic segmentation

    Alexander Kirillov, Kaiming He, Ross Girshick, Carsten Rother, and Piotr Doll ´ar. Panoptic segmentation. In Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 9404–9413, 2019. 3

    2019

  14. [22]

    Cutpaste: Self-supervised learning for anomaly de- tection and localization

    Chun-Liang Li, Kihyuk Sohn, Jinsung Yoon, and Tomas Pfister. Cutpaste: Self-supervised learning for anomaly de- tection and localization. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 9664–9674, 2021. 2, 3

    2021

  15. [23]

    Feature pyra- mid networks for object detection

    Tsung-Yi Lin, Piotr Doll ´ar, Ross Girshick, Kaiming He, Bharath Hariharan, and Serge Belongie. Feature pyra- mid networks for object detection. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 2117–2125, 2017. 3

    2017

  16. [24]

    Swin transformer: Hierarchical vision transformer using shifted windows

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF international conference on computer vision, pages 10012–10022, 2021. 4

    2021

  17. [25]

    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. 2, 3, 4, 6

    2023

  18. [26]

    Decoupled weight decay regularization

    Ilya Loshchilov and Frank Hutter. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101, 2017. 6

    2017 arXiv

  19. [27]

    Hierarchical vector quantized transformer for multi-class unsupervised anomaly detection

    Ruiying Lu, YuJie Wu, Long Tian, Dongsheng Wang, Bo Chen, Xiyang Liu, and Ruimin Hu. Hierarchical vector quantized transformer for multi-class unsupervised anomaly detection. Advances in Neural Information Processing Sys- tems, 36:8487–8500, 2023. 1, 2, 3, 6, 7

    2023

  20. [28]

    On pixel-level performance assessment in anomaly detection

    Mehdi Rafiei, Toby P Breckon, and Alexandros Iosifidis. On pixel-level performance assessment in anomaly detection. arXiv preprint arXiv:2310.16435, 2023. 2, 3

    2023 arXiv

  21. [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. 1, 2, 3, 4, 6, 7

    2022

  22. [30]

    Fully convolutional cross-scale-flows for image- based defect detection

    Marco Rudolph, Tom Wehrbein, Bodo Rosenhahn, and Bas- tian Wandt. Fully convolutional cross-scale-flows for image- based defect detection. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision , pages 1088–1097, 2022. 2

    2022

  23. [31]

    The precision-recall plot is more informative than the roc plot when evaluat- ing binary classifiers on imbalanced datasets

    Takaya Saito and Marc Rehmsmeier. The precision-recall plot is more informative than the roc plot when evaluat- ing binary classifiers on imbalanced datasets. PloS one , 10(3):e0118432, 2015. 3

    2015

  24. [32]

    Multiresolution knowledge distillation for anomaly detection

    Mohammadreza Salehi, Niousha Sadjadi, Soroosh Baselizadeh, Mohammad H Rohban, and Hamid R Ra- biee. Multiresolution knowledge distillation for anomaly detection. In Proceedings of the IEEE/CVF confer- ence on computer vision and pattern recognition , pages 14902–14912, 2021. 5

    2021

  25. [33]

    Unsupervised anomaly detection with generative adversarial networks to guide marker discovery

    Thomas Schlegl, Philipp Seeb ¨ock, Sebastian M Waldstein, Ursula Schmidt-Erfurth, and Georg Langs. Unsupervised anomaly detection with generative adversarial networks to guide marker discovery. In International conference on in- formation processing in medical imaging , pages ...

    2017

  26. [34]

    Efficientnet: Rethinking model scaling for convolutional neural networks

    Mingxing Tan and Quoc Le. Efficientnet: Rethinking model scaling for convolutional neural networks. In International conference on machine learning, pages 6105–6114. PMLR,

  27. [35]

    Mobileutr: Revisiting the relationship between light-weight cnn and transformer for efficient medical image segmentation

    Fenghe Tang, Bingkun Nian, Jianrui Ding, Quan Quan, Jie Yang, Wei Liu, and S Kevin Zhou. Mobileutr: Revisiting the relationship between light-weight cnn and transformer for efficient medical image segmentation. arXiv preprint arXiv:2312.01740, 2023. 4

    2023 arXiv

  28. [36]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszko- reit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. Advances in neural information processing systems, 30, 2017. 4

    2017

  29. [37]

    Student-teacher feature pyramid matching for anomaly de- tection

    Guodong Wang, Shumin Han, Errui Ding, and Di Huang. Student-teacher feature pyramid matching for anomaly de- tection. arXiv preprint arXiv:2103.04257, 2021. 2

    2021 arXiv

  30. [38]

    Uncertainty-inspired open set learning for retinal anomaly identification

    Meng Wang, Tian Lin, Lianyu Wang, Aidi Lin, Ke Zou, Xinxing Xu, Yi Zhou, Yuanyuan Peng, Qingquan Meng, Yiming Qian, et al. Uncertainty-inspired open set learning for retinal anomaly identification. Nature Communications, 14(1):6757, 2023. 2

    2023

  31. [39]

    Cbam: Convolutional block attention module

    Sanghyun Woo, Jongchan Park, Joon-Young Lee, and In So Kweon. Cbam: Convolutional block attention module. In Proceedings of the European conference on computer vision (ECCV), pages 3–19, 2018. 5

    2018

  32. [40]

    Squid: Deep feature in-painting for unsupervised anomaly detec- tion

    Tiange Xiang, Yixiao Zhang, Yongyi Lu, Alan L Yuille, Chaoyi Zhang, Weidong Cai, and Zongwei Zhou. Squid: Deep feature in-painting for unsupervised anomaly detec- tion. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition , pages 23890–23901,

  33. [41]

    Adversarial medical im- age with hierarchical feature hiding

    Qingsong Yao, Zecheng He, Yuexiang Li, Yi Lin, Kai Ma, Yefeng Zheng, and S Kevin Zhou. Adversarial medical im- age with hierarchical feature hiding. IEEE Transactions on Medical Imaging, 2023. 1

    2023

  34. [42]

    Label-free segmentation of covid-19 lesions in lung ct.IEEE transactions on medical imaging , 40(10):2808–2819, 2021

    Qingsong Yao, Li Xiao, Peihang Liu, and S Kevin Zhou. Label-free segmentation of covid-19 lesions in lung ct.IEEE transactions on medical imaging , 40(10):2808–2819, 2021. 1, 5

    2021

  35. [43]

    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. 1, 2

    2023

  36. [44]

    One-for-all: Proposal masked cross-class anomaly detection

    Xincheng Yao, Chongyang Zhang, Ruoqi Li, Jun Sun, and Zhenyu Liu. One-for-all: Proposal masked cross-class anomaly detection. In Proceedings of the AAAI Confer- ence on Artificial Intelligence, volume 37, pages 4792–4800,

  37. [45]

    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. 1, 2, 3, 6, 8

    2022

  38. [46]

    Adtr: Anomaly detection transformer with feature reconstruction

    Zhiyuan You, Kai Yang, Wenhan Luo, Lei Cui, Yu Zheng, and Xinyi Le. Adtr: Anomaly detection transformer with feature reconstruction. In International Conference on Neu- ral Information Processing, pages 298–310. Springer, 2022. 3

  39. [47]

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

    Vitjan Zavrtanik, Matej Kristan, and Danijel Skoˇcaj. Draem- a discriminatively trained reconstruction embedding for sur- face anomaly detection. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 8330– 8339, 2021. 2, 5, 6

    2021

  40. [48]

    Defect-gan: High-fidelity defect synthesis for automated defect inspection

    Gongjie Zhang, Kaiwen Cui, Tzu-Yi Hung, and Shijian Lu. Defect-gan: High-fidelity defect synthesis for automated defect inspection. In Proceedings of the IEEE/CVF Win- ter Conference on Applications of Computer Vision , pages 2524–2534, 2021. 2

    2021

  41. [49]

    Prototypical residual networks for anomaly detection and localization

    Hui Zhang, Zuxuan Wu, Zheng Wang, Zhineng Chen, and Yu-Gang Jiang. Prototypical residual networks for anomaly detection and localization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 16281–16291, 2023. 2, 3

    2023

  42. [50]

    Exploring plain vit reconstruction for multi- class unsupervised anomaly detection

    Jiangning Zhang, Xuhai Chen, Yabiao Wang, Chengjie Wang, Yong Liu, Xiangtai Li, Ming-Hsuan Yang, and Dacheng Tao. Exploring plain vit reconstruction for multi- class unsupervised anomaly detection. arXiv preprint arXiv:2312.07495, 2023. 6

    2023 arXiv

  43. [51]

    Destseg: Segmentation guided denoising student-teacher for anomaly detection

    Xuan Zhang, Shiyu Li, Xi Li, Ping Huang, Jiulong Shan, and Ting Chen. Destseg: Segmentation guided denoising student-teacher for anomaly detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3914–3923, 2023. 2, 3, 5, 6

    2023

  44. [52]

    Omnial: A unified cnn framework for unsuper- vised anomaly localization

    Ying Zhao. Omnial: A unified cnn framework for unsuper- vised anomaly localization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 3924–3933, 2023. 1, 6

    2023

  45. [53]

    Deep autoen- coding gaussian mixture model for unsupervised anomaly detection

    Bo Zong, Qi Song, Martin Renqiang Min, Wei Cheng, Cris- tian Lumezanu, Daeki Cho, and Haifeng Chen. Deep autoen- coding gaussian mixture model for unsupervised anomaly detection. In International conference on learning represen- tations, 2018. 1

    2018

  46. [54]

    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. 6 Towards Accurate Unified Anomaly S...

    2022 arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.