{"url":"/sota/multi-class-anomaly-detection-on-mvtec-ad","task":{"name":"Multi-class Anomaly Detection","url":"/task/multi-class-anomaly-detection","note":null},"dataset":{"name":"MVTec AD","url":"/dataset/mvtecad"},"category":"Computer Vision","categories":["Computer Vision","Graphs","Methodology","Miscellaneous"],"category_note":null,"description":"Multi-class Anomaly Detection is a task that identifies anomalies by jointly learning and detecting outliers across multiple classes, in contrast to traditional Anomaly Detection, which typically focuses on identifying anomalies within a single class.","description_from":"task","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","rank":"the archive's row order at snapshot; not re-ranked","rows_end_at":"2025-07-28","rows_withheld_as_spam":0,"metric_values":"the archive's strings, untouched"},"metrics":["Detection AUROC","Segmentation AUROC"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"Detection AUROC":null,"Segmentation AUROC":null}},"counts":{"rows":13,"rows_with_code":11,"rows_with_paper_page":13,"rows_dated":13,"rows_using_additional_data":0},"rows":[{"rank_in_archive_order":1,"model":"INP-Former-Large","metrics":{"Detection AUROC":"99.8","Segmentation AUROC":"98.6"},"uses_additional_data":false,"paper_date":"2025-03-04","paper":"/paper/exploring-intrinsic-normal-prototypes-within","paper_url":"https://arxiv.org/abs/2503.02424v1","paper_title":"Exploring Intrinsic Normal Prototypes within a Single Image for Universal Anomaly Detection","code":"https://github.com/luow23/inp-former","n_code_links":2,"syntology":{"n_ran":4,"n_unverified":1,"n_samples":5,"n_pointer_only_licence":0}},{"rank_in_archive_order":2,"model":"Dinomaly-Large","metrics":{"Detection AUROC":"99.8","Segmentation AUROC":"98.5"},"uses_additional_data":false,"paper_date":"2024-05-23","paper":"/paper/dinomaly-the-less-is-more-philosophy-in-multi","paper_url":"https://arxiv.org/abs/2405.14325v4","paper_title":"Dinomaly: The Less Is More Philosophy in Multi-Class Unsupervised Anomaly Detection","code":"https://github.com/guojiajeremy/dinomaly","n_code_links":2,"syntology":{"n_ran":9,"n_unverified":1,"n_samples":10,"n_pointer_only_licence":0}},{"rank_in_archive_order":3,"model":"INP-Former-Base","metrics":{"Detection AUROC":"99.7","Segmentation AUROC":"98.5"},"uses_additional_data":false,"paper_date":"2025-03-04","paper":"/paper/exploring-intrinsic-normal-prototypes-within","paper_url":"https://arxiv.org/abs/2503.02424v1","paper_title":"Exploring Intrinsic Normal Prototypes within a Single Image for Universal Anomaly Detection","code":"https://github.com/luow23/inp-former","n_code_links":2,"syntology":{"n_ran":4,"n_unverified":1,"n_samples":5,"n_pointer_only_licence":0}},{"rank_in_archive_order":4,"model":"Dinomaly-Base","metrics":{"Detection AUROC":"99.6","Segmentation AUROC":"98.4"},"uses_additional_data":false,"paper_date":"2024-05-23","paper":"/paper/dinomaly-the-less-is-more-philosophy-in-multi","paper_url":"https://arxiv.org/abs/2405.14325v4","paper_title":"Dinomaly: The Less Is More Philosophy in Multi-Class Unsupervised Anomaly Detection","code":"https://github.com/guojiajeremy/dinomaly","n_code_links":2,"syntology":{"n_ran":9,"n_unverified":1,"n_samples":10,"n_pointer_only_licence":0}},{"rank_in_archive_order":5,"model":"RLR","metrics":{"Detection AUROC":"98.6","Segmentation AUROC":"98.5"},"uses_additional_data":false,"paper_date":"2024-03-18","paper":"/paper/learning-unified-reference-representation-for","paper_url":"https://arxiv.org/abs/2403.11561v2","paper_title":"Learning Unified Reference Representation for Unsupervised Multi-class Anomaly Detection","code":"https://github.com/hlr7999/rlr","n_code_links":1,"syntology":{"n_ran":6,"n_unverified":1,"n_samples":7,"n_pointer_only_licence":7}},{"rank_in_archive_order":6,"model":"MambaAD","metrics":{"Detection AUROC":"98.6","Segmentation AUROC":"97.7"},"uses_additional_data":false,"paper_date":"2024-04-09","paper":"/paper/mambaad-exploring-state-space-models-for","paper_url":"https://arxiv.org/abs/2404.06564v3","paper_title":"MambaAD: Exploring State Space Models for Multi-class Unsupervised Anomaly Detection","code":"https://github.com/zhangzjn/ader","n_code_links":3,"syntology":{"n_ran":3,"n_unverified":0,"n_samples":3,"n_pointer_only_licence":3}},{"rank_in_archive_order":7,"model":"HGAD","metrics":{"Detection AUROC":"98.4","Segmentation AUROC":"97.9"},"uses_additional_data":false,"paper_date":"2024-03-20","paper":"/paper/hierarchical-gaussian-mixture-normalizing","paper_url":"https://arxiv.org/abs/2403.13349v2","paper_title":"Hierarchical Gaussian Mixture Normalizing Flow Modeling for Unified Anomaly Detection","code":"https://github.com/xcyao00/hgad","n_code_links":1,"syntology":{"n_ran":1,"n_unverified":2,"n_samples":3,"n_pointer_only_licence":3}},{"rank_in_archive_order":8,"model":"PNPT","metrics":{"Detection AUROC":"98.3","Segmentation AUROC":"98.1"},"uses_additional_data":false,"paper_date":"2024-06-17","paper":"/paper/prior-normality-prompt-transformer-for-multi","paper_url":"https://arxiv.org/abs/2406.11507v1","paper_title":"Prior Normality Prompt Transformer for Multi-class Industrial Image Anomaly Detection","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":9,"model":"CRAS","metrics":{"Detection AUROC":"98.3","Segmentation AUROC":"98.0"},"uses_additional_data":false,"paper_date":"2025-05-23","paper":"/paper/center-aware-residual-anomaly-synthesis-for","paper_url":"https://arxiv.org/abs/2505.17551v1","paper_title":"Center-aware Residual Anomaly Synthesis for Multi-class Industrial Anomaly Detection","code":"https://github.com/cqylunlun/CRAS","n_code_links":1,"syntology":null},{"rank_in_archive_order":10,"model":"HVQ-Trans","metrics":{"Detection AUROC":"98.0","Segmentation AUROC":"97.3"},"uses_additional_data":false,"paper_date":"2023-10-22","paper":"/paper/hierarchical-vector-quantized-transformer-for-1","paper_url":"https://arxiv.org/abs/2310.14228v1","paper_title":"Hierarchical Vector Quantized Transformer for Multi-class Unsupervised Anomaly Detection","code":"https://github.com/ruiyinglu/hvq-trans","n_code_links":1,"syntology":{"n_ran":2,"n_unverified":0,"n_samples":2,"n_pointer_only_licence":2}},{"rank_in_archive_order":11,"model":"OminAL","metrics":{"Detection AUROC":"97.2","Segmentation AUROC":"98.3"},"uses_additional_data":false,"paper_date":"2023-01-01","paper":"/paper/omnial-a-unified-cnn-framework-for","paper_url":"http://openaccess.thecvf.com//content/CVPR2023/html/Zhao_OmniAL_A_Unified_CNN_Framework_for_Unsupervised_Anomaly_Localization_CVPR_2023_paper.html","paper_title":"OmniAL: A Unified CNN Framework for Unsupervised Anomaly Localization","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":12,"model":"DiAD","metrics":{"Detection AUROC":"97.2","Segmentation AUROC":"96.8"},"uses_additional_data":false,"paper_date":"2023-12-11","paper":"/paper/diad-a-diffusion-based-framework-for-multi","paper_url":"https://arxiv.org/abs/2312.06607v1","paper_title":"DiAD: A Diffusion-based Framework for Multi-class Anomaly Detection","code":"https://github.com/lewandofskee/DiAD","n_code_links":1,"syntology":{"n_ran":8,"n_unverified":3,"n_samples":11,"n_pointer_only_licence":0}},{"rank_in_archive_order":13,"model":"UniAD","metrics":{"Detection AUROC":"96.5","Segmentation AUROC":"96.8"},"uses_additional_data":false,"paper_date":"2022-06-08","paper":"/paper/a-unified-model-for-multi-class-anomaly","paper_url":"https://arxiv.org/abs/2206.03687v3","paper_title":"A Unified Model for Multi-class Anomaly Detection","code":"https://github.com/zhiyuanyou/uniad","n_code_links":1,"syntology":{"n_ran":8,"n_unverified":7,"n_samples":15,"n_pointer_only_licence":0}}],"since_archive":{"claim":"Results that newer papers report for their own method, placed here by Syntology. A model pointed at the cell in the paper's own table; the number was read from that cell and checked against this leaderboard's metric, dataset, split and scale; an independent check that saw this leaderboard's other rows and every other leaderboard on the same dataset accepted it. Not reviewed by the paper's authors or by the archive's editors, and not ranked against the archive rows.","extraction_file_present":true,"measurement":{"test_papers":883,"papers_with_output":881,"judged_true":108,"judged":110,"wilson95_lower":0.9361,"measured_on":"2026-09-24","frozen_commit":"0e3de0df94"},"measurement_note":"blind adjudication of accepted entries on a held-out split of archive papers, rules frozen before the test","coverage":{"sentence":"Syntology has checked 6,885 of the 9,623 papers on this site that are newer than the archive; results from the others appear after they are checked.","complete":false,"papers_newer_than_archive":9623,"papers_checked":6885,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":2737},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-25T09:33:49+00:00","claim":"Per row: N of M harvested code samples from that row's paper executed on a synthesized fixture; the other M-N are unverified. Not a reproduction of the row's number; not a correctness claim. n_pointer_only_licence counts samples the site points at rather than redistributes (a licence axis, independent of ran/unverified).","rows_with_graph_line":10,"rows_with_any_sample_ran":10,"distinct_papers_with_graph_line":8,"distinct_papers_with_any_sample_ran":8,"samples_over_distinct_papers":{"n_ran":41,"n_unverified":15,"n_samples":56,"n_pointer_only_licence":15,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":54,"n_unverified":17,"n_samples":71,"n_pointer_only_licence":15,"note":"row-weighted: a paper behind several rows is counted once per row; inflated relative to samples_over_distinct_papers by design, kept for readers summing the per-row syntology blocks"}}}