Disentangled Pseudo-Bag Augmentation for Whole Slide Image Multiple Instance Learning

As the predominant approach for pathological whole slide image (WSI) classification, multiple instance learning (MIL) methods struggle with limited labeled WSIs. Although MIL has achieved notable progress with pseudo-bag-oriented augmentation methods, their effectiveness is often constrained by nois...

Full description

Saved in:

Bibliographic Details
Published in:	IEEE transactions on medical imaging Vol. 44; no. 11; pp. 4181 - 4197
Main Authors:	Dong, Jiuyang, Jiang, Junjun, Jiang, Kui, Li, Jiahan, Cai, Linghan, Zhang, Yongbing
Format:	Journal Article
Language:	English
Published:	United States IEEE 01.11.2025
Subjects:	Accuracy Algorithms Cancer Data augmentation Data mining Feature extraction Humans Image Interpretation, Computer-Assisted - methods Machine Learning multiple instance learning Multiple-Instance Learning Algorithms Neural Networks, Computer Pathology Phenotypes Probability Training Tumors Whole slide image classification
ISSN:	0278-0062, 1558-254X, 1558-254X
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Abstract	As the predominant approach for pathological whole slide image (WSI) classification, multiple instance learning (MIL) methods struggle with limited labeled WSIs. Although MIL has achieved notable progress with pseudo-bag-oriented augmentation methods, their effectiveness is often constrained by noisy pseudo-labels and low-quality pseudo-bags. To overcome these problems, we revisit the use of pseudo-bags for WSI data augmentation and propose a new pseudo-bag generation paradigm, dubbed DPBAug. Its distinctive features can be summarized as: i) We develop an intra-slide pseudo-bag generation module, which separates the heterogeneous instances within each slide through phenotype partitioning. Moreover, to ensure accurate label inheritance when generating pseudo-bags, we propose an instance sampling algorithm with replacement. ii) An inter-slide pseudo-bag fusion module is designed to integrate heterogeneous information across multiple WSIs, producing high-quality training samples that better leverage the potential of neural networks. iii) A pseudo-bag memory update module prioritizes valuable synthetic pseudo-bags. This further enhances the network's classification performance. Extensive experiments demonstrate that DPBAug surpasses existing augmentation methods, enhancing the classification performance and reliability of multiple MIL baselines across various public datasets. DPBAug also improves the generalization and data efficiency of existing MIL methods, facilitating their adoption in clinical practice and rare cancer research The project is available at: https://github.com/JiuyangDong/DPBAug .
AbstractList	As the predominant approach for pathological whole slide image (WSI) classification, multiple instance learning (MIL) methods struggle with limited labeled WSIs. Although MIL has achieved notable progress with pseudo-bag-oriented augmentation methods, their effectiveness is often constrained by noisy pseudo-labels and low-quality pseudo-bags. To overcome these problems, we revisit the use of pseudo-bags for WSI data augmentation and propose a new pseudo-bag generation paradigm, dubbed DPBAug. Its distinctive features can be summarized as: i) We develop an intra-slide pseudo-bag generation module, which separates the heterogeneous instances within each slide through phenotype partitioning. Moreover, to ensure accurate label inheritance when generating pseudo-bags, we propose an instance sampling algorithm with replacement. ii) An inter-slide pseudo-bag fusion module is designed to integrate heterogeneous information across multiple WSIs, producing high-quality training samples that better leverage the potential of neural networks. iii) A pseudo-bag memory update module prioritizes valuable synthetic pseudo-bags. This further enhances the network's classification performance. Extensive experiments demonstrate that DPBAug surpasses existing augmentation methods, enhancing the classification performance and reliability of multiple MIL baselines across various public datasets. DPBAug also improves the generalization and data efficiency of existing MIL methods, facilitating their adoption in clinical practice and rare cancer research The project is available at: https://github.com/JiuyangDong/DPBAug. As the predominant approach for pathological whole slide image (WSI) classification, multiple instance learning (MIL) methods struggle with limited labeled WSIs. Although MIL has achieved notable progress with pseudo-bag-oriented augmentation methods, their effectiveness is often constrained by noisy pseudo-labels and low-quality pseudo-bags. To overcome these problems, we revisit the use of pseudo-bags for WSI data augmentation and propose a new pseudo-bag generation paradigm, dubbed DPBAug. Its distinctive features can be summarized as: i) We develop an intra-slide pseudobag generation module, which separates the heterogeneous instances within each slide through phenotype partitioning. Moreover, to ensure accurate label inheritance when generating pseudo-bags, we propose an instance sampling algorithm with replacement. ii) An inter-slide pseudo-bag fusion module is designed to integrate heterogeneous information across multiple WSIs, producing high-quality training samples that better leverage the potential of neural networks. iii) A pseudo-bag memory update module prioritizes valuable synthetic pseudo-bags. This further enhances the network's classification performance. Extensive experiments demonstrate that DPBAug surpasses existing augmentation methods, enhancing the classification performance and reliability of multiple MIL baselines across various public datasets. DPBAug also improves the generalization and data efficiency of existing MIL methods, facilitating their adoption in clinical practice and rare cancer research. The project is available at: https://github.com/JiuyangDong/DPBAug.As the predominant approach for pathological whole slide image (WSI) classification, multiple instance learning (MIL) methods struggle with limited labeled WSIs. Although MIL has achieved notable progress with pseudo-bag-oriented augmentation methods, their effectiveness is often constrained by noisy pseudo-labels and low-quality pseudo-bags. To overcome these problems, we revisit the use of pseudo-bags for WSI data augmentation and propose a new pseudo-bag generation paradigm, dubbed DPBAug. Its distinctive features can be summarized as: i) We develop an intra-slide pseudobag generation module, which separates the heterogeneous instances within each slide through phenotype partitioning. Moreover, to ensure accurate label inheritance when generating pseudo-bags, we propose an instance sampling algorithm with replacement. ii) An inter-slide pseudo-bag fusion module is designed to integrate heterogeneous information across multiple WSIs, producing high-quality training samples that better leverage the potential of neural networks. iii) A pseudo-bag memory update module prioritizes valuable synthetic pseudo-bags. This further enhances the network's classification performance. Extensive experiments demonstrate that DPBAug surpasses existing augmentation methods, enhancing the classification performance and reliability of multiple MIL baselines across various public datasets. DPBAug also improves the generalization and data efficiency of existing MIL methods, facilitating their adoption in clinical practice and rare cancer research. The project is available at: https://github.com/JiuyangDong/DPBAug.
Author	Li, Jiahan Jiang, Kui Cai, Linghan Zhang, Yongbing Dong, Jiuyang Jiang, Junjun
Author_xml	– sequence: 1 givenname: Jiuyang orcidid: 0009-0008-2114-3525 surname: Dong fullname: Dong, Jiuyang email: jiuyang.dong@stu.hit.edu.cn organization: School of Computer Science and Technology, Harbin Institute of Technology, Harbin, China – sequence: 2 givenname: Junjun orcidid: 0000-0002-5694-505X surname: Jiang fullname: Jiang, Junjun email: jiangjunjun@hit.edu.cn organization: School of Computer Science and Technology, Harbin Institute of Technology, Harbin, China – sequence: 3 givenname: Kui orcidid: 0000-0002-4055-7503 surname: Jiang fullname: Jiang, Kui email: jiangkui@hit.edu.cn organization: School of Computer Science and Technology, Harbin Institute of Technology, Harbin, China – sequence: 4 givenname: Jiahan orcidid: 0000-0003-2541-5280 surname: Li fullname: Li, Jiahan email: jiahan.li@stu.hit.edu.cn organization: School of Computer Science and Technology, Harbin Institute of Technology, Harbin, China – sequence: 5 givenname: Linghan surname: Cai fullname: Cai, Linghan email: cailh@stu.hit.edu.cn organization: School of Computer Science and Technology, Harbin Institute of Technology, Shenzhen, China – sequence: 6 givenname: Yongbing orcidid: 0000-0003-3320-2904 surname: Zhang fullname: Zhang, Yongbing email: ybzhang08@hit.edu.cn organization: School of Computer Science and Technology, Harbin Institute of Technology, Shenzhen, China
BackLink	https://www.ncbi.nlm.nih.gov/pubmed/40366836$$D View this record in MEDLINE/PubMed
BookMark	eNpFkE1Pg0AQhjemxn7o3YMxHL1QZz-Bo9Yvkjaa2EZvZIFZxMBSWTj476Vp1cxhkpnnfQ_PlIxsY5GQcwpzSiG6Xq_iOQMm51yqKBL0iEyolKHPpHgfkQmwIPQBFBuTqXOfAFRIiE7IWABXKuRqQjZ3pUPbaVtUmHsvDvu88W914d30Rb17dGVjPdO03ttHU6H3WpU5enGtC_RWfdWV2-EYWzc0ZOgtUbe2tMUpOTa6cnh22DOyebhfL5785fNjvLhZ-hkLZOebMNU8D7mOUiaU1JEKRBDyLM2BRwFmJoOcZwGKYQwXUarAQC5RmJQaISmfkat977Ztvnp0XVKXLsOq0hab3iWcgeBMhYEY0MsD2qc15sm2LWvdfie_KgYA9kDWNs61aP4QCsnOdjLYTna2k4PtIXKxj5SI-I9TAEGV4D9NVnqS
CODEN	ITMID4
Cites_doi	10.2307/2346830 10.1016/j.media.2024.103289 10.48550/arxiv.1710.09412 10.1016/j.media.2019.101544 10.1109/tai.2024.3454591 10.1038/s41551-020-00682-w 10.1038/s41586-024-07441-w 10.1109/CVPR46437.2021.01409 10.1109/CVPR52688.2022.01824 10.1038/s41591-024-02857-3 10.1038/s41467-021-25296-x 10.1002/(SICI)1097-0258(19990915/30)18:17/18<2529::AID-SIM274>3.0.CO;2-5 10.1080/00224065.1986.11979014 10.1109/CVPR52729.2023.01503 10.5555/3454287.3455008 10.1109/CVPR.2015.7298780 10.1007/978-3-319-66179-7_69 10.1038/s41591-023-02504-3 10.1001/jama.2017.14585 10.1016/j.ccell.2022.07.004 10.1109/CVPR.2016.90 10.1007/978-3-031-72083-3_28 10.1007/978-3-031-43987-2_46 10.1007/978-3-030-87199-4_11 10.1093/bioinformatics/btw252 10.1038/s43018-022-00416-8 10.1109/CVPR52733.2024.01070 10.1109/TMI.2018.2867350 10.1109/CVPR52729.2023.02292 10.1038/s44222-023-00096-8 10.1109/CVPR.2009.5206848 10.1007/978-3-031-16434-7_4 10.1109/CVPR52688.2022.01567 10.1109/TMI.2021.3097319 10.1016/j.patcog.2017.08.026 10.1038/s41467-024-51465-9 10.1609/aaai.v31i1.10890 10.1109/TMI.2024.3351213 10.1109/ISBI53787.2023.10230723
ContentType	Journal Article
DBID	97E RIA RIE AAYXX CITATION CGR CUY CVF ECM EIF NPM 7X8
DOI	10.1109/TMI.2025.3569941
DatabaseName	IEEE All-Society Periodicals Package (ASPP) 2005–Present IEEE All-Society Periodicals Package (ASPP) 1998–Present IEEE Electronic Library (IEL) 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 MEDLINE - Academic
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	4197
ExternalDocumentID	40366836 10_1109_TMI_2025_3569941 11004164
Genre	orig-research Journal Article
GrantInformation_xml	– fundername: National Natural Science Foundation of China grantid: 62031023; 62331011 – fundername: Shenzhen Science and Technology Project grantid: GXWD20220818170353009 funderid: 10.13039/501100001809
GroupedDBID	--- -DZ -~X .GJ 0R~ 29I 4.4 53G 5GY 5RE 5VS 6IK 97E AAJGR 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 7X8
ID	FETCH-LOGICAL-c275t-f8ba3d83a9b2465a9674783cbd0397ecfc0d3c7e4e4ef349b60f0d5e4fb1f4513
IEDL.DBID	RIE
ISICitedReferencesCount	0
ISICitedReferencesURI	http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001616144400033&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	Thu Oct 02 23:23:19 EDT 2025 Fri Nov 07 01:55:23 EST 2025 Sat Nov 29 06:52:22 EST 2025 Wed Nov 19 08:27:20 EST 2025
IsPeerReviewed	false
IsScholarly	true
Issue	11
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-c275t-f8ba3d83a9b2465a9674783cbd0397ecfc0d3c7e4e4ef349b60f0d5e4fb1f4513
Notes	ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23
ORCID	0000-0003-2541-5280 0009-0008-2114-3525 0000-0002-4055-7503 0000-0003-3320-2904 0000-0002-5694-505X
PMID	40366836
PQID	3204326874
PQPubID	23479
PageCount	17
ParticipantIDs	pubmed_primary_40366836 crossref_primary_10_1109_TMI_2025_3569941 ieee_primary_11004164 proquest_miscellaneous_3204326874
PublicationCentury	2000
PublicationDate	2025-11-01
PublicationDateYYYYMMDD	2025-11-01
PublicationDate_xml	– month: 11 year: 2025 text: 2025-11-01 day: 01
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 ref12 ref15 ref14 ref11 ref10 Liu (ref32) ref17 ref16 ref18 Maron (ref5); 10 ref46 ref45 ref48 ref47 ref42 Dosovitskiy (ref29) 2020 ref41 ref44 ref43 ref49 Maaten (ref36) 2008; 9 ref7 ref9 ref4 Yang (ref19) 2023 Ilse (ref25) ref40 Shao (ref26) ref35 ref34 ref37 ref31 ref30 ref33 Keeler (ref6); 3 ref2 ref1 ref39 ref38 ref24 ref23 ref20 Wang (ref3); 35 ref22 ref21 ref28 ref27 Ramon (ref8)
References_xml	– ident: ref39 doi: 10.2307/2346830 – ident: ref42 doi: 10.1016/j.media.2024.103289 – ident: ref34 doi: 10.48550/arxiv.1710.09412 – start-page: 53 volume-title: Proc. ICML- Workshop Attribute-Value Relational Learn. ident: ref8 article-title: Multi instance neural networks – ident: ref11 doi: 10.1016/j.media.2019.101544 – ident: ref16 doi: 10.1109/tai.2024.3454591 – ident: ref27 doi: 10.1038/s41551-020-00682-w – ident: ref41 doi: 10.1038/s41586-024-07441-w – ident: ref24 doi: 10.1109/CVPR46437.2021.01409 – ident: ref20 doi: 10.1109/CVPR52688.2022.01824 – start-page: 253 volume-title: Proc. Asian Conf. Mach. Learn. ident: ref32 article-title: Key instance detection in multi-instance learning – ident: ref44 doi: 10.1038/s41591-024-02857-3 – ident: ref12 doi: 10.1038/s41467-021-25296-x – ident: ref49 doi: 10.1002/(SICI)1097-0258(19990915/30)18:17/18<2529::AID-SIM274>3.0.CO;2-5 – ident: ref38 doi: 10.1080/00224065.1986.11979014 – ident: ref2 doi: 10.1109/CVPR52729.2023.01503 – ident: ref40 doi: 10.5555/3454287.3455008 – ident: ref9 doi: 10.1109/CVPR.2015.7298780 – ident: ref7 doi: 10.1007/978-3-319-66179-7_69 – start-page: 2127 volume-title: Proc. Int. Conf. Mach. Learn. ident: ref25 article-title: Attention-based deep multiple instance learning – ident: ref43 doi: 10.1038/s41591-023-02504-3 – ident: ref31 doi: 10.1001/jama.2017.14585 – ident: ref45 doi: 10.1016/j.ccell.2022.07.004 – ident: ref28 doi: 10.1109/CVPR.2016.90 – volume: 9 start-page: 2579 issue: 86 year: 2008 ident: ref36 article-title: Visualizing data using t-SNE publication-title: J. Mach. Learn. Res. – volume: 3 start-page: 557 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref6 article-title: Integrated segmentation and recognition of hand-printed numerals – volume: 35 start-page: 18009 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref3 article-title: SCL-WC: Cross-slide contrastive learning for weakly-supervised whole-slide image classification – ident: ref37 doi: 10.1007/978-3-031-72083-3_28 – ident: ref23 doi: 10.1007/978-3-031-43987-2_46 – ident: ref13 doi: 10.1007/978-3-030-87199-4_11 – year: 2020 ident: ref29 article-title: An image is worth 16×16 words: Transformers for image recognition at scale publication-title: arXiv:2010.11929 – ident: ref4 doi: 10.1093/bioinformatics/btw252 – ident: ref46 doi: 10.1038/s43018-022-00416-8 – volume: 10 start-page: 570 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref5 article-title: A framework for multiple-instance learning – ident: ref14 doi: 10.1109/CVPR52733.2024.01070 – ident: ref33 doi: 10.1109/TMI.2018.2867350 – ident: ref18 doi: 10.1109/CVPR52729.2023.02292 – start-page: 2136 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref26 article-title: TransMIL: Transformer based correlated multiple instance learning for whole slide image classification – ident: ref47 doi: 10.1038/s44222-023-00096-8 – ident: ref30 doi: 10.1109/CVPR.2009.5206848 – ident: ref17 doi: 10.1007/978-3-031-16434-7_4 – year: 2023 ident: ref19 article-title: ProtoDiv: Prototype-guided division of consistent pseudo-bags for whole-slide image classification publication-title: arXiv:2304.06652 – ident: ref1 doi: 10.1109/CVPR52688.2022.01567 – ident: ref22 doi: 10.1109/TMI.2021.3097319 – ident: ref35 doi: 10.1016/j.patcog.2017.08.026 – ident: ref48 doi: 10.1038/s41467-024-51465-9 – ident: ref10 doi: 10.1609/aaai.v31i1.10890 – ident: ref21 doi: 10.1109/TMI.2024.3351213 – ident: ref15 doi: 10.1109/ISBI53787.2023.10230723
SSID	ssj0014509
Score	2.512236
Snippet	As the predominant approach for pathological whole slide image (WSI) classification, multiple instance learning (MIL) methods struggle with limited labeled...
SourceID	proquest pubmed crossref ieee
SourceType	Aggregation Database Index Database Publisher
StartPage	4181
SubjectTerms	Accuracy Algorithms Cancer Data augmentation Data mining Feature extraction Humans Image Interpretation, Computer-Assisted - methods Machine Learning multiple instance learning Multiple-Instance Learning Algorithms Neural Networks, Computer Pathology Phenotypes Probability Training Tumors Whole slide image classification
Title	Disentangled Pseudo-Bag Augmentation for Whole Slide Image Multiple Instance Learning
URI	https://ieeexplore.ieee.org/document/11004164 https://www.ncbi.nlm.nih.gov/pubmed/40366836 https://www.proquest.com/docview/3204326874
Volume	44
WOSCitedRecordID	wos001616144400033&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/eLvHCXMwlV1LT-MwEB5BhVbLgd0tLJRlkVfaC4dAGj8SH8tL20NRpW1Fb1Fsj6tKkK5ow-_HdlzgwmGVSw7OQzMT-3Nmvm8AfitDBTJhE8N9C7MKRVIwmieao9thG4myr0KzifzurpjN5DiS1QMXBhFD8Rme-9OQyzdL3fhfZRdB3szh-23YznPRkrVeUwaMt_UcmZeMTUW2yUmm8mIyGrqdYMbPKRdSMt8dhrmZW7TCzG_LUeiv8jHUDEvO7Zf_fNmvsBexJRm0wfANtrDuwu47xcEufBrFXPo-TK8XgXhUzx_QkPEKG7NMLqs5GTTzx8hIqonDtOTe99Alfx8WBsnw0U1AZBTLEMkwoEuNJOq0zg9genszufqTxCYLic5yvk5soSpqClpJlTHBKym8oj7VyqQOqqC2OjVU58jcYSmTSqQ2NRyZVX3LeJ9-h069rPEISO6wj7IKK-tQTeUs3ddezo9yJaUxyvbgbGPr8l-rpVGGPUgqS-ei0ruojC7qwYE36du4aM0e_Np4p3Qfgs9uVDUum1VJPcs3E0Xuxhy2bnu9euPt4w_u-gM--4e3HMMT6KyfGvwJO_p5vVg9nbpomxWnIdpeAPrszr4
linkProvider	IEEE
linkToHtml	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LT9wwEB7xqEo50JZHWaCtK3HpIUs2fiQ-8hSrsiukLoJbFNvj1UqQrdhNfz-24wUuHFAuOTiRNePY32Tm-wbgUBkqkAmbGO5bmFUokoLRPNEcXYRtJMqeCs0m8uGwuLuT15GsHrgwiBiKz7Drb0Mu30x143-VHQV5M4fvl2GVM5alLV3rOWnAeFvRkXnR2FRki6xkKo9Gg76LBTPepVxIyXx_GOb2btFKM78cSKHDyttgMxw6F5_fOd0vsBHRJTlul8NXWMJ6E9ZfaQ5uwsdBzKZvwc3ZJFCP6vE9GnI9w8ZMk5NqTI6b8UPkJNXEoVpy67vokr_3E4Ok_-C2IDKIhYikH_ClRhKVWsfbcHNxPjq9TGKbhURnOZ8ntlAVNQWtpMqY4JUUXlOfamVSB1ZQW50aqnNk7rKUSSVSmxqOzKqeZbxHd2Clnta4CyR36EdZhZV1uKZylu5pL-hHuZLSGGU78Hth6_Jfq6ZRhigklaVzUeldVEYXdWDbm_RlXLRmB34tvFO6T8HnN6oap82spJ7nm4kid2O-tW57fnrh7b033voT1i5Hg6vyqj_8sw-f_ERaxuEBrMwfG_wOH_T_-WT2-COsuScHHtEd
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=Disentangled+Pseudo-bag+Augmentation+for+Whole+Slide+Image+Multiple+Instance+Learning&rft.jtitle=IEEE+transactions+on+medical+imaging&rft.au=Dong%2C+Jiuyang&rft.au=Jiang%2C+Junjun&rft.au=Jiang%2C+Kui&rft.au=Li%2C+Jiahan&rft.date=2025-11-01&rft.issn=1558-254X&rft.eissn=1558-254X&rft.volume=PP&rft_id=info:doi/10.1109%2FTMI.2025.3569941&rft.externalDBID=NO_FULL_TEXT
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