{"about":{"non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","site":"https://codewithpapers.app","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page","syntology":{"site":"https://syntology.ai","developers":"https://syntology.ai/developers","mcp":{"server":"https://syntology.ai/mcp","transport":"streamable-http","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","auth":{"type":"trial token, no account","trial_token":"https://syntology.ai/api/oauth/trial/token","method":"POST","docs":"https://syntology.ai/developers"}},"have":"https://syntology.ai/api/graph/have?x=<method, arXiv id or title> (free, answers coverage only)","paper_base":"https://syntology.ai/paper/","atlas_base":"https://app.syntology.ai/?focus="},"machine_readable":[{"url":"https://codewithpapers.app/llms.txt","what":"the machine catalog: every machine-readable file, counted"},{"url":"https://codewithpapers.app/index/manifest.json","what":"paper-to-code index by arXiv id, with Syntology's counts"},{"url":"https://codewithpapers.app/search/manifest.json","what":"site search index (titles, authors) and its files"},{"url":"https://codewithpapers.app/download","what":"bulk files: Syntology's layer, described there"},{"url":"https://codewithpapers.app/build_manifest.json","what":"the build record: inputs, counts, exclusions, probes"}]},"url":"/paper/panda-adapting-pretrained-features-for","title":"PANDA: Adapting Pretrained Features for Anomaly Detection and Segmentation","arxiv_id":"2010.05903","date":"2020-10-12","proceeding":"CVPR 2021 1","authors":["Tal Reiss","Niv Cohen","Liron Bergman","Yedid Hoshen"],"abstract":"Anomaly detection methods require high-quality features. In recent years, the anomaly detection community has attempted to obtain better features using advances in deep self-supervised feature learning. Surprisingly, a very promising direction, using pretrained deep features, has been mostly overlooked. In this paper, we first empirically establish the perhaps expected, but unreported result, that combining pretrained features with simple anomaly detection and segmentation methods convincingly outperforms, much more complex, state-of-the-art methods. In order to obtain further performance gains in anomaly detection, we adapt pretrained features to the target distribution. Although transfer learning methods are well established in multi-class classification problems, the one-class classification (OCC) setting is not as well explored. It turns out that naive adaptation methods, which typically work well in supervised learning, often result in catastrophic collapse (feature deterioration) and reduce performance in OCC settings. A popular OCC method, DeepSVDD, advocates using specialized architectures, but this limits the adaptation performance gain. We propose two methods for combating collapse: i) a variant of early stopping that dynamically learns the stopping iteration ii) elastic regularization inspired by continual learning. Our method, PANDA, outperforms the state-of-the-art in the OCC, outlier exposure and anomaly segmentation settings by large margins.","url_abs":"https://arxiv.org/abs/2010.05903v3","url_pdf":"https://arxiv.org/pdf/2010.05903v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"panda-adapting-pretrained-features-for","repo_url":"https://github.com/talreiss/PANDA","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}}],"tasks":[{"task_slug":"anomaly-detection","task_name":"Anomaly Detection"},{"task_slug":"anomaly-segmentation","task_name":"Anomaly Segmentation"},{"task_slug":"continual-learning","task_name":"Continual Learning"},{"task_slug":"multi-class-classification","task_name":"Multi-class Classification"},{"task_slug":"one-class-classification","task_name":"One-Class Classification"},{"task_slug":"transfer-learning","task_name":"Transfer Learning"}],"methods":[{"method_slug":"early-stopping","method_name":"Early Stopping"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/anomaly-detection-on-cats-and-dogs","task":"Anomaly Detection","dataset":"Cats-and-Dogs","model":"PANDA","rank_in_archive_order":1,"of":4,"metrics":{"ROC AUC":"97.3"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-cats-and-dogs","task":"Anomaly Detection","dataset":"Cats-and-Dogs","model":"PANDA-OE","rank_in_archive_order":2,"of":4,"metrics":{"ROC AUC":"94.5"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-cats-and-dogs","task":"Anomaly Detection","dataset":"Cats-and-Dogs","model":"Self-Supervised One-class SVM, RBF kernel","rank_in_archive_order":3,"of":4,"metrics":{"ROC AUC":"51.7"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-cats-and-dogs","task":"Anomaly Detection","dataset":"Cats-and-Dogs","model":"Self-Supervised DeepSVDD","rank_in_archive_order":4,"of":4,"metrics":{"ROC AUC":"50.5"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-dior","task":"Anomaly Detection","dataset":"DIOR","model":"PANDA-OE","rank_in_archive_order":1,"of":4,"metrics":{"ROC AUC":"95.9"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-dior","task":"Anomaly Detection","dataset":"DIOR","model":"PANDA","rank_in_archive_order":2,"of":4,"metrics":{"ROC AUC":"94.3"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-dior","task":"Anomaly Detection","dataset":"DIOR","model":"Self-Supervised One-class SVM, RBF kernel","rank_in_archive_order":3,"of":4,"metrics":{"ROC AUC":"70.7"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-dior","task":"Anomaly Detection","dataset":"DIOR","model":"Self-Supervised DeepSVDD","rank_in_archive_order":4,"of":4,"metrics":{"ROC AUC":"70"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-fashion-mnist","task":"Anomaly Detection","dataset":"Fashion-MNIST","model":"PANDA","rank_in_archive_order":2,"of":12,"metrics":{"ROC AUC":"95.6"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-fashion-mnist","task":"Anomaly Detection","dataset":"Fashion-MNIST","model":"Self-Supervised One-class SVM, RBF kernel","rank_in_archive_order":6,"of":12,"metrics":{"ROC AUC":"92.8"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-fashion-mnist","task":"Anomaly Detection","dataset":"Fashion-MNIST","model":"PANDA-OE","rank_in_archive_order":10,"of":12,"metrics":{"ROC AUC":"91.8"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-fashion-mnist","task":"Anomaly Detection","dataset":"Fashion-MNIST","model":"Self-Supervised DeepSVDD","rank_in_archive_order":11,"of":12,"metrics":{"ROC AUC":"84.8"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-hyper-kvasir-dataset","task":"Anomaly Detection","dataset":"Hyper-Kvasir Dataset","model":"PANDA","rank_in_archive_order":3,"of":6,"metrics":{"AUC":"0.937"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-one-class-cifar-10","task":"Anomaly Detection","dataset":"One-class CIFAR-10","model":"PANDA-OE","rank_in_archive_order":5,"of":36,"metrics":{"AUROC":"98.9"},"uses_additional_data":true},{"leaderboard":"/sota/anomaly-detection-on-one-class-cifar-10","task":"Anomaly Detection","dataset":"One-class CIFAR-10","model":"PANDA","rank_in_archive_order":11,"of":36,"metrics":{"AUROC":"96.2"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-one-class-cifar-10","task":"Anomaly Detection","dataset":"One-class CIFAR-10","model":"Self-Supervised DeepSVDD","rank_in_archive_order":35,"of":36,"metrics":{"AUROC":"64.8"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-one-class-cifar-10","task":"Anomaly Detection","dataset":"One-class CIFAR-10","model":"Self-Supervised One-class SVM, RBF kernel","rank_in_archive_order":36,"of":36,"metrics":{"AUROC":"64.7"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-one-class-cifar-100","task":"Anomaly Detection","dataset":"One-class CIFAR-100","model":"PANDA-OE","rank_in_archive_order":3,"of":15,"metrics":{"AUROC":"97.3"},"uses_additional_data":true},{"leaderboard":"/sota/anomaly-detection-on-one-class-cifar-100","task":"Anomaly Detection","dataset":"One-class CIFAR-100","model":"PANDA","rank_in_archive_order":5,"of":15,"metrics":{"AUROC":"94.1"},"uses_additional_data":true},{"leaderboard":"/sota/anomaly-detection-on-one-class-cifar-100","task":"Anomaly Detection","dataset":"One-class CIFAR-100","model":"Self-Supervised Multi-Head RotNet","rank_in_archive_order":12,"of":15,"metrics":{"AUROC":"80.1"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-one-class-cifar-100","task":"Anomaly Detection","dataset":"One-class CIFAR-100","model":"Self-Supervised DeepSVDD","rank_in_archive_order":14,"of":15,"metrics":{"AUROC":"67"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-one-class-cifar-100","task":"Anomaly Detection","dataset":"One-class CIFAR-100","model":"Self-Supervised One-class SVM, RBF kernel","rank_in_archive_order":15,"of":15,"metrics":{"AUROC":"62.6"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2010.05903","atlas_url":"https://app.syntology.ai/?focus=2010.05903","mcp":null,"developers":"https://syntology.ai/developers","held":true,"reason":"arxiv_id_in_atlas","checked_against":{"atlas_ids":549449,"atlas_ids_sha256":"afd72399dafd9906e9062927f70655952385a3d75488558e1f0484e950ab09b7","read_at":"2026-10-05T01:46:21Z"},"agent_calls":[{"tool":"get_citation_path","arguments":{"paper_1":"2010.05903"},"arguments_in":{"paper_2":"another paper's arXiv id or title"},"call":"get_citation_path(paper_1=\"2010.05903\", paper_2=\"…\")"},{"tool":"get_concepts_for_paper","arguments":{"arxiv_id":"2010.05903"},"call":"get_concepts_for_paper(arxiv_id=\"2010.05903\")"}],"by_repo":[{"repo":"talreiss/PANDA","kind":"official","samples":0,"ran":0,"instrument":0,"unverified":0}],"by_repo_is":"one row per repository in the archive's code links, Syntology's graph links or the samples; samples 0 means no sample is linked to this paper, a statement about Syntology's coverage, not about the repository: with harvested_for_other_papers present, Syntology harvested the repository for that many other papers, and without it, Syntology harvested nothing from it; placed_by_identical_code counts samples whose link records no repository, placed here because identical code was harvested from this repository for another paper; ran counts executions on a synthesized input and instrument counts failures of Syntology's instrument, not of the code"},"code_links_note":null,"arxiv_metadata":null,"syntology_extracted_results":null,"record_sha256":"2d3d8b27ba35b69fe847ef75af35d1d28445a0906f7db63a65ff9bef7cca20d0","record_changed_at":"2026-09-28","record_changed_at_basis":"first_hashed"}