Pith. sign in

REVIEW 3 major objections 5 minor 49 references

NOVA: A Benchmark for Anomaly Localization and Clinical Reasoning in Brain MRI

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read NOVA is a new evaluation-only benchmark of rare brain MRI pathologies where GPT-4o, Gemini 2.0 Flash, and Qwen2.5-VL-72B all exhibit large performance drops across anomaly localization, image captioning, and diagnostic reasoning.

desk verdict NOVA is a genuinely useful new benchmark that is undermined by a false 'double-blinded' claim in its annotation protocol, yet the dataset and baseline results are worth serious consideration. read the letter →

arxiv 2505.14064 v1 pith:ADFRWIEA submitted 2025-05-20 eess.IV cs.AIcs.CVcs.LG

Cosmin I. Bercea , Jun Li , Philipp Raffler , Evamaria O. Riedel , Lena Schmitzer , Angela Kurz , Felix Bitzer , Paula Ro{ss}müller
show 7 more authors
Julian Canisius Mirjam L. Beyrle Che Liu Wenjia Bai Bernhard Kainz Julia A. Schnabel Benedikt Wiestler
This is my paper · ORCID
classification eess.IVcs.AIcs.CVcs.LG
keywords modelsnovaclinicalacrossanomalybenchmarkbraindistribution
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

NOVA is a new benchmark for testing AI systems on brain MRI scans. It contains about 906 images taken from Eurorad, a public teaching platform, covering 281 rare brain diseases. Each image has a bounding box drawn by at least two radiologists around the abnormal area, a short caption describing the findings, and a clinical history. The images were kept in their original form, in various planes and sequences, to mimic real hospital conditions.

The benchmark asks AI models to do three things: point to the abnormal region in an image (localization), write a short description of what they see (captioning), and make a diagnosis when given the image plus the clinical history (reasoning). The authors tested three leading vision-language models: GPT-4o, Gemini 2.0 Flash, and Qwen2.5-VL-72B. Performance was low across the board. For localization, the best model found only about 38% of abnormalities at a lenient overlap threshold and produced hundreds of false positives. For captioning, clinical term F1 scores were below 20%, and for diagnosis, top-1 accuracy was around 24%.

The results suggest that these large, general-purpose models do not transfer well to rare medical conditions. However, the benchmark has some caveats. The radiologists who drew the bounding boxes had access to the original clinical description and diagnosis, while the models only see the image in the localization task. Also, each case is a single 2D slice, not a full 3D volume, and the dataset comes from a single European teaching repository. These factors could make the localization task harder or less fair than intended.

Extended reading notes

Core claim

The paper claims NOVA is the first benchmark to jointly evaluate anomaly localization, visual captioning, and diagnostic reasoning in brain MRI with 281 rare pathologies, and that leading VLMs show 'substantial performance drops across all tasks', establishing it as a rigorous testbed. If true, current vision-language models are unreliable for rare brain MRI analysis in zero-shot settings.

Load-bearing premise

The ground-truth bounding boxes in Task 1 were drawn by radiologists who had access to the full Eurorad clinical description and diagnosis (Section 3.2), whereas models receive only the MRI image. If the anomalies are visually identifiable only with clinical context, the localization scores do not purely measure visual anomaly detection, undermining the benchmark's validity for Task 1.

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

3 major / 5 minor

Summary. NOVA is an evaluation-only benchmark compiled from 906 Eurorad brain MRI cases covering 281 rare diagnoses. The paper defines three tasks: anomaly localization from a single released MRI slice via bounding boxes, image captioning, and diagnostic reasoning from clinical history plus image caption. The authors benchmark GPT-4o, Gemini 2.0 Flash, and Qwen2.5-VL-72B and report substantial performance drops on all three tasks relative to natural-image benchmarks, arguing that current vision-language models are unreliable for rare brain MRI analysis in zero-shot settings. The dataset and annotations are released under a CC BY-NC-SA license.

Significance. If the validity concerns are resolved, NOVA addresses a genuine gap: no existing benchmark jointly evaluates anomaly localization, clinical captioning, and diagnostic reasoning on rare brain MRI. The breadth of diagnoses (281), the real-world acquisition heterogeneity, the expert multi-reader annotation protocol, and the inference-only design are strengths, and the authors explicitly acknowledge possible pretraining contamination. The paper also avoids fitted parameters, so the benchmark is not self-fulfilling in the usual sense. However, the two central validity issues—unblinded ground-truth boxes for Task 1 and GPT-4o serving as both judge and contestant in Task 3—must be addressed before the benchmark can support the headline claim that VLMs are unreliable in open-world clinical brain MRI.

major comments (3)
  1. [§3.2, Abstract] The abstract states that NOVA provides 'double-blinded expert bounding-box annotations,' but Section 3.2 reports that 'Each case was independently labeled by two readers, who reviewed the full original Eurorad clinical description and associated metadata to inform their annotations.' The readers therefore knew the clinical history and likely the diagnosis before drawing boxes, and the senior adjudicator also had this context. Since Task 1 (Section 4.1) gives models only the image, low localization scores conflate visual anomaly detection with the information asymmetry between annotators and models. This is an internal inconsistency: the stated protocol contradicts the 'double-blinded' claim. The authors should either re-annotate a representative subset with image-only readers to quantify the effect, or rescope Task 1 as localization with clinical priors and soften the corresponding claims about VLM visual unreliability.
  2. [§4.3, Table 3] Task 3 uses GPT-4o 'to perform semantic matching between predictions and ground truth labels,' while GPT-4o is itself one of the models evaluated on the same task. This introduces a self-scoring circularity: GPT-4o's own predictions may be judged more leniently than those of other models if the judge favors its own output distribution. The authors should use a different judge model (e.g., a strong open-source model) or human raters, and should report inter-annotator agreement on a sample of diagnostic predictions to confirm that the semantic matching is not biased.
  3. [§3.3, §6] The release format is described as 'uniformly sized 480×480 grayscale PNG slices,' and Section 6 acknowledges that only 2D slices are provided, yet the manuscript repeatedly refers to 'scans' and claims to evaluate anomaly localization in brain MRI. A single axial, sagittal, or coronal slice may not contain the full lesion burden or the most representative plane, and the paper does not state how the slice was selected for each case. This limits the external validity of the localization and captioning results. The authors should document the slice-selection rule and report the proportion of cases in which the selected slice actually contains the annotated abnormality.
minor comments (5)
  1. [Throughout] The spelling 'NOVA' and 'NOV A' is used inconsistently; please unify the dataset name.
  2. [§5.1, Table 1] The sentence 'over 600 false-positive boxes were recorded' is ambiguous because Table 1 reports FP30 values of 899, 1163, and 672 per model; the text should explicitly state whether this is per-model or total.
  3. [§3.3] Please clarify whether each 'case' is represented by exactly one slice or by multiple slices, and explain how multiple ground-truth bounding boxes per case map onto the released slice(s).
  4. [Table 3] The column header 'Cov.' is not defined in the caption; please state explicitly that it denotes the fraction of ground-truth labels covered by the model's prediction vocabulary.
  5. [§4.2, Table 2] The binary normal-versus-abnormal F1 values in Table 2 are extremely low (2.4–11.3%); since all Eurorad cases are pathological, the paper should explain how the binary label is derived and whether the metric measures any meaningful signal for these models.
Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The benchmark introduces no new physical entities. Its central claim rests on domain assumptions about data representativeness, the use of 2D slices, the validity of annotations drawn with clinical history, and the applicability of standard metrics.

free parameters (2)
  • Consensus IoU threshold = 0.3
    Annotations with IoU > 0.3 were merged into consensus labels, an arbitrary threshold that directly determines the ground truth and evaluation targets.
  • Single 2D slice selection = 1 per case
    Each case is reduced to one 480x480 PNG slice; the choice of slice is not explained and can change the presence and location of the anomaly.
assumptions (5)
  • domain assumption Eurorad cases are representative of rare brain MRI pathologies in clinical practice.
    All data is sourced from one European teaching repository, which may not generalize globally (acknowledged in Section 6).
  • domain assumption A single 2D slice per case is sufficient for the three benchmark tasks.
    The paper deliberately releases 2D slices, but this discards volumetric context and may miss the lesion if the slice is off-center (Section 6).
  • ad hoc to paper Radiologist bounding boxes drawn with knowledge of the clinical history are a valid visual ground truth for anomaly localization.
    Annotators reviewed the full clinical description and diagnosis before drawing boxes (Section 3.2), so the boxes may encode information not present in the image.
  • domain assumption The 281 diagnosis labels are mutually exclusive and correctly assigned.
    Diagnoses come from Eurorad, a peer-reviewed teaching platform, but no independent re-validation is reported.
  • standard math Standard evaluation metrics (mAP, BLEU, METEOR, exact keyword matching) are appropriate for the three tasks.
    These are widely used metrics, though keyword matching and GPT-4o semantic matching have known reliability limits.

how reviews work

0 comments
Cite this review

Pith. "Pith review of NOVA: A Benchmark for Anomaly Localization and Clinical Reasoning in Brain MRI." pith.science (2026). https://pith.science/paper/ADFRWIEA

@misc{pith2026250514064,
  author       = {Pith},
  title        = {Pith review of: NOVA: A Benchmark for Anomaly Localization and Clinical Reasoning in Brain MRI},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ADFRWIEA}},
  note         = {Machine review of arXiv:2505.14064}
}
abstract

In many real-world applications, deployed models encounter inputs that differ from the data seen during training. Out-of-distribution detection identifies whether an input stems from an unseen distribution, while open-world recognition flags such inputs to ensure the system remains robust as ever-emerging, previously $unknown$ categories appear and must be addressed without retraining. Foundation and vision-language models are pre-trained on large and diverse datasets with the expectation of broad generalization across domains, including medical imaging. However, benchmarking these models on test sets with only a few common outlier types silently collapses the evaluation back to a closed-set problem, masking failures on rare or truly novel conditions encountered in clinical use. We therefore present $NOVA$, a challenging, real-life $evaluation-only$ benchmark of $\sim$900 brain MRI scans that span 281 rare pathologies and heterogeneous acquisition protocols. Each case includes rich clinical narratives and double-blinded expert bounding-box annotations. Together, these enable joint assessment of anomaly localisation, visual captioning, and diagnostic reasoning. Because NOVA is never used for training, it serves as an $extreme$ stress-test of out-of-distribution generalisation: models must bridge a distribution gap both in sample appearance and in semantic space. Baseline results with leading vision-language models (GPT-4o, Gemini 2.0 Flash, and Qwen2.5-VL-72B) reveal substantial performance drops across all tasks, establishing NOVA as a rigorous testbed for advancing models that can detect, localize, and reason about truly unknown anomalies.

Figures

Figures reproduced from arXiv: 2505.14064 by the authors.

Figure 1
Figure 1. Overview of the NOVA benchmark. Task 1: Anomaly localization: models predict bounding [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Representative brain MRI scans from the NOVA dataset illustrating the diversity of [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Dataset composition and annotation quality in NOVA. (a) Distribution of cases across [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Examples of model predictions for anomaly grounding on NOVA. Ground truth and model [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Task2: Image Captioning. (Top) Example image-caption pairs showing model predictions [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Distribution of diagnostic label frequencies [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

49 extracted references · 41 canonical work pages

  1. [1]

    https://brain-development.org/ixi-dataset/

    Ixi dataset. https://brain-development.org/ixi-dataset/ . Accessed: 2023-02-15

    work page 2023

  2. [2]

    METEOR: An automatic metric for MT evaluation with improved correlation with human judgments

    Satanjeev Banerjee and Alon Lavie. METEOR: An automatic metric for MT evaluation with improved correlation with human judgments. In Jade Goldstein, Alon Lavie, Chin-Yew Lin, and Clare V oss, editors,Proceedings of the ACL Workshop on Intrinsic and Extrinsic Evaluation Measures for Machine Translation and/or Summarization, pages 65–72, Ann Arbor, Michigan,...

    2005

  3. [3]

    Bercea, Benedikt Wiestler, Daniel Rueckert, and Julia A Schnabel

    Cosmin I. Bercea, Benedikt Wiestler, Daniel Rueckert, and Julia A Schnabel. Generalizing unsupervised anomaly detection: Towards unbiased pathology screening. In Medical Imaging with Deep Learning, 2023

    work page 2023

  4. [4]

    Diffusion models with implicit guidance for medical anomaly detection

    Cosmin I Bercea, Benedikt Wiestler, Daniel Rueckert, and Julia A Schnabel. Diffusion models with implicit guidance for medical anomaly detection. In International Conference on Medical Image Computing and Computer-Assisted Intervention, pages 211–220. Springer, 2024

    work page 2024

  5. [5]

    Evaluating normative representation learning in generative ai for robust anomaly detection in brain imaging

    Cosmin I Bercea, Benedikt Wiestler, Daniel Rueckert, and Julia A Schnabel. Evaluating normative representation learning in generative ai for robust anomaly detection in brain imaging. Nature Communications, 16(1):1624, 2025

    work page 2025

  6. [6]

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

    Paul Bergmann, Michael Fauser, David Sattlegger, and Carsten Steger. Mvtec ad — a compre- hensive real-world dataset for unsupervised anomaly detection. In 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 9584–9592, 2019

    work page 2019

  7. [7]

    Salinas, and María de la Iglesia-Vayá

    Agustín Bustos, Antonio Pertusa, Jose M. Salinas, and María de la Iglesia-Vayá. Padchest: A large chest x-ray image dataset with multi-label annotated reports. Medical Image Analysis, 66:101797, 2020

    work page 2020

  8. [8]

    Padchest-gr: A bilingual chest x-ray dataset for grounded radiology report generation

    Daniel C Castro, Aurelia Bustos, Shruthi Bannur, Stephanie L Hyland, Kenza Bouzid, Maria Teodora Wetscherek, Maria Dolores Sánchez-Valverde, Lara Jaques-Pérez, Lourdes Pérez-Rodríguez, Kenji Takeda, et al. Padchest-gr: A bilingual chest x-ray dataset for grounded radiology report generation. arXiv preprint arXiv:2411.05085, 2024

    arXiv 2024

Show all 49 references
  1. [9]

    Unsupervised lesion detection via image restoration with a normative prior

    Xiaoran Chen, Suhang You, Kerem Can Tezcan, and Ender Konukoglu. Unsupervised lesion detection via image restoration with a normative prior. Medical Image Analysis, 64:101713, 2020

    2020

  2. [10]

    An image is worth 16x16 words: Transformers for image recognition at scale

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

    2010 arXiv

  3. [11]

    Promptdet: Towards open-vocabulary detection using uncurated images

    Chengjian Feng, Yujie Zhong, Zequn Jie, Xiangxiang Chu, Haibing Ren, Xiaolin Wei, Weidi Xie, and Lin Ma. Promptdet: Towards open-vocabulary detection using uncurated images. In European conference on computer vision, pages 701–717. Springer, 2022

    2022

  4. [12]

    Recontrast: Domain-specific anomaly detection via contrastive reconstruction

    Jia Guo, Shuai Lu, Lize Jia, Weihang Zhang, and Huiqi Li. Recontrast: Domain-specific anomaly detection via contrastive reconstruction. In A. Oh, T. Naumann, A. Globerson, K. Saenko, M. Hardt, and S. Levine, editors, Advances in Neural Information Processing Systems, volume 36...

    2023

  5. [13]

    Developing generalist foun- dation models from a multimodal dataset for 3d computed tomography

    Ibrahim Ethem Hamamci, Sezgin Er, Furkan Almas, et al. Developing generalist foun- dation models from a multimodal dataset for 3d computed tomography. arXiv preprint arXiv:2403.17834, 2024

    2024

  6. [14]

    Deep anomaly detection with outlier exposure

    Dan Hendrycks, Mantas Mazeika, and Thomas Dietterich. Deep anomaly detection with outlier exposure. arXiv preprint arXiv:1812.04606, 2018

    2018 arXiv

  7. [15]

    Isles 2022: A multi-center magnetic resonance imaging stroke lesion segmentation dataset

    Moritz R Hernandez Petzsche, Ezequiel de la Rosa, Uta Hanning, Roland Wiest, Waldo Valenzuela, Mauricio Reyes, Maria Meyer, Sook-Lei Liew, Florian Kofler, Ivan Ezhov, et al. Isles 2022: A multi-center magnetic resonance imaging stroke lesion segmentation dataset. Scientific Da...

    2022

  8. [16]

    Out-of-distribution detection in medical image analysis: A survey

    Zesheng Hong, Yubiao Yue, Yubin Chen, Lele Cong, Huanjie Lin, Yuanmei Luo, Mini Han Wang, Weidong Wang, Jialong Xu, Xiaoqi Yang, et al. Out-of-distribution detection in medical image analysis: A survey. arXiv preprint arXiv:2404.18279, 2024

    2024 arXiv

  9. [17]

    Alistair E. W. Johnson, Tom J. Pollard, Seth J. Berkowitz, et al. MIMIC-CXR, a de-identified publicly available database of chest radiographs with free-text reports. Scientific Data, 6:317, 2019

    2019

  10. [18]

    Y . W. Kim and L. T. Mansfield. Fool me twice: Delayed diagnoses in radiology with emphasis on perpetuated errors. AJR. American Journal of Roentgenology, 202(3):465–470, 2014

    2014

  11. [19]

    Wilds: A benchmark of in-the-wild distribution shifts

    Pang Wei Koh, Shiori Sagawa, Henrik Marklund, Sang Michael Xie, Marvin Zhang, Akshay Balsubramani, Weihua Hu, Michihiro Yasunaga, Richard Lanas Phillips, Irena Gao, Tony Lee, Etienne David, Ian Stavness, Wei Guo, Berton Earnshaw, Imran Haque, Sara M Beery, Jure Leskovec, Anshu...

    2021

  12. [20]

    Emergency triage of brain computed tomography via anomaly detection with a deep generative model

    Seungjun Lee, Boryeong Jeong, Minjee Kim, Ryoungwoo Jang, Wooyul Paik, Jiseon Kang, Won Chung, Gil-Sun Hong, and Namkug Kim. Emergency triage of brain computed tomography via anomaly detection with a deep generative model. Nature Communications, 13:4251, 07 2022

    2022

  13. [21]

    A novel public MR image dataset of multiple sclerosis patients with lesion segmentations based on multi-rater consensus

    Žiga Lesjak, Alina Galimzianova, Andrej Koren, et al. A novel public MR image dataset of multiple sclerosis patients with lesion segmentations based on multi-rater consensus. Neuroin- formatics, 16(1):51–63, 2018

    2018

  14. [22]

    Cutpaste: Self-supervised learning for anomaly detection and localization

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

    2021

  15. [23]

    Lo, ., and et al

    Sook-Lei Liew, Bethany P. Lo, ., and et al. Miarnda R. Donnelly. A large, curated, open-source stroke neuroimaging dataset to improve lesion segmentation algorithms. Scientific Data, 9, 2022

    2022

  16. [24]

    A benchmarking crisis in biomedical machine learning

    Faisal Mahmood. A benchmarking crisis in biomedical machine learning. Nature Medicine, 31(4):1060–1060, 2025

    2025

  17. [25]

    Marcus, Aditya F

    Daniel S. Marcus, Aditya F. Fotenos, John G. Csernansky, John C. Morris, and Randy L. Buckner. Open access series of imaging studies: longitudinal MRI data in nondemented and demented older adults. Journal of Cognitive Neuroscience, 22(12):2677–2684, 2010

    2010

  18. [26]

    Bjoern H. Menze, Andras Jakab, Stefan Bauer, Jayashree Kalpathy-Cramer, Keyvan Farahani, Justin Kirby, Yuliya Burren, Nicole Porz, Johannes Slotboom, Roland Wiest, Levente Lanczi, Elizabeth Gerstner, Marc-André Weber, Tal Arbel, Brian B. Avants, Nicholas Ayache, Patricia Buend...

    1993

  19. [27]

    Bleu: a method for automatic evaluation of machine translation

    Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. Bleu: a method for automatic evaluation of machine translation. In Pierre Isabelle, Eugene Charniak, and Dekang Lin, editors, Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics , ...

    2002

  20. [28]

    Petersen, Paul S

    Ronald C. Petersen, Paul S. Aisen, Laurel A. Beckett, Michael C. Donohue, Anthony C. Gamst, Danielle J. Harvey, Clifford R. Jr. Jack, William J. Jagust, Leslie M. Shaw, Arthur W. Toga, John Q. Trojanowski, and Michael W. Weiner. Alzheimer’s disease neuroimaging initiative (adn...

    2010

  21. [29]

    Generative probabilistic novelty detection with adversarial autoencoders

    Stanislav Pidhorskyi, Ranya Almohsen, and Gianfranco Doretto. Generative probabilistic novelty detection with adversarial autoencoders. Advances in neural information processing systems, 31, 2018

    2018

  22. [30]

    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, et al. Learning transferable visual models from natural language supervision. In International Conference on Machine Learning, pa...

    2021

  23. [31]

    Do imagenet classifiers generalize to imagenet? In Proceedings of the International Conference on Machine Learning, pages 5389–5400

    Benjamin Recht, Rebecca Roelofs, Ludwig Schmidt, and Vaishaal Shankar. Do imagenet classifiers generalize to imagenet? In Proceedings of the International Conference on Machine Learning, pages 5389–5400. PMLR, 2019

    2019

  24. [32]

    Buizer, Antonio Federico, Thomas Gasser, Samuel Groeschel, Sanja Hermanns, Thomas Klockgether, Ingeborg Krägeloh-Mann, G

    Carola Reinhard, Anne-Catherine Bachoud-Lévi, Tobias Bäumer, Enrico Bertini, Alicia Brunelle, Annemieke I. Buizer, Antonio Federico, Thomas Gasser, Samuel Groeschel, Sanja Hermanns, Thomas Klockgether, Ingeborg Krägeloh-Mann, G. Bernhard Landwehrmeyer, Is- abelle Leber, Alfons...

    2020

  25. [33]

    Towards total recall in industrial anomaly detection

    Karsten Roth, Latha Pemula, Joaquin Zepeda, Bernhard Schölkopf, Thomas Brox, and Peter Gehler. Towards total recall in industrial anomaly detection. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 14318–14328, 2022

    2022

  26. [34]

    A unifying review of deep and shallow anomaly detection

    Lukas Ruff, Jacob R Kauffmann, Robert A Vandermeulen, Grégoire Montavon, Wojciech Samek, Marius Kloft, Thomas G Dietterich, and Klaus-Robert Müller. A unifying review of deep and shallow anomaly detection. Proceedings of the IEEE, 109(5):756–795, 2021

    2021

  27. [35]

    f-anogan: Fast unsupervised anomaly detection with generative adversarial networks

    Thomas Schlegl, Philipp Seeböck, Sebastian M Waldstein, Georg Langs, and Ursula Schmidt- Erfurth. f-anogan: Fast unsupervised anomaly detection with generative adversarial networks. Medical Image Analysis, 54:30–44, 2019

    2019

  28. [36]

    Natural synthetic anomalies for self-supervised anomaly detection and localization

    Hannah M Schlüter, Jeremy Tan, Benjamin Hou, and Bernhard Kainz. Natural synthetic anomalies for self-supervised anomaly detection and localization. In European Conference on Computer Vision, pages 474–489. Springer, 2022

    2022

  29. [37]

    Uk biobank: An open access resource for identifying the causes of a wide range of complex diseases of middle and old age

    Cathie Sudlow, John Gallacher, Naomi Allen, Valerie Beral, Paul Burton, John Danesh, et al. Uk biobank: An open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Medicine, 12(3):e1001779, 2015

    2015

  30. [38]

    Taylor, Nitin Williams, Rhodri Cusack, Tibor Auer, Meredith A

    Jason R. Taylor, Nitin Williams, Rhodri Cusack, Tibor Auer, Meredith A. Shafto, Marie Dixon, Lorraine K. Tyler, Cam-CAN, and Richard N. Henson. The cambridge centre for ageing and neuroscience (cam-can) data repository: Structural and functional mri, meg, and cognitive data fr...

    2017

  31. [39]

    Bridging ood detection and generalization: A graph-theoretic view

    Han Wang and Yixuan Li. Bridging ood detection and generalization: A graph-theoretic view. Advances in Neural Information Processing Systems, 2024

    2024

  32. [40]

    Diffusion models for medical anomaly detection

    Julia Wolleb, Florentin Bieder, Robin Sandkühler, and Philippe C Cattin. Diffusion models for medical anomaly detection. In International Conference on Medical Image Computing and Computer-Assisted Intervention, pages 35–45. Springer, 2022

    2022

  33. [41]

    An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, et al. Qwen2. 5 technical report. arXiv preprint arXiv:2412.15115, 2024. 12

    2024 arXiv

  34. [42]

    Openood: Benchmarking generalized out-of- distribution detection

    Jingkang Yang, Pengyun Wang, Dejian Zou, Zitang Zhou, Kunyuan Ding, WENXUAN PENG, Haoqi Wang, Guangyao Chen, Bo Li, Yiyou Sun, Xuefeng Du, Kaiyang Zhou, Wayne Zhang, Dan Hendrycks, Yixuan Li, and Ziwei Liu. Openood: Benchmarking generalized out-of- distribution detection. In S...

    2022

  35. [43]

    Modeling context in referring expressions

    Licheng Yu, Patrick Poirson, Shan Yang, Alexander C Berg, and Tamara L Berg. Modeling context in referring expressions. In European Conference on Computer Vision, pages 69–85. Springer, 2016

    2016

  36. [44]

    fastmri: An open dataset and benchmarks for accelerated mri

    Jure Zbontar, Florian Knoll, Anuroop Sriram, Tullie Murrell, Zhengnan Huang, Matthew J Muckley, Aaron Defazio, Ruben Stern, Patricia Johnson, Mary Bruno, et al. fastmri: An open dataset and benchmarks for accelerated mri. arXiv preprint arXiv:1811.08839, 2018

    2018 arXiv

  37. [45]

    fastmri+, clinical pathology annotations for knee and brain fully sampled magnetic resonance imaging data

    Ruiyang Zhao, Burhaneddin Yaman, Yuxin Zhang, Russell Stewart, Austin Dixon, Florian Knoll, Zhengnan Huang, Yvonne W Lui, Michael S Hansen, and Matthew P Lungren. fastmri+, clinical pathology annotations for knee and brain fully sampled magnetic resonance imaging data. Scienti...

    2022

  38. [46]

    Out-of-distribution detection learning with unreliable out-of-distribution sources

    Haotian Zheng, Qizhou Wang, Zhen Fang, Xiaobo Xia, Feng Liu, Tongliang Liu, and Bo Han. Out-of-distribution detection learning with unreliable out-of-distribution sources. In A. Oh, T. Naumann, A. Globerson, K. Saenko, M. Hardt, and S. Levine, editors,Advances in Neural Inform...

    2023

  39. [47]

    A survey on open-vocabulary detection and segmentation: Past, present, and future

    Chaoyang Zhu and Long Chen. A survey on open-vocabulary detection and segmentation: Past, present, and future. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024

    2024

  40. [48]

    Mood 2020: A public benchmark for out-of-distribution detection and localization on medical images

    David Zimmerer, Peter M Full, Fabian Isensee, Paul Jäger, Tim Adler, Jens Petersen, Gregor Köhler, Tobias Ross, Annika Reinke, Antanas Kascenas, et al. Mood 2020: A public benchmark for out-of-distribution detection and localization on medical images. IEEE Transactions on Medi...

    2020

  41. [49]

    Unsu- pervised anomaly localization using variational auto-encoders

    David Zimmerer, Fabian Isensee, Jens Petersen, Simon Kohl, and Klaus Maier-Hein. Unsu- pervised anomaly localization using variational auto-encoders. In Medical Image Computing and Computer Assisted Intervention–MICCAI 2019: 22nd International Conference, Shenzhen, China, Octo...

    2019

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.