FR-MIL: Distribution Re-Calibration-Based Multiple Instance Learning With Transformer for Whole Slide Image Classification

In digital pathology, whole slide images (WSI) are crucial for cancer prognostication and treatment planning. WSI classification is generally addressed using multiple instance learning (MIL), alleviating the challenge of processing billions of pixels and curating rich annotations. Though recent MIL...

Ausführliche Beschreibung

Gespeichert in:

Bibliographische Detailangaben
Veröffentlicht in:	IEEE transactions on medical imaging Jg. 44; H. 1; S. 409 - 421
Hauptverfasser:	Chikontwe, Philip, Kim, Meejeong, Jeong, Jaehoon, Jung Sung, Hyun, Go, Heounjeong, Jeong Nam, Soo, Park, Sang Hyun
Format:	Journal Article
Sprache:	Englisch
Veröffentlicht:	United States IEEE 01.01.2025
Schlagworte:	Algorithms Benchmark testing Feature extraction Histopathology Humans Image Interpretation, Computer-Assisted - methods Machine Learning Medical diagnostic imaging multiple instance learning Multiple-Instance Learning Algorithms Neoplasms - diagnostic imaging Neoplasms - pathology Task analysis Transformers Uncertainty Vectors weakly supervised learning whole slide images
ISSN:	0278-0062, 1558-254X, 1558-254X
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!

Abstract	In digital pathology, whole slide images (WSI) are crucial for cancer prognostication and treatment planning. WSI classification is generally addressed using multiple instance learning (MIL), alleviating the challenge of processing billions of pixels and curating rich annotations. Though recent MIL approaches leverage variants of the attention mechanism to learn better representations, they scarcely study the properties of the data distribution itself i.e., different staining and acquisition protocols resulting in intra-patch and inter-slide variations. In this work, we first introduce a distribution re-calibration strategy to shift the feature distribution of a WSI bag (instances) using the statistics of the max-instance (critical) feature. Second, we enforce class (bag) separation via a metric loss assuming that positive bags exhibit larger magnitudes than negatives. We also introduce a generative process leveraging Vector Quantization (VQ) for improved instance discrimination i.e., VQ helps model bag latent factors for improved classification. To model spatial and context information, a position encoding module (PEM) is employed with transformer-based pooling by multi-head self-attention (PMSA). Evaluation of popular WSI benchmark datasets reveals our approach improves over state-of-the-art MIL methods. Further, we validate the general applicability of our method on classic MIL benchmark tasks and for point cloud classification with limited points. https://github.com/PhilipChicco/FRMIL
AbstractList	In digital pathology, whole slide images (WSI) are crucial for cancer prognostication and treatment planning. WSI classification is generally addressed using multiple instance learning (MIL), alleviating the challenge of processing billions of pixels and curating rich annotations. Though recent MIL approaches leverage variants of the attention mechanism to learn better representations, they scarcely study the properties of the data distribution itself i.e., different staining and acquisition protocols resulting in intra-patch and inter-slide variations. In this work, we first introduce a distribution re-calibration strategy to shift the feature distribution of a WSI bag (instances) using the statistics of the max-instance (critical) feature. Second, we enforce class (bag) separation via a metric loss assuming that positive bags exhibit larger magnitudes than negatives. We also introduce a generative process leveraging Vector Quantization (VQ) for improved instance discrimination i.e., VQ helps model bag latent factors for improved classification. To model spatial and context information, a position encoding module (PEM) is employed with transformer-based pooling by multi-head self-attention (PMSA). Evaluation of popular WSI benchmark datasets reveals our approach improves over state-of-the-art MIL methods. Further, we validate the general applicability of our method on classic MIL benchmark tasks and for point cloud classification with limited points. https://github.com/PhilipChicco/FRMIL In digital pathology, whole slide images (WSI) are crucial for cancer prognostication and treatment planning. WSI classification is generally addressed using multiple instance learning (MIL), alleviating the challenge of processing billions of pixels and curating rich annotations. Though recent MIL approaches leverage variants of the attention mechanism to learn better representations, they scarcely study the properties of the data distribution itself i.e., different staining and acquisition protocols resulting in intra-patch and inter-slide variations. In this work, we first introduce a distribution re-calibration strategy to shift the feature distribution of a WSI bag (instances) using the statistics of the max-instance (critical) feature. Second, we enforce class (bag) separation via a metric loss assuming that positive bags exhibit larger magnitudes than negatives. We also introduce a generative process leveraging Vector Quantization (VQ) for improved instance discrimination i.e., VQ helps model bag latent factors for improved classification. To model spatial and context information, a position encoding module (PEM) is employed with transformer-based pooling by multi-head self-attention (PMSA). Evaluation of popular WSI benchmark datasets reveals our approach improves over state-of-the-art MIL methods. Further, we validate the general applicability of our method on classic MIL benchmark tasks and for point cloud classification with limited points. https://github.com/PhilipChicco/FRMIL.In digital pathology, whole slide images (WSI) are crucial for cancer prognostication and treatment planning. WSI classification is generally addressed using multiple instance learning (MIL), alleviating the challenge of processing billions of pixels and curating rich annotations. Though recent MIL approaches leverage variants of the attention mechanism to learn better representations, they scarcely study the properties of the data distribution itself i.e., different staining and acquisition protocols resulting in intra-patch and inter-slide variations. In this work, we first introduce a distribution re-calibration strategy to shift the feature distribution of a WSI bag (instances) using the statistics of the max-instance (critical) feature. Second, we enforce class (bag) separation via a metric loss assuming that positive bags exhibit larger magnitudes than negatives. We also introduce a generative process leveraging Vector Quantization (VQ) for improved instance discrimination i.e., VQ helps model bag latent factors for improved classification. To model spatial and context information, a position encoding module (PEM) is employed with transformer-based pooling by multi-head self-attention (PMSA). Evaluation of popular WSI benchmark datasets reveals our approach improves over state-of-the-art MIL methods. Further, we validate the general applicability of our method on classic MIL benchmark tasks and for point cloud classification with limited points. https://github.com/PhilipChicco/FRMIL. In digital pathology, whole slide images (WSI) are crucial for cancer prognostication and treatment planning. WSI classification is generally addressed using multiple instance learning (MIL), alleviating the challenge of processing billions of pixels and curating rich annotations. Though recent MIL approaches leverage variants of the attention mechanism to learn better representations, they scarcely study the properties of the data distribution itself i.e., different staining and acquisition protocols resulting in intra-patch and inter-slide variations. In this work, we first introduce a distribution re-calibration strategy to shift the feature distribution of a WSI bag (instances) using the statistics of the max-instance (critical) feature. Second, we enforce class (bag) separation via a metric loss assuming that positive bags exhibit larger magnitudes than negatives. We also introduce a generative process leveraging Vector Quantization (VQ) for improved instance discrimination i.e., VQ helps model bag latent factors for improved classification. To model spatial and context information, a position encoding module (PEM) is employed with transformer-based pooling by multi-head self-attention (PMSA). Evaluation of popular WSI benchmark datasets reveals our approach improves over state-of-the-art MIL methods. Further, we validate the general applicability of our method on classic MIL benchmark tasks and for point cloud classification with limited points. https://github.com/PhilipChicco/FRMIL.
Author	Go, Heounjeong Jeong, Jaehoon Kim, Meejeong Park, Sang Hyun Chikontwe, Philip Jung Sung, Hyun Jeong Nam, Soo
Author_xml	– sequence: 1 givenname: Philip orcidid: 0000-0002-6995-2312 surname: Chikontwe fullname: Chikontwe, Philip email: Philip_Chikontwe@hms.harvard.edu organization: Daegu Gyeongbuk Institute of Science and Technology (DGIST), Daegu, South Korea – sequence: 2 givenname: Meejeong surname: Kim fullname: Kim, Meejeong email: altec0372@gmail.com organization: St. Mary's Hospital and the College of Medicine, The Catholic University of Korea, Seoul, South Korea – sequence: 3 givenname: Jaehoon surname: Jeong fullname: Jeong, Jaehoon email: j.hoon@dgist.ac.kr organization: Daegu Gyeongbuk Institute of Science and Technology (DGIST), Daegu, South Korea – sequence: 4 givenname: Hyun surname: Jung Sung fullname: Jung Sung, Hyun email: shj78730@naver.com organization: Asan Medical Center, Seoul, South Korea – sequence: 5 givenname: Heounjeong surname: Go fullname: Go, Heounjeong email: damul37@amc.seoul.kr organization: Asan Medical Center, Seoul, South Korea – sequence: 6 givenname: Soo surname: Jeong Nam fullname: Jeong Nam, Soo email: soojeong_nam@amc.seoul.kr organization: Asan Medical Center, Seoul, South Korea – sequence: 7 givenname: Sang Hyun orcidid: 0000-0001-7476-1046 surname: Park fullname: Park, Sang Hyun email: shpark13135@dgist.ac.kr organization: Daegu Gyeongbuk Institute of Science and Technology (DGIST), Daegu, South Korea
BackLink	https://www.ncbi.nlm.nih.gov/pubmed/39163176$$D View this record in MEDLINE/PubMed
BookMark	eNp9kc9v0zAYhq1paOsGdw7T5OMuKf6ROAk3VhhUaoU0isYtsp3PmyfH6WznAH897lrQxIHTp096ns_y-56hYz96QOgtJXNKSftus17OGWHlnJelqKk4QjNaVU3BqvLHMZoRVjcFIYKdorMYHwmhZUXaE3TKWyo4rcUM_bq5LdbL1Xv80cYUrJqSHT2-hWIhnVVB7tbiWkbo8XpyyW4d4KWPSXoNeAUyeOvv8Z1ND3gTpI9mDAMEnAe-exgz_M3ZPiuDvAe8cDJGa6x-PvsavTLSRXhzmOfo-82nzeJLsfr6ebn4sCo0ZyQVTAklWt2YsqVcGWVI3_SMaSor2gtFlOSCKCNMDkJUnNGmEb3itZGaK1lTfo6u9ne3YXyaIKZusFGDc9LDOMWOk7aiNeNVmdHLAzqpAfpuG-wgw8_uT14ZIHtAhzHGAOYvQkm3q6TLlXS7SrpDJVkR_yjapucAUpDW_U-82IsWAF68I0pC80d_A2PKmM8
CODEN	ITMID4
CitedBy_id	crossref_primary_10_1016_j_eswa_2025_126831 crossref_primary_10_1016_j_prp_2025_156006 crossref_primary_10_1038_s41597_025_05586_5 crossref_primary_10_3389_fphar_2025_1592950
Cites_doi	10.1016/S0004-3702(96)00034-3 10.1016/j.media.2021.102105 10.1109/TPAMI.2016.2535231 10.1016/j.patcog.2017.08.026 10.1038/s41551-020-00682-w 10.1007/978-3-030-87237-3_19 10.1016/j.cmpb.2011.12.007 10.1109/TAI.2024.3365779 10.1109/TMI.2022.3176598 10.1609/aaai.v35i3.16280 10.1016/j.artint.2013.06.003 10.1109/CVPR52688.2022.01567 10.1016/j.cviu.2021.103299 10.1609/aaai.v34i04.6030 10.5555/3524938.3525087 10.1109/CVPR.2018.00813 10.1098/rsos.220638 10.61603/ceas.v2i1.33 10.1109/CVPR52729.2023.01503 10.1109/CVPR46437.2021.01379 10.1109/CVPR.2016.90 10.5555/3495724.3497510 10.3389/fmed.2019.00264 10.1007/978-3-030-87237-3_57 10.1001/jama.2017.14585 10.1002/widm.1439 10.1016/j.media.2020.101813 10.1038/s41591-019-0508-1 10.1609/aaai.v36i2.20051 10.1053/j.gastro.2009.12.064 10.1007/978-3-030-87237-3_20 10.1109/WACV48630.2021.00176 10.1109/TMI.2022.3171418 10.1007/978-3-031-16434-7_41 10.1109/JPROC.2021.3058954 10.1007/978-3-031-16434-7_25 10.1007/978-3-031-16434-7_4 10.1007/978-3-030-59722-1_50 10.1016/j.media.2022.102482 10.1109/CVPR46437.2021.01409 10.48550/ARXIV.1706.03762 10.1016/j.patcog.2017.10.009 10.1109/CVPR52688.2022.01824 10.1007/s10462-021-10121-0 10.1109/CVPR52729.2023.01899
ContentType	Journal Article
DBID	97E RIA RIE AAYXX CITATION CGR CUY CVF ECM EIF NPM 7X8
DOI	10.1109/TMI.2024.3446716
DatabaseName	IEEE Xplore (IEEE) IEEE All-Society Periodicals Package (ASPP) 1998–Present IEEE Xplore Digital Library CrossRef Medline MEDLINE MEDLINE (Ovid) MEDLINE MEDLINE PubMed MEDLINE - Academic
DatabaseTitle	CrossRef MEDLINE Medline Complete MEDLINE with Full Text PubMed MEDLINE (Ovid) MEDLINE - Academic
DatabaseTitleList	MEDLINE - Academic MEDLINE
Database_xml	– sequence: 1 dbid: NPM name: PubMed url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed sourceTypes: Index Database – sequence: 2 dbid: RIE name: IEEE Electronic Library (IEL) url: https://ieeexplore.ieee.org/ sourceTypes: Publisher – sequence: 3 dbid: 7X8 name: MEDLINE - Academic url: https://search.proquest.com/medline sourceTypes: Aggregation Database
DeliveryMethod	fulltext_linktorsrc
Discipline	Medicine Engineering
EISSN	1558-254X
EndPage	421
ExternalDocumentID	39163176 10_1109_TMI_2024_3446716 10640165
Genre	orig-research Research Support, Non-U.S. Gov't Journal Article
GrantInformation_xml	– fundername: Smart HealthCare Program funderid: 10.13039/501100003600 – fundername: Korean National Police Agency grantid: 220222M01 funderid: 10.13039/501100003600 – fundername: Korean Government [Ministry of Science and ICT (MSIT)] through the Artificial Intelligence Innovation Hub grantid: 2021-0-02068 – fundername: Institute for Information and Communication Technology Planning and Evaluation (IITP)
GroupedDBID	--- -DZ -~X .GJ 0R~ 29I 4.4 53G 5GY 5RE 5VS 6IK 97E AAJGR AARMG AASAJ AAWTH ABAZT ABQJQ ABVLG ACGFO ACGFS ACIWK ACNCT ACPRK AENEX AETIX AFRAH AGQYO AGSQL AHBIQ AI. AIBXA AKJIK AKQYR ALLEH ALMA_UNASSIGNED_HOLDINGS ASUFR ATWAV BEFXN BFFAM BGNUA BKEBE BPEOZ CS3 DU5 EBS EJD F5P HZ~ H~9 IBMZZ ICLAB IFIPE IFJZH IPLJI JAVBF LAI M43 MS~ O9- OCL P2P PQQKQ RIA RIE RNS RXW TAE TN5 VH1 AAYXX CITATION CGR CUY CVF ECM EIF NPM RIG 7X8
ID	FETCH-LOGICAL-c320t-2b6b69c8f4913bfbf0d8d22c1a51d6b0ba360bf6f20265321886db37fac3ba713
IEDL.DBID	RIE
ISICitedReferencesCount	6
ISICitedReferencesURI	http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001389746700008&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN	0278-0062 1558-254X
IngestDate	Mon Sep 29 04:47:16 EDT 2025 Wed Jul 30 01:47:44 EDT 2025 Sat Nov 29 05:14:12 EST 2025 Tue Nov 18 21:43:03 EST 2025 Wed Aug 27 02:02:01 EDT 2025
IsPeerReviewed	false
IsScholarly	true
Issue	1
Language	English
License	https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html https://doi.org/10.15223/policy-029 https://doi.org/10.15223/policy-037
LinkModel	DirectLink
MergedId	FETCHMERGED-LOGICAL-c320t-2b6b69c8f4913bfbf0d8d22c1a51d6b0ba360bf6f20265321886db37fac3ba713
Notes	ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23
ORCID	0000-0002-6995-2312 0000-0001-7476-1046
PMID	39163176
PQID	3095172354
PQPubID	23479
PageCount	13
ParticipantIDs	ieee_primary_10640165 crossref_primary_10_1109_TMI_2024_3446716 crossref_citationtrail_10_1109_TMI_2024_3446716 pubmed_primary_39163176 proquest_miscellaneous_3095172354
PublicationCentury	2000
PublicationDate	2025-Jan. 2025-1-00 2025-01-00 20250101
PublicationDateYYYYMMDD	2025-01-01
PublicationDate_xml	– month: 01 year: 2025 text: 2025-Jan.
PublicationDecade	2020
PublicationPlace	United States
PublicationPlace_xml	– name: United States
PublicationTitle	IEEE transactions on medical imaging
PublicationTitleAbbrev	TMI
PublicationTitleAlternate	IEEE Trans Med Imaging
PublicationYear	2025
Publisher	IEEE
Publisher_xml	– name: IEEE
References	ref13 Zhang (ref29) 2021 ref15 ref59 ref14 ref53 ref52 Chang (ref60) 2015 ref11 ref55 ref10 ref54 Shao (ref19); 34 ref16 ref18 ref51 ref50 Agustsson (ref58) 2017 Tu (ref62) 2019 ref45 ref47 ref42 Sharma (ref12) 2021 ref41 Yan (ref64) Chalapathy (ref24) 2019 Ilse (ref17) Zaheer (ref48); 30 ref49 ref8 ref7 Theis (ref57) ref9 ref4 ref3 ref6 ref5 ref40 Yang (ref26) ref35 ref34 ref37 ref36 ref30 ref33 ref2 ref1 Dosovitskiy (ref22) ref39 ref38 Van Den Oord (ref31); 30 Brehmer (ref46); 35 ref23 Zhang (ref27); 35 ref25 ref20 ref63 ref66 ref21 ref65 Lei Ba (ref56) 2016 Andrews (ref44); 15 Maron (ref43); 10 Lee (ref32) Ghaffarzadegan (ref28) ref61
References_xml	– ident: ref59 doi: 10.1016/S0004-3702(96)00034-3 – year: 2021 ident: ref29 article-title: Non-IID multi-instance learning for predicting instance and bag labels using variational auto-encoder publication-title: arXiv:2105.01276 – ident: ref40 doi: 10.1016/j.media.2021.102105 – ident: ref41 doi: 10.1109/TPAMI.2016.2535231 – ident: ref5 doi: 10.1016/j.patcog.2017.08.026 – start-page: 1 volume-title: Proc. Int. Conf. Learn. Represent. ident: ref57 article-title: Lossy image compression with compressive autoencoders – ident: ref15 doi: 10.1038/s41551-020-00682-w – ident: ref10 doi: 10.1007/978-3-030-87237-3_19 – ident: ref2 doi: 10.1016/j.cmpb.2011.12.007 – ident: ref53 doi: 10.1109/TAI.2024.3365779 – ident: ref55 doi: 10.1109/TMI.2022.3176598 – ident: ref23 doi: 10.1609/aaai.v35i3.16280 – ident: ref6 doi: 10.1016/j.artint.2013.06.003 – ident: ref51 doi: 10.1109/CVPR52688.2022.01567 – ident: ref42 doi: 10.1016/j.cviu.2021.103299 – ident: ref16 doi: 10.1609/aaai.v34i04.6030 – year: 2015 ident: ref60 article-title: Shapenet: An information-rich 3D model repository publication-title: arXiv:1512.03012 – volume: 15 start-page: 1 volume-title: Proc. NeurIPS ident: ref44 article-title: Support vector machines for multiple-instance learning – ident: ref49 doi: 10.5555/3524938.3525087 – ident: ref63 doi: 10.1109/CVPR.2018.00813 – ident: ref45 doi: 10.1098/rsos.220638 – volume: 35 start-page: 38319 volume-title: Proc. NeurIPS ident: ref46 article-title: Weakly supervised causal representation learning – ident: ref30 doi: 10.61603/ceas.v2i1.33 – ident: ref54 doi: 10.1109/CVPR52729.2023.01503 – ident: ref25 doi: 10.1109/CVPR46437.2021.01379 – start-page: 1 volume-title: Proc. 32nd AAAI Conf. Artif. Intell. ident: ref28 article-title: Deep multiple instance feature learning via variational autoencoder – ident: ref9 doi: 10.1109/CVPR.2016.90 – ident: ref50 doi: 10.5555/3495724.3497510 – volume: 30 start-page: 1 volume-title: Proc. NeurIPS ident: ref31 article-title: Neural discrete representation learning – ident: ref1 doi: 10.3389/fmed.2019.00264 – ident: ref11 doi: 10.1007/978-3-030-87237-3_57 – start-page: 1 volume-title: Proc. ICLR ident: ref22 article-title: An image is worth 16×16 words: Transformers for image recognition at scale – start-page: 1 volume-title: Proc. ICLR ident: ref26 article-title: Free lunch for few-shot learning: Distribution calibration – ident: ref36 doi: 10.1001/jama.2017.14585 – ident: ref3 doi: 10.1002/widm.1439 – ident: ref7 doi: 10.1016/j.media.2020.101813 – ident: ref21 doi: 10.1038/s41591-019-0508-1 – ident: ref65 doi: 10.1609/aaai.v36i2.20051 – start-page: 3744 volume-title: Proc. Int. Conf. Mach. Learn. (ICML) ident: ref32 article-title: Set Transformer: A framework for attention-based permutation-invariant neural networks – ident: ref61 doi: 10.1053/j.gastro.2009.12.064 – year: 2017 ident: ref58 article-title: Soft-to-hard vector quantization for end-to-end learned compression of images and neural networks publication-title: arXiv:1704.00648 – ident: ref20 doi: 10.1007/978-3-030-87237-3_20 – ident: ref13 doi: 10.1109/WACV48630.2021.00176 – ident: ref38 doi: 10.1109/TMI.2022.3171418 – ident: ref33 doi: 10.1007/978-3-031-16434-7_41 – volume: 34 start-page: 1 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref19 article-title: Transmil: Transformer based correlated multiple instance learning for whole slide image classification – ident: ref47 doi: 10.1109/JPROC.2021.3058954 – volume: 35 start-page: 34940 volume-title: Proc. NeurIPS ident: ref27 article-title: Multi-instance causal representation learning for instance label prediction and out-of-distribution generalization – ident: ref52 doi: 10.1007/978-3-031-16434-7_25 – start-page: 682 volume-title: Medical Imaging With Deep Learning year: 2021 ident: ref12 article-title: Cluster-to-conquer: A framework for end-to-end multi-instance learning for whole slide image classification – start-page: 2127 volume-title: Proc. Int. Conf. Mach. Learn. ident: ref17 article-title: Attention-based deep multiple instance learning – year: 2016 ident: ref56 article-title: Layer normalization publication-title: arXiv:1607.06450 – ident: ref37 doi: 10.1007/978-3-031-16434-7_4 – volume: 30 start-page: 1 volume-title: Proc. NeurIPS ident: ref48 article-title: Deep sets – volume: 10 start-page: 1 volume-title: Proc. NeurIPS ident: ref43 article-title: A framework for multiple-instance learning – year: 2019 ident: ref62 article-title: Multiple instance learning with graph neural networks publication-title: arXiv:1906.04881 – ident: ref14 doi: 10.1007/978-3-030-59722-1_50 – start-page: 662 volume-title: Proc. Asian Conf. Mach. Learn. ident: ref64 article-title: Deep multi-instance learning with dynamic pooling – ident: ref35 doi: 10.1016/j.media.2022.102482 – ident: ref18 doi: 10.1109/CVPR46437.2021.01409 – year: 2019 ident: ref24 article-title: Deep learning for anomaly detection: A survey publication-title: arXiv:1901.03407 – ident: ref8 doi: 10.48550/ARXIV.1706.03762 – ident: ref34 doi: 10.1016/j.patcog.2017.10.009 – ident: ref39 doi: 10.1109/CVPR52688.2022.01824 – ident: ref4 doi: 10.1007/s10462-021-10121-0 – ident: ref66 doi: 10.1109/CVPR52729.2023.01899
SSID	ssj0014509
Score	2.4956114
Snippet	In digital pathology, whole slide images (WSI) are crucial for cancer prognostication and treatment planning. WSI classification is generally addressed using...
SourceID	proquest pubmed crossref ieee
SourceType	Aggregation Database Index Database Enrichment Source Publisher
StartPage	409
SubjectTerms	Algorithms Benchmark testing Feature extraction Histopathology Humans Image Interpretation, Computer-Assisted - methods Machine Learning Medical diagnostic imaging multiple instance learning Multiple-Instance Learning Algorithms Neoplasms - diagnostic imaging Neoplasms - pathology Task analysis Transformers Uncertainty Vectors weakly supervised learning whole slide images
Title	FR-MIL: Distribution Re-Calibration-Based Multiple Instance Learning With Transformer for Whole Slide Image Classification
URI	https://ieeexplore.ieee.org/document/10640165 https://www.ncbi.nlm.nih.gov/pubmed/39163176 https://www.proquest.com/docview/3095172354
Volume	44
WOSCitedRecordID	wos001389746700008&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText	1
inHoldings	1
isFullTextHit
isPrint
journalDatabaseRights	– providerCode: PRVIEE databaseName: IEEE Electronic Library (IEL) customDbUrl: eissn: 1558-254X dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0014509 issn: 0278-0062 databaseCode: RIE dateStart: 19820101 isFulltext: true titleUrlDefault: https://ieeexplore.ieee.org/ providerName: IEEE
link	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1NbxMxEB3RClXlUKAUCIXKSFw4uPV6vV6bW_mIiNRUqAQ1t9V6bEOkkqA06YFfj-31RuVQJE67B9ta7Rt7ZjzzZgDeOI61LiyjKISnQgYHRXnmKCqhPXLLlE1E4bP6_FxNp_pLJqsnLoxzLiWfueP4mmL5doHreFUWdrgUkX6zBVt1LTuy1iZkIKoun4PHkrFM8j4myfTJZDwKniAXx2VwfoKDsAs7kW8aVKf8Sx2l_ip3m5pJ5Qwf_ufHPoK9bFuS004YHsM9N9-HB7cqDu7DzjjH0p_A7-EFHY_O3pGPsXZubntFLhyNdC3TCQZ9H5ScJeOcdUhGyZhER3JZ1u_kcrb6QSa99euWJDzIZWy6S75ezWyY8jOcWCT13oxZSWnZA_g2_DT58JnmTgwUS85WlBtppEblhS5K441nVlnOsWirwkrDTFtKZrz04S_Lqgxmg5LWlLVvsTRt8IOfwvZ8MXfPgaDhCqVELataKORaeoZodGG0sYWuB3DSA9JgLlMeu2VcNcldYboJaDYRzSajOYC3mxm_uhId_xh7EJG6Na4DaQCve9CbsL9i0KSdu8X6uimjDVrzshIDeNZJw2Z2L0Qv7lj1EHZ5bBecbmxewvZquXav4D7erGbXy6MgxFN1lIT4D5nN6yo
linkProvider	IEEE
linkToHtml	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LbxMxEB5BQaUcePQB4WkkLhzcem2v1-bGK2pENkIlqL2t1i8aqSQoTTjw67G93qgcisRp92Bbq52xZ8Yz33wArx01lSoswYZzj7kIAYr0xGEjufKGWiJtAgqPq8lEnp2pLxmsnrAwzrlUfOYO42vK5duFWcersrDDBY_wm5twK1JnlR1ca5M04GVX0UFj01giaJ-VJOpoWo9CLEj5IQvhTwgRdmA7Ik6D8RR_GaTEsHK9s5mMzvD-f37uA7iXvUv0rlOHh3DDzXfh7pWeg7uwXeds-h78Hp7gejR-iz7G7rmZ-AqdOBwBW7pTDfw-mDmL6lx3iEbJnTQO5cas39HpbHWOpr3_65YoPNBppN1FXy9mNkz5Ec4slNg3Y11SWnYfvg0_TT8c48zFgA2jZIWpFlooIz1XBdNee2KlpdQUbVlYoYlumSDaCx_-sihZcByksJpVvjVMtyESPoCt-WLuHgMymkojhFGirLg0VAlPjNGq0ErbQlUDOOoF0pjcqDzyZVw0KWAhqgnSbKI0myzNAbzZzPjZNen4x9j9KKkr4zohDeBVL_Qm7LCYNmnnbrG-bFj0QivKSj6AR502bGb3SvTkmlVfwp3jaT1uxqPJ56ewQyN5cLq_eQZbq-XaPYfb5tdqdrl8kVT5D3vJ7Y0
openUrl	ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=FR-MIL%3A+Distribution+Re-Calibration-Based+Multiple+Instance+Learning+With+Transformer+for+Whole+Slide+Image+Classification&rft.jtitle=IEEE+transactions+on+medical+imaging&rft.au=Chikontwe%2C+Philip&rft.au=Kim%2C+Meejeong&rft.au=Jeong%2C+Jaehoon&rft.au=Jung+Sung%2C+Hyun&rft.date=2025-01-01&rft.eissn=1558-254X&rft.volume=44&rft.issue=1&rft.spage=409&rft_id=info:doi/10.1109%2FTMI.2024.3446716&rft_id=info%3Apmid%2F39163176&rft.externalDocID=39163176
thumbnail_l	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0278-0062&client=summon
thumbnail_m	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0278-0062&client=summon
thumbnail_s	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0278-0062&client=summon