{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,31]],"date-time":"2026-10-31T21:32:06Z","timestamp":1793482326337,"version":"4.5.0"},"reference-count":74,"publisher":"Association for Computing Machinery (ACM)","issue":"2","license":[{"start":{"date-parts":[[2025,1,9]],"date-time":"2025-01-09T00:00:00Z","timestamp":1736380800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/www.acm.org\/publications\/policies\/copyright_policy#Background"}],"funder":[{"DOI":"10.13039\/501100001809","name":"NSFC","doi-asserted-by":"crossref","award":["62206052, 62125602, 62076063"],"award-info":[{"award-number":["62206052, 62125602, 62076063"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/501100002858","name":"China Postdoctoral Science Foundation","doi-asserted-by":"crossref","award":["2024M750424"],"award-info":[{"award-number":["2024M750424"]}],"id":[{"id":"10.13039\/501100002858","id-type":"DOI","asserted-by":"crossref"}]},{"name":"Postdoctoral Fellowship Program of CPSF","award":["GZC20240252"],"award-info":[{"award-number":["GZC20240252"]}]},{"name":"Jiangsu Funding Program for Excellent Postdoctoral Talent","award":["2024ZB242"],"award-info":[{"award-number":["2024ZB242"]}]}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["ACM Trans. Multimedia Comput. Commun. Appl."],"published-print":{"date-parts":[[2025,2,28]]},"abstract":"<jats:p>\n            In industrial anomaly detection, model efficiency and mobile-friendliness become the primary concerns in real-world applications. Simultaneously, the impressive generalization capabilities of Segment Anything (SAM) have garnered broad academic attention, making it an ideal choice for localizing unseen anomalies and diverse real-world patterns. In this paper, considering these two critical factors, we propose a SAM-guided Two-stream Lightweight Model for unsupervised anomaly detection (STLM) that not only aligns with the two practical application requirements but also harnesses the robust generalization capabilities of SAM. We employ two lightweight image encoders, i.e., our two-stream lightweight module, guided by SAM\u2019s knowledge. To be specific, one stream is trained to generate discriminative and general feature representations in both normal and anomalous regions, while the other stream reconstructs the same images without anomalies, which effectively enhances the differentiation of two-stream representations when facing anomalous regions. Furthermore, we employ a shared mask decoder and a feature aggregation module to generate anomaly maps. Our experiments conducted on MVTec AD benchmark show that STLM, with about 16M parameters and achieving an inference time in 20ms, competes effectively with state-of-the-art methods in terms of performance, 98.26% on pixel-level AUC and 94.92% on PRO. We further experiment on more difficult datasets, e.g., VisA and DAGM, to demonstrate the effectiveness and generalizability of STLM. Codes are available online at\n            <jats:ext-link xmlns:xlink=\"http:\/\/www.w3.org\/1999\/xlink\" ext-link-type=\"uri\" xlink:href=\"https:\/\/github.com\/Qi5Lei\/STLM\">https:\/\/github.com\/Qi5Lei\/STLM<\/jats:ext-link>\n            .\n          <\/jats:p>","DOI":"10.1145\/3706574","type":"journal-article","created":{"date-parts":[[2024,11,29]],"date-time":"2024-11-29T15:24:28Z","timestamp":1732893868000},"page":"1-23","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":21,"title":["A SAM-guided Two-stream Lightweight Model for Anomaly Detection"],"prefix":"10.1145","volume":"21","author":[{"ORCID":"https:\/\/orcid.org\/0009-0003-6130-1210","authenticated-orcid":false,"given":"Chenghao","family":"Li","sequence":"first","affiliation":[{"name":"Southeast University, Nanjing, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7091-0702","authenticated-orcid":false,"given":"Lei","family":"Qi","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering, Southeast University, Nanjing, China, and Key Laboratory of New Generation Artificial Intelligence Technology and Its Interdisciplinary Applications, Southeast University, Ministry of Education, Nanjing, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7729-0622","authenticated-orcid":false,"given":"Xin","family":"Geng","sequence":"additional","affiliation":[{"name":"Southeast University, Nanjing, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"320","published-online":{"date-parts":[[2025,1,9]]},"reference":[{"key":"e_1_3_1_2_2","article-title":"Statistics notes: Diagnostic tests 2: Predictive values","volume":"309","author":"Altman Douglas G.","year":"1994","unstructured":"Douglas G. Altman and J. Martin Bland. 1994. Statistics notes: Diagnostic tests 2: Predictive values. British Medical Journal 309 (1994), 102.","journal-title":"British Medical Journal"},{"key":"e_1_3_1_3_2","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-022-01578-9"},{"key":"e_1_3_1_4_2","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr.2019.00982"},{"key":"e_1_3_1_5_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.00424"},{"key":"e_1_3_1_6_2","doi-asserted-by":"publisher","DOI":"10.5220\/0007364503720380"},{"key":"e_1_3_1_7_2","unstructured":"Rishi Bommasani Drew A. Hudson Ehsan Adeli Russ Altman Simran Arora Sydney von Arx Michael S. Bernstein Jeannette Bohg Antoine Bosselut Emma Brunskill et al. 2021. On the opportunities and risks of foundation models. arXiv:2108.07258. Retrieved from https:\/\/arxiv.org\/abs\/2108.07258"},{"key":"e_1_3_1_8_2","first-page":"1877","article-title":"Language models are few-shot learners","volume":"33","author":"Brown Tom","year":"2020","unstructured":"Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. 2020. Language models are few-shot learners. Advances in Neural Information Processing Systems 33 (2020), 1877\u20131901.","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_3_1_9_2","doi-asserted-by":"publisher","DOI":"10.1007\/s13042-024-02355-5"},{"key":"e_1_3_1_10_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00951"},{"key":"e_1_3_1_11_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.engappai.2024.107886"},{"key":"e_1_3_1_12_2","doi-asserted-by":"publisher","DOI":"10.1145\/3574131.3574463"},{"key":"e_1_3_1_13_2","doi-asserted-by":"publisher","DOI":"10.1145\/3394171.3413813"},{"key":"e_1_3_1_14_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.00951"},{"key":"e_1_3_1_15_2","volume-title":"Proceedings of the International Conference on Learning Representations","author":"Dosovitskiy Alexey","year":"2021","unstructured":"Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, and Neil Houlsby. 2021. An image is worth 16x16 words: Transformers for image recognition at scale. In Proceedings of the International Conference on Learning Representations. Retrieved from https:\/\/openreview.net\/forum?id=YicbFdNTTy"},{"key":"e_1_3_1_16_2","doi-asserted-by":"publisher","DOI":"10.1145\/3464423"},{"key":"e_1_3_1_17_2","doi-asserted-by":"publisher","DOI":"10.1007\/s11704-021-1237-4"},{"key":"e_1_3_1_18_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2019.00179"},{"key":"e_1_3_1_19_2","doi-asserted-by":"publisher","DOI":"10.3156\/JSOFT.29.5_177_2"},{"key":"e_1_3_1_20_2","doi-asserted-by":"publisher","DOI":"10.1109\/WACV51458.2022.00188"},{"key":"e_1_3_1_21_2","doi-asserted-by":"publisher","DOI":"10.1145\/3131342"},{"key":"e_1_3_1_22_2","first-page":"0","volume-title":"Proceedings of the European Conference on Computer Vision","author":"Hinterstoisser Stefan","year":"2018","unstructured":"Stefan Hinterstoisser, Vincent Lepetit, Paul Wohlhart, and Kurt Konolige. 2018. On Pre-trained image features and synthetic images for deep learning. In Proceedings of the European Conference on Computer Vision. Springer-Verlag, 0\u20130."},{"key":"e_1_3_1_23_2","unstructured":"Geoffrey Hinton Oriol Vinyals and Jeff Dean. 2015. Distilling the knowledge in a neural network. arXiv:1503.02531. Retrieved from https:\/\/arxiv.org\/abs\/1503.02531"},{"key":"e_1_3_1_24_2","first-page":"2790","article-title":"Parameter-efficient transfer learning for NLP","author":"Houlsby Neil","year":"2019","unstructured":"Neil Houlsby, Andrei Giurgiu, Stanislaw Jastrzebski, Bruna Morrone, Quentin De Laroussilhe, Andrea Gesmundo, Mona Attariyan, and Sylvain Gelly. 2019. Parameter-efficient transfer learning for NLP. In International conference on machine learning. PMLR, 2790\u20132799.","journal-title":"International conference on machine learning"},{"key":"e_1_3_1_25_2","unstructured":"Andrew G Howard Menglong Zhu Bo Chen Dmitry Kalenichenko Weijun Wang Tobias Weyand Marco Andreetto and Hartwig Adam. 2017. Mobilenets: Efficient convolutional neural networks for Mobile vision applications. arXiv:1704.04861. Retrieved from https:\/\/arxiv.org\/abs\/1704.04861"},{"key":"e_1_3_1_26_2","volume-title":"Proceedings of the International Conference on Learning Representations","author":"Hu J. Edward","year":"2022","unstructured":"J. Edward Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, and Weizhu Chen. 2022. LoRA: Low-rank adaptation of large language models. In Proceedings of the International Conference on Learning Representations. ICLR Press."},{"key":"e_1_3_1_27_2","unstructured":"Wei Ji Jingjing Li Qi Bi Wenbo Li and Li Cheng. 2023. Segment anything is not always perfect: An investigation of Sam on different Real-world applications. arXiv:2304.05750. Retrieved from https:\/\/arxiv.org\/abs\/2304.05750"},{"key":"e_1_3_1_28_2","first-page":"75434","article-title":"SANFlow: Semantic-aware normalizing flow for anomaly detection","volume":"36","author":"Kim Daehyun","year":"2024","unstructured":"Daehyun Kim, Sungyong Baik, and Tae Hyun Kim. 2024. SANFlow: Semantic-aware normalizing flow for anomaly detection. Advances in Neural Information Processing Systems 36 (2024), 75434\u201375454.","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_3_1_29_2","doi-asserted-by":"crossref","unstructured":"Alexander Kirillov Eric Mintun Nikhila Ravi Hanzi Mao Chloe Rolland Laura Gustafson Tete Xiao Spencer Whitehead Alexander C. Berg Wan-Yen Lo et al. 2023. Segment anything. arXiv:2304.02643. Retrieved from https:\/\/arxiv.org\/abs\/2304.02643","DOI":"10.1109\/ICCV51070.2023.00371"},{"key":"e_1_3_1_30_2","doi-asserted-by":"publisher","DOI":"10.1109\/access.2022.3193699"},{"key":"e_1_3_1_31_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR46437.2021.00954"},{"key":"e_1_3_1_32_2","doi-asserted-by":"publisher","DOI":"10.1162\/dint_a_00232"},{"key":"e_1_3_1_33_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.01549"},{"key":"e_1_3_1_34_2","doi-asserted-by":"publisher","DOI":"10.1145\/3431768"},{"key":"e_1_3_1_35_2","doi-asserted-by":"publisher","DOI":"10.1162\/dint_a_00144"},{"key":"e_1_3_1_36_2","doi-asserted-by":"publisher","DOI":"10.1145\/3343031.3350998"},{"key":"e_1_3_1_37_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.00867"},{"key":"e_1_3_1_38_2","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v35i10.17054"},{"key":"e_1_3_1_39_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2019.00339"},{"key":"e_1_3_1_40_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.01954"},{"key":"e_1_3_1_41_2","first-page":"1","article-title":"Dinov2: Learning robust visual features without supervision","author":"Oquab Maxime","year":"2024","unstructured":"Maxime Oquab, Timoth\u00e9e Darcet, Th\u00e9o Moutakanni, Huy Vo, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, et al. 2024. Dinov2: Learning robust visual features without supervision. Transactions on Machine Learning Research (2024), 1\u201331.","journal-title":"Transactions on Machine Learning Research"},{"key":"e_1_3_1_42_2","doi-asserted-by":"publisher","DOI":"10.1145\/3284750"},{"key":"e_1_3_1_43_2","first-page":"606","volume-title":"Proceedings of Machine Learning and Systems","volume":"5","author":"Pope Reiner","year":"2023","unstructured":"Reiner Pope, Sholto Douglas, Aakanksha Chowdhery, Jacob Devlin, James Bradbury, Jonathan Heek, Kefan Xiao, Shivani Agrawal, and Jeff Dean. 2023. Efficiently scaling transformer inference. Proceedings of Machine Learning and Systems 5 (2023), 606\u2013624."},{"key":"e_1_3_1_44_2","first-page":"8748","volume-title":"Proceedings of the International Conference on Machine Learning","author":"Radford Alec","year":"2021","unstructured":"Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. 2021. Learning transferable visual models from natural language supervision. In Proceedings of the International Conference on Machine Learning. PMLR, 8748\u20138763."},{"key":"e_1_3_1_45_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01392"},{"key":"e_1_3_1_46_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR46437.2021.01466"},{"key":"e_1_3_1_47_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-19821-2_27"},{"key":"e_1_3_1_48_2","doi-asserted-by":"publisher","DOI":"10.1007\/s11704-022-2134-1"},{"key":"e_1_3_1_49_2","doi-asserted-by":"publisher","DOI":"10.1109\/tim.2022.3196436"},{"key":"e_1_3_1_50_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.02348"},{"key":"e_1_3_1_51_2","unstructured":"Guodong Wang Shumin Han Errui Ding and Di Huang. 2021. Student-teacher feature Pyramid matching for anomaly detection. arXiv:2103.04257. Retrieved from https:\/\/arxiv.org\/abs\/2103.04257"},{"key":"e_1_3_1_52_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.00110"},{"key":"e_1_3_1_53_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCSVT.2023.3344099"},{"key":"e_1_3_1_54_2","article-title":"Weakly supervised learning for industrial optical inspection","author":"Wieler Matthias","year":"2007","unstructured":"Matthias Wieler and Tobias Hahn. 2007. Weakly supervised learning for industrial optical inspection. In Proceedings of the DAGM Symposium.","journal-title":"Proceedings of the DAGM Symposium"},{"key":"e_1_3_1_55_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-19803-8_5"},{"key":"e_1_3_1_56_2","doi-asserted-by":"publisher","DOI":"10.1145\/3614435"},{"key":"e_1_3_1_57_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.00626"},{"key":"e_1_3_1_58_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.02346"},{"key":"e_1_3_1_59_2","volume-title":"Proceedings of the International Conference on Machine Learning","author":"Yao Zhewei","year":"2021","unstructured":"Zhewei Yao, Zhen Dong, Zhangcheng Zheng, Amir Gholami, Jiali Yu, Eric Tan, Leyuan Wang, Qijing Huang, Yida Wang, Michael W. Mahoney, and Kurt Keutzer. 2021. HAWQ-V3: Dyadic neural network quantization. In Proceedings of the International Conference on Machine Learning. PMLR."},{"key":"e_1_3_1_60_2","unstructured":"Jiawei Yu Ye Zheng Xiang Wang Wei Li Yushuang Wu Rui Zhao and Liwei Wu. 2021. Fastflow: Unsupervised anomaly detection and localization via 2d normalizing flows. arXiv:2111.07677. Retrieved from https:\/\/arxiv.org\/abs\/2111.07677"},{"key":"e_1_3_1_61_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00822"},{"key":"e_1_3_1_62_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.patcog.2020.107706"},{"key":"e_1_3_1_63_2","unstructured":"Chaoning Zhang Dongshen Han Yu Qiao Jung Uk Kim Sung-Ho Bae Seungkyu Lee and Choong Seon Hong. 2023. Faster segment anything: Towards lightweight SAM for Mobile applications. arXiv:2306.14289. Retrieved from https:\/\/arxiv.org\/abs\/2306.14289"},{"key":"e_1_3_1_64_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.01562"},{"key":"e_1_3_1_65_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.00381"},{"key":"e_1_3_1_66_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00716"},{"key":"e_1_3_1_67_2","doi-asserted-by":"crossref","unstructured":"Zhengbin Zhang Zhenhao Xu Xingsheng Gu and Juan Xiong. 2023d. Cross-CBAM: A lightweight network for scene segmentation. arXiv:2306.02306. Retrieved from https:\/\/arxiv.org\/abs\/2306.02306","DOI":"10.2139\/ssrn.4370101"},{"key":"e_1_3_1_68_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01165"},{"key":"e_1_3_1_69_2","doi-asserted-by":"publisher","DOI":"10.1162\/dint_a_00146"},{"key":"e_1_3_1_70_2","doi-asserted-by":"publisher","DOI":"10.1145\/3097983.3098052"},{"key":"e_1_3_1_71_2","volume-title":"Proceedings of the International Conference on Learning Representations","author":"Zhou Jinghao","year":"2022","unstructured":"Jinghao Zhou, Chen Wei, Huiyu Wang, Wei Shen, Cihang Xie, Alan Yuille, and Tao Kong. 2022. Image BERT Pre-training with online tokenizer. In Proceedings of the International Conference on Learning Representations. Retrieved from https:\/\/openreview.net\/forum?id=ydopy-e6Dg"},{"key":"e_1_3_1_72_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.01075"},{"key":"e_1_3_1_73_2","doi-asserted-by":"publisher","DOI":"10.1007\/s11704-022-1349-5"},{"key":"e_1_3_1_74_2","first-page":"19769","article-title":"Segment everything everywhere all at Once","volume":"36","author":"Zou Xueyan","year":"2024","unstructured":"Xueyan Zou, Jianwei Yang, Hao Zhang, Feng Li, Linjie Li, Jianfeng Wang, Lijuan Wang, Jianfeng Gao, and Yong Jae Lee. 2024. Segment everything everywhere all at Once. Advances in Neural Information Processing Systems 36 (2024), 19769\u201319782.","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_3_1_75_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-20056-4_23"}],"container-title":["ACM Transactions on Multimedia Computing, Communications, and Applications"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3706574","content-type":"unspecified","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3706574","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,19]],"date-time":"2025-06-19T01:56:51Z","timestamp":1750298211000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3706574"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,1,9]]},"references-count":74,"journal-issue":{"issue":"2","published-print":{"date-parts":[[2025,2,28]]}},"alternative-id":["10.1145\/3706574"],"URL":"https:\/\/doi.org\/10.1145\/3706574","relation":{},"ISSN":["1551-6857","1551-6865"],"issn-type":[{"value":"1551-6857","type":"print"},{"value":"1551-6865","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,1,9]]},"assertion":[{"value":"2024-07-03","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2024-11-18","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-01-09","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}