@article{
 title = {Student-Teacher Feature Pyramid Matching for Anomaly Detection},
 type = {article},
 year = {2021},
 websites = {https://arxiv.org/abs/2103.04257v3},
 month = {3},
 day = {7},
 id = {7139b18e-352a-382f-8928-c0230482818a},
 created = {2023-06-12T12:11:28.927Z},
 accessed = {2023-06-12},
 file_attached = {true},
 profile_id = {f1f70cad-e32d-3de2-a3c0-be1736cb88be},
 group_id = {5ec9cc91-a5d6-3de5-82f3-3ef3d98a89c1},
 last_modified = {2023-06-13T05:14:56.274Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {false},
 hidden = {false},
 folder_uuids = {5a010301-acb6-4642-a6b2-8afaee1b741c,4b11b59b-846d-45b5-82b7-81d8bc3108b2,bd3c6f2e-3514-47cf-bc42-12db8b9abe45},
 private_publication = {false},
 abstract = {Anomaly detection is a challenging task and usually formulated as an
one-class learning problem for the unexpectedness of anomalies. This paper
proposes a simple yet powerful approach to this issue, which is implemented in
the student-teacher framework for its advantages but substantially extends it
in terms of both accuracy and efficiency. Given a strong model pre-trained on
image classification as the teacher, we distill the knowledge into a single
student network with the identical architecture to learn the distribution of
anomaly-free images and this one-step transfer preserves the crucial clues as
much as possible. Moreover, we integrate the multi-scale feature matching
strategy into the framework, and this hierarchical feature matching enables the
student network to receive a mixture of multi-level knowledge from the feature
pyramid under better supervision, thus allowing to detect anomalies of various
sizes. The difference between feature pyramids generated by the two networks
serves as a scoring function indicating the probability of anomaly occurring.
Due to such operations, our approach achieves accurate and fast pixel-level
anomaly detection. Very competitive results are delivered on the MVTec anomaly
detection dataset, superior to the state of the art ones.},
 bibtype = {article},
 author = {Wang, Guodong and Han, Shumin and Ding, Errui and Huang, Di}
}

{"_id":"uMH2ittxKEA4QM4hf","bibbaseid":"wang-han-ding-huang-studentteacherfeaturepyramidmatchingforanomalydetection-2021","author_short":["Wang, G.","Han, S.","Ding, E.","Huang, D."],"bibdata":{"title":"Student-Teacher Feature Pyramid Matching for Anomaly Detection","type":"article","year":"2021","websites":"https://arxiv.org/abs/2103.04257v3","month":"3","day":"7","id":"7139b18e-352a-382f-8928-c0230482818a","created":"2023-06-12T12:11:28.927Z","accessed":"2023-06-12","file_attached":"true","profile_id":"f1f70cad-e32d-3de2-a3c0-be1736cb88be","group_id":"5ec9cc91-a5d6-3de5-82f3-3ef3d98a89c1","last_modified":"2023-06-13T05:14:56.274Z","read":false,"starred":false,"authored":false,"confirmed":false,"hidden":false,"folder_uuids":"5a010301-acb6-4642-a6b2-8afaee1b741c,4b11b59b-846d-45b5-82b7-81d8bc3108b2,bd3c6f2e-3514-47cf-bc42-12db8b9abe45","private_publication":false,"abstract":"Anomaly detection is a challenging task and usually formulated as an\none-class learning problem for the unexpectedness of anomalies. This paper\nproposes a simple yet powerful approach to this issue, which is implemented in\nthe student-teacher framework for its advantages but substantially extends it\nin terms of both accuracy and efficiency. Given a strong model pre-trained on\nimage classification as the teacher, we distill the knowledge into a single\nstudent network with the identical architecture to learn the distribution of\nanomaly-free images and this one-step transfer preserves the crucial clues as\nmuch as possible. Moreover, we integrate the multi-scale feature matching\nstrategy into the framework, and this hierarchical feature matching enables the\nstudent network to receive a mixture of multi-level knowledge from the feature\npyramid under better supervision, thus allowing to detect anomalies of various\nsizes. The difference between feature pyramids generated by the two networks\nserves as a scoring function indicating the probability of anomaly occurring.\nDue to such operations, our approach achieves accurate and fast pixel-level\nanomaly detection. Very competitive results are delivered on the MVTec anomaly\ndetection dataset, superior to the state of the art ones.","bibtype":"article","author":"Wang, Guodong and Han, Shumin and Ding, Errui and Huang, Di","bibtex":"@article{\n title = {Student-Teacher Feature Pyramid Matching for Anomaly Detection},\n type = {article},\n year = {2021},\n websites = {https://arxiv.org/abs/2103.04257v3},\n month = {3},\n day = {7},\n id = {7139b18e-352a-382f-8928-c0230482818a},\n created = {2023-06-12T12:11:28.927Z},\n accessed = {2023-06-12},\n file_attached = {true},\n profile_id = {f1f70cad-e32d-3de2-a3c0-be1736cb88be},\n group_id = {5ec9cc91-a5d6-3de5-82f3-3ef3d98a89c1},\n last_modified = {2023-06-13T05:14:56.274Z},\n read = {false},\n starred = {false},\n authored = {false},\n confirmed = {false},\n hidden = {false},\n folder_uuids = {5a010301-acb6-4642-a6b2-8afaee1b741c,4b11b59b-846d-45b5-82b7-81d8bc3108b2,bd3c6f2e-3514-47cf-bc42-12db8b9abe45},\n private_publication = {false},\n abstract = {Anomaly detection is a challenging task and usually formulated as an\none-class learning problem for the unexpectedness of anomalies. This paper\nproposes a simple yet powerful approach to this issue, which is implemented in\nthe student-teacher framework for its advantages but substantially extends it\nin terms of both accuracy and efficiency. Given a strong model pre-trained on\nimage classification as the teacher, we distill the knowledge into a single\nstudent network with the identical architecture to learn the distribution of\nanomaly-free images and this one-step transfer preserves the crucial clues as\nmuch as possible. Moreover, we integrate the multi-scale feature matching\nstrategy into the framework, and this hierarchical feature matching enables the\nstudent network to receive a mixture of multi-level knowledge from the feature\npyramid under better supervision, thus allowing to detect anomalies of various\nsizes. The difference between feature pyramids generated by the two networks\nserves as a scoring function indicating the probability of anomaly occurring.\nDue to such operations, our approach achieves accurate and fast pixel-level\nanomaly detection. Very competitive results are delivered on the MVTec anomaly\ndetection dataset, superior to the state of the art ones.},\n bibtype = {article},\n author = {Wang, Guodong and Han, Shumin and Ding, Errui and Huang, Di}\n}","author_short":["Wang, G.","Han, S.","Ding, E.","Huang, D."],"urls":{"Paper":"https://bibbase.org/service/mendeley/bfbbf840-4c42-3914-a463-19024f50b30c/file/388871ac-d212-ddeb-1fde-ef0b25d9770e/full_text.pdf.pdf","Website":"https://arxiv.org/abs/2103.04257v3"},"biburl":"https://bibbase.org/service/mendeley/bfbbf840-4c42-3914-a463-19024f50b30c","bibbaseid":"wang-han-ding-huang-studentteacherfeaturepyramidmatchingforanomalydetection-2021","role":"author","metadata":{"authorlinks":{}},"downloads":0},"bibtype":"article","biburl":"https://bibbase.org/service/mendeley/bfbbf840-4c42-3914-a463-19024f50b30c","dataSources":["2252seNhipfTmjEBQ"],"keywords":[],"search_terms":["student","teacher","feature","pyramid","matching","anomaly","detection","wang","han","ding","huang"],"title":"Student-Teacher Feature Pyramid Matching for Anomaly Detection","year":2021}