{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,9]],"date-time":"2026-05-09T00:13:59Z","timestamp":1778285639487,"version":"3.51.4"},"reference-count":57,"publisher":"Institution of Engineering and Technology (IET)","issue":"1","license":[{"start":{"date-parts":[[2025,7,16]],"date-time":"2025-07-16T00:00:00Z","timestamp":1752624000000},"content-version":"vor","delay-in-days":196,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2025,1,1]],"date-time":"2025-01-01T00:00:00Z","timestamp":1735689600000},"content-version":"tdm","delay-in-days":0,"URL":"http:\/\/doi.wiley.com\/10.1002\/tdm_license_1.1"}],"funder":[{"DOI":"10.13039\/501100001691","name":"Japan Society for the Promotion of Science","doi-asserted-by":"publisher","award":["JP23K19989"],"award-info":[{"award-number":["JP23K19989"]}],"id":[{"id":"10.13039\/501100001691","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001691","name":"Japan Society for the Promotion of Science","doi-asserted-by":"publisher","award":["JP24K20828"],"award-info":[{"award-number":["JP24K20828"]}],"id":[{"id":"10.13039\/501100001691","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["ietresearch.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["IET Image Processing"],"published-print":{"date-parts":[[2025,1]]},"abstract":"<jats:title>ABSTRACT<\/jats:title>\n                  <jats:p>Unsupervised anomaly localization aims to identify anomalous regions that deviate from normal sample patterns. Most recent methods perform feature matching or reconstruction for the target sample with pre\u2010trained deep neural networks. However, they still struggle to address challenging anomalies because the deep embeddings stored in the memory bank can be less powerful and informative. Specifically, prior methods often overly rely on the finite resources stored in the memory bank, which leads to low robustness to unseen targets. In this paper, we propose a novel subspace\u2010guided feature reconstruction framework to pursue adaptive feature approximation for anomaly localization. It first learns to construct low\u2010dimensional subspaces from the given nominal samples, and then learns to reconstruct the given deep target embedding by linearly combining the subspace basis vectors using the self\u2010expressive model. Our core is that, despite the limited resources in the memory bank, the out\u2010of\u2010bank features can be alternatively \u201cmimicked\u201d to adaptively model the target. Moreover, we propose a sampling method that leverages the sparsity of subspaces and allows the feature reconstruction to depend only on a small resource subset, contributing to less memory overhead. Extensive experiments on three benchmark datasets demonstrate that our approach generally achieves state\u2010of\u2010the\u2010art anomaly localization\u00a0performance.<\/jats:p>","DOI":"10.1049\/ipr2.70157","type":"journal-article","created":{"date-parts":[[2025,7,17]],"date-time":"2025-07-17T04:19:33Z","timestamp":1752725973000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":3,"title":["Subspace\u2010Guided Feature Reconstruction for Unsupervised Anomaly Localization"],"prefix":"10.1049","volume":"19","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-8148-1445","authenticated-orcid":false,"given":"Katsuya","family":"Hotta","sequence":"first","affiliation":[{"name":"Iwate University Morioka\u2010shi Iwate Japan"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0845-9217","authenticated-orcid":false,"given":"Chao","family":"Zhang","sequence":"additional","affiliation":[{"name":"University of Toyama Toyama\u2010shi Toyama Japan"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yoshihiro","family":"Hagihara","sequence":"additional","affiliation":[{"name":"Iwate University Morioka\u2010shi Iwate Japan"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Takuya","family":"Akashi","sequence":"additional","affiliation":[{"name":"Okayama University Kita\u2010ku Okayama Japan"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"265","published-online":{"date-parts":[[2025,7,16]]},"reference":[{"key":"e_1_2_10_2_1","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-022-01578-9"},{"key":"e_1_2_10_3_1","doi-asserted-by":"crossref","unstructured":"Y.Zou J.Jeong L.Pemula D.Zhang andO.Dabeer \u201cSpot\u2010the\u2010Difference Self\u2010Supervised Pre\u2010Training for Anomaly Detection and Segmentation \u201d inEuropean Conference on Computer Vision(Springer 2022) 392\u2013408.","DOI":"10.1007\/978-3-031-20056-4_23"},{"key":"e_1_2_10_4_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.imavis.2023.104817"},{"issue":"1","key":"e_1_2_10_5_1","first-page":"18","article-title":"Anomaly Detection and Localization in Crowded Scenes","volume":"36","author":"Li W.","year":"2013","journal-title":"IEEE Transactions on Pattern Analysis and Machine Intelligence"},{"key":"e_1_2_10_6_1","doi-asserted-by":"crossref","unstructured":"W.Liu W.Luo D.Lian andS.Gao \u201cFuture Frame Prediction for Anomaly Detection\u2013a New Baseline \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2018) 6536\u20136545.","DOI":"10.1109\/CVPR.2018.00684"},{"key":"e_1_2_10_7_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.imavis.2023.104629"},{"key":"e_1_2_10_8_1","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1109\/TIM.2022.3196436","article-title":"Deep Learning for Unsupervised Anomaly Localization in Industrial Images: A Survey","volume":"71","author":"Tao X.","year":"2022","journal-title":"IEEE Transactions on Instrumentation and Measurement"},{"key":"e_1_2_10_9_1","doi-asserted-by":"crossref","unstructured":"H.DengandX.Li \u201cAnomaly Detection via Reverse Distillation From One\u2010Class Embedding \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2022) 9737\u20139746.","DOI":"10.1109\/CVPR52688.2022.00951"},{"key":"e_1_2_10_10_1","doi-asserted-by":"crossref","unstructured":"Z.Liu Y.Zhou Y.Xu andZ.Wang \u201cSimpleNet: A Simple Network for Image Anomaly Detection and Localization \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2023) 20402\u201320411.","DOI":"10.1109\/CVPR52729.2023.01954"},{"key":"e_1_2_10_11_1","doi-asserted-by":"crossref","unstructured":"J.Deng W.Dong R.Socher L. J.Li K.Li andL.Fei\u2010Fei \u201cImagenet: A Large\u2010Scale Hierarchical Image Database \u201d in:IEEE Conference on Computer Vision and Pattern Recognition(IEEE 2009) 248\u2013255.","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"e_1_2_10_12_1","unstructured":"N.CohenandY.Hoshen \u201cSub\u2010Image Anomaly Detection With Deep Pyramid Correspondences \u201dCoRR: abs\/2005.02357(2020)."},{"key":"e_1_2_10_13_1","doi-asserted-by":"crossref","unstructured":"K.Roth L.Pemula J.Zepeda B.Sch\u00f6lkopf T.Brox andP.Gehler \u201cTowards Total Recall in Industrial Anomaly Detection \u201d inIEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)(IEEE 2022) 14298\u201314308.","DOI":"10.1109\/CVPR52688.2022.01392"},{"key":"e_1_2_10_14_1","doi-asserted-by":"publisher","DOI":"10.1049\/cvi2.12258"},{"key":"e_1_2_10_15_1","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2022.3193699"},{"key":"e_1_2_10_16_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-1-4615-0953-0_4"},{"issue":"1","key":"e_1_2_10_17_1","first-page":"1","article-title":"Geometric Approximation via Coresets","volume":"52","author":"Agarwal P. K.","year":"2005","journal-title":"Combinatorial and Computational Geometry"},{"key":"e_1_2_10_18_1","doi-asserted-by":"crossref","unstructured":"P.Bergmann S.L\u00f6we M.Fauser D.Sattlegger andC.Steger \u201cImproving Unsupervised Defect Segmentation by Applying Structural Similarity to Autoencoders \u201dCoRR:abs\/1807.02011(2018).","DOI":"10.5220\/0007364503720380"},{"key":"e_1_2_10_19_1","doi-asserted-by":"crossref","unstructured":"P.Bergmann M.Fauser D.Sattlegger andC.Steger \u201cMVTec AD\u2013A Comprehensive Real\u2010World Dataset for Unsupervised Anomaly Detection \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2019) 9592\u20139600.","DOI":"10.1109\/CVPR.2019.00982"},{"key":"e_1_2_10_20_1","unstructured":"J.Yang Y.Shi andZ.Qi \u201cDfr: Deep Feature Reconstruction for Unsupervised Anomaly Segmentation \u201dCoRR:abs\/2012.07122(2020)."},{"key":"e_1_2_10_21_1","doi-asserted-by":"publisher","DOI":"10.1049\/ipr2.12518"},{"key":"e_1_2_10_22_1","doi-asserted-by":"publisher","DOI":"10.1007\/s41095-022-0293-5"},{"key":"e_1_2_10_23_1","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2013.57"},{"key":"e_1_2_10_24_1","doi-asserted-by":"publisher","DOI":"10.1002\/cpa.20132"},{"key":"e_1_2_10_25_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.eswa.2022.117528"},{"key":"e_1_2_10_26_1","unstructured":"L.Bergman N.Cohen andY.Hoshen \u201cDeep Nearest Neighbor Anomaly Detection \u201dCoRR:abs\/2002.10445(2020)."},{"key":"e_1_2_10_27_1","doi-asserted-by":"crossref","unstructured":"M.Rudolph T.Wehrbein B.Rosenhahn andB.Wandt \u201cFully Convolutional Cross\u2010Scale\u2010Flows for Image\u2010Based Defect Detection \u201d inProceedings of the IEEE\/CVF Winter Conference on Applications of Computer Vision(IEEE 2022) 1088\u20131097.","DOI":"10.1109\/WACV51458.2022.00189"},{"key":"e_1_2_10_28_1","doi-asserted-by":"crossref","unstructured":"S.Akcay A.Atapour\u2010Abarghouei T. P.Breckon \u201cGanomaly: Semi\u2010Supervised Anomaly Detection via Adversarial Training \u201d inProceedings of the Asian Conference on Computer Vision(Springer 2019) 622\u2013637.","DOI":"10.1007\/978-3-030-20893-6_39"},{"key":"e_1_2_10_29_1","doi-asserted-by":"crossref","unstructured":"S.Ak\u00e7ay A.Atapour\u2010Abarghouei andT. P.Breckon \u201cSkip\u2010Ganomaly: Skip Connected and Adversarially Trained Encoder\u2010Decoder Anomaly Detection \u201d in2019 International Joint Conference on Neural Networks (IJCNN)(IEEE 2019) 1\u20138.","DOI":"10.1109\/IJCNN.2019.8851808"},{"key":"e_1_2_10_30_1","unstructured":"D. P.KingmaandM.Welling \u201cAuto\u2010Encoding Variational Bayes \u201d inProceedings of the International Conference on Learning Representations(IEEE Information Theory Society 2014)."},{"key":"e_1_2_10_31_1","doi-asserted-by":"crossref","unstructured":"P.Bergmann M.Fauser D.Sattlegger andC.Steger \u201cUninformed Students: Student\u2010Teacher Anomaly Detection With Discriminative Latent Embeddings \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2020) 4183\u20134192.","DOI":"10.1109\/CVPR42600.2020.00424"},{"key":"e_1_2_10_32_1","unstructured":"G.Hinton O.Vinyals andJ.Dean \u201cDistilling the Knowledge in a Neural Network \u201dCoRR:abs\/1503.02531(2015)."},{"key":"e_1_2_10_33_1","doi-asserted-by":"crossref","unstructured":"G.Wang S.Han E.Ding andD.Huang \u201cStudent\u2010Teacher Feature Pyramid Matching for Anomaly Detection \u201d inProceedings of the British Machine Vision Conference(The British Machine Vision Association 2021).","DOI":"10.5244\/C.35.349"},{"key":"e_1_2_10_34_1","doi-asserted-by":"crossref","unstructured":"L.Capogrosso F.Girella F.Taioli et\u00a0al. \u201cDiffusion\u2010Based Image Generation for In\u2010Distribution Data Augmentation in Surface Defect Detection \u201d in19th International Conference on Computer Vision Theory and Applications(IEEE 2024).","DOI":"10.5220\/0012350400003660"},{"key":"e_1_2_10_35_1","unstructured":"C. L.Li K.Sohn J.Yoon andT.Pfister \u201cCutpaste: Self\u2010Supervised Learning for Anomaly Detection and Localization \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2021) 9664\u20139674."},{"key":"e_1_2_10_36_1","doi-asserted-by":"crossref","unstructured":"X.Zhang S.Li X.Li P.Huang J.Shan andT.Chen \u201cDeSTSeg: Segmentation Guided Denoising Student\u2010Teacher for Anomaly Detection \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2023) 3914\u20133923.","DOI":"10.1109\/CVPR52729.2023.00381"},{"key":"e_1_2_10_37_1","doi-asserted-by":"crossref","unstructured":"V.Zavrtanik M.Kristan andD.Sko\u010daj \u201cDraem\u2010a Discriminatively Trained Reconstruction Embedding for Surface Anomaly Detection \u201d inProceedings of the IEEE\/CVF International Conference on Computer Vision(IEEE 2021) 8330\u20138339.","DOI":"10.1109\/ICCV48922.2021.00822"},{"key":"e_1_2_10_38_1","doi-asserted-by":"crossref","unstructured":"V.Zavrtanik M.Kristan andD.Sko\u010daj \u201cDsr\u2013a Dual Subspace Re\u2010Projection Network for Surface Anomaly Detection \u201d inEuropean Conference on Computer Vision(Springer 2022) 539\u2013554.","DOI":"10.1007\/978-3-031-19821-2_31"},{"key":"e_1_2_10_39_1","doi-asserted-by":"publisher","DOI":"10.1109\/TII.2024.3438261"},{"key":"e_1_2_10_40_1","doi-asserted-by":"crossref","unstructured":"X.Zhang N.Li J.Li T.Dai Y.Jiang andS. T.Xia \u201cUnsupervised Surface Anomaly Detection With Diffusion Probabilistic Model \u201d inProceedings of the IEEE\/CVF International Conference on Computer Vision(IEEE 2023) 6782\u20136791.","DOI":"10.1109\/ICCV51070.2023.00624"},{"key":"e_1_2_10_41_1","doi-asserted-by":"crossref","unstructured":"R.Rombach A.Blattmann D.Lorenz P.Esser andB.Ommer \u201cHigh\u2010Resolution Image Synthesis With Latent Diffusion Models \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2022) 10684\u201310695.","DOI":"10.1109\/CVPR52688.2022.01042"},{"key":"e_1_2_10_42_1","doi-asserted-by":"crossref","unstructured":"X.Zhang M.Xu andX.Zhou \u201cRealNet: A Feature Selection Network With Realistic Synthetic Anomaly for Anomaly Detection \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2024) 16699\u201316708.","DOI":"10.1109\/CVPR52733.2024.01580"},{"key":"e_1_2_10_43_1","doi-asserted-by":"crossref","unstructured":"T.Defard A.Setkov A.Loesch andR.Audigier \u201cPadim: A Patch Distribution Modeling Framework for Anomaly Detection and Localization \u201d inInternational Conference on Pattern Recognition(Springer 2021) 475\u2013489.","DOI":"10.1007\/978-3-030-68799-1_35"},{"key":"e_1_2_10_44_1","first-page":"S1","article-title":"On the Generalized Distance in Statistics","volume":"80","author":"Mahalanobis P. C.","year":"2018","journal-title":"Sankhy\u0101: The Indian Journal of Statistics, Series A (2008\u2010)"},{"key":"e_1_2_10_45_1","doi-asserted-by":"publisher","DOI":"10.1109\/MSP.2010.939739"},{"key":"e_1_2_10_46_1","doi-asserted-by":"crossref","unstructured":"K.He X.Zhang S.Ren andJ.Sun \u201cDeep Residual Learning for Image Recognition \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2016) 770\u2013778.","DOI":"10.1109\/CVPR.2016.90"},{"key":"e_1_2_10_47_1","unstructured":"C.You andR.Vidal \u201cGeometric Conditions for Subspace\u2010Sparse Recovery \u201d inInternational Conference on Machine Learning(IEEE Information Theory Society 2015) 1585\u20131593."},{"key":"e_1_2_10_48_1","doi-asserted-by":"crossref","unstructured":"C.You D.Robinson andR.Vidal \u201cScalable Sparse Subspace Clustering by Orthogonal Matching Pursuit \u201d inProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2016) 3918\u20133927.","DOI":"10.1109\/CVPR.2016.425"},{"key":"e_1_2_10_49_1","doi-asserted-by":"crossref","unstructured":"Y. C.Pati R.Rezaiifar andP. S.Krishnaprasad \u201cOrthogonal Matching Pursuit: Recursive Function Approximation With Applications to Wavelet Decomposition \u201d inProceedings of Asilomar Conference on Signals Systems and Computers(IEEE 1993) 40\u201344.","DOI":"10.1109\/ACSSC.1993.342465"},{"key":"e_1_2_10_50_1","doi-asserted-by":"publisher","DOI":"10.1109\/TIT.2004.834793"},{"key":"e_1_2_10_51_1","doi-asserted-by":"publisher","DOI":"10.1109\/TIT.2010.2054653"},{"key":"e_1_2_10_52_1","doi-asserted-by":"publisher","DOI":"10.1007\/s00371-018-1588-5"},{"key":"e_1_2_10_53_1","doi-asserted-by":"crossref","unstructured":"P.Mishra R.Verk D.Fornasier C.Piciarelli G. L.Foresti \u201cVT\u2010ADL: A vision Transformer Network for Image Anomaly Detection and Localization \u201d inProceedings of the IEEE 30th International Symposium on Industrial Electronics (ISIE)(IEEE 2021) 01\u201306.","DOI":"10.1109\/ISIE45552.2021.9576231"},{"key":"e_1_2_10_54_1","unstructured":"J.Yu Y.Zheng X.Wang et\u00a0al. \u201cFastflow: Unsupervised Anomaly Detection and Localization via 2D Normalizing Flows \u201dCoRR: abs\/2111.07677(2021)."},{"key":"e_1_2_10_55_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.patcog.2020.107706"},{"key":"e_1_2_10_56_1","doi-asserted-by":"crossref","unstructured":"S.Zagoruyko andN.Komodakis \u201cWide Residual Networks \u201d inProceedings of the British Machine Vision Conference(The British Machine Vision Association 2016) 87.1\u201387.12.","DOI":"10.5244\/C.30.87"},{"key":"e_1_2_10_57_1","doi-asserted-by":"crossref","unstructured":"M.Rudolph B.Wandt andB.Rosenhahn \u201cSame Same but Differnet: Semi\u2010Supervised Defect Detection With Normalizing Flows \u201d inProceedings of the IEEE\/CVF Winter Conference on Applications of Computer Vision(IEEE 2021) 1907\u20131916.","DOI":"10.1109\/WACV48630.2021.00195"},{"key":"e_1_2_10_58_1","unstructured":"M.Tan andQ.Le \u201cEfficientnet: Rethinking Model Scaling for Convolutional Neural Networks \u201d inInternational Conference on Machine Learning(PMLR 2019) 6105\u20136114."}],"container-title":["IET Image Processing"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/ietresearch.onlinelibrary.wiley.com\/doi\/pdf\/10.1049\/ipr2.70157","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/ietresearch.onlinelibrary.wiley.com\/doi\/full-xml\/10.1049\/ipr2.70157","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/ietresearch.onlinelibrary.wiley.com\/doi\/pdf\/10.1049\/ipr2.70157","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,8]],"date-time":"2026-05-08T23:40:12Z","timestamp":1778283612000},"score":1,"resource":{"primary":{"URL":"https:\/\/ietresearch.onlinelibrary.wiley.com\/doi\/10.1049\/ipr2.70157"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,1]]},"references-count":57,"journal-issue":{"issue":"1","published-print":{"date-parts":[[2025,1]]}},"alternative-id":["10.1049\/ipr2.70157"],"URL":"https:\/\/doi.org\/10.1049\/ipr2.70157","archive":["Portico"],"relation":{},"ISSN":["1751-9659","1751-9667"],"issn-type":[{"value":"1751-9659","type":"print"},{"value":"1751-9667","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,1]]},"assertion":[{"value":"2024-03-05","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-07-06","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-07-16","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e70157"}}