LPD: Learnable Prototypes with Diversity Regularization for Weakly Supervised Histopathology Segmentation Article Swipe

PDF

Khang Le , Tuan A. Vu , Ha Thach , Ngoc Bui Lam Quang , Thanh-Huy Nguyen , Minh Hoang Le , Zhu Han , Chandra Mohan , Hien Van Nguyen ·

YOU? · · 2025 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2512.05922

Weakly supervised semantic segmentation (WSSS) in histopathology reduces pixel-level labeling by learning from image-level labels, but it is hindered by inter-class homogeneity, intra-class heterogeneity, and CAM-induced region shrinkage (global pooling-based class activation maps whose activations highlight only the most distinctive areas and miss nearby class regions). Recent works address these challenges by constructing a clustering prototype bank and then refining masks in a separate stage; however, such two-stage pipelines are costly, sensitive to hyperparameters, and decouple prototype discovery from segmentation learning, limiting their effectiveness and efficiency. We propose a cluster-free, one-stage learnable-prototype framework with diversity regularization to enhance morphological intra-class heterogeneity coverage. Our approach achieves state-of-the-art (SOTA) performance on BCSS-WSSS, outperforming prior methods in mIoU and mDice. Qualitative segmentation maps show sharper boundaries and fewer mislabels, and activation heatmaps further reveal that, compared with clustering-based prototypes, our learnable prototypes cover more diverse and complementary regions within each class, providing consistent qualitative evidence for their effectiveness.

Related Topics

Fire With Fire (2012 Film)

Die With A Smile

My Life With The Walter Boys

Dementia With Lewy Bodies

List Of Actors With Two Or More Academy Awards In Acting Categories

Bestiality With A Donkey

How To With John Wilson

Late Night With Conan O'Brien

Lucy In The Sky With Diamonds

The Man With The Iron Fists

What's Love Got To Do With It (Song)

The House With A Clock In Its Walls (Film)

Concepts

No concepts available.

Metadata

Type: preprint
Landing Page: http://arxiv.org/abs/2512.05922
PDF: https://arxiv.org/pdf/2512.05922
OA Status: green
OpenAlex ID: https://openalex.org/W4417144663

All OpenAlex metadata

Raw OpenAlex JSON

OpenAlex ID: https://openalex.org/W4417144663

Canonical identifier for this work in OpenAlex
DOI: https://doi.org/10.48550/arxiv.2512.05922

Digital Object Identifier
Title: LPD: Learnable Prototypes with Diversity Regularization for Weakly Supervised Histopathology Segmentation

Work title
Type: preprint

OpenAlex work type
Publication year: 2025

Year of publication
Publication date: 2025-12-05

Full publication date if available
Authors: Khang Le, Tuan A. Vu, Ha Thach, Ngoc Bui Lam Quang, Thanh-Huy Nguyen, Minh Hoang Le, Zhu Han, Chandra Mohan, Hien Van Nguyen

List of authors in order
Landing page: https://arxiv.org/abs/2512.05922

Publisher landing page
PDF URL: https://arxiv.org/pdf/2512.05922

Direct link to full text PDF
Open access: Yes

Whether a free full text is available
OA status: green

Open access status per OpenAlex
OA URL: https://arxiv.org/pdf/2512.05922

Direct OA link when available
Cited by: 0

Total citation count in OpenAlex

Full payload

id	https://openalex.org/W4417144663
doi	https://doi.org/10.48550/arxiv.2512.05922
ids.doi	https://doi.org/10.48550/arxiv.2512.05922
ids.openalex	https://openalex.org/W4417144663
fwci
type	preprint
title	LPD: Learnable Prototypes with Diversity Regularization for Weakly Supervised Histopathology Segmentation
biblio.issue
biblio.volume
biblio.last_page
biblio.first_page
is_xpac	False
apc_list
apc_paid
language
locations[0].id	pmh:oai:arXiv.org:2512.05922
locations[0].is_oa	True
locations[0].source.id	https://openalex.org/S4306400194
locations[0].source.issn
locations[0].source.type	repository
locations[0].source.is_oa	True
locations[0].source.issn_l
locations[0].source.is_core	False
locations[0].source.is_in_doaj	False
locations[0].source.display_name	arXiv (Cornell University)
locations[0].source.host_organization	https://openalex.org/I205783295
locations[0].source.host_organization_name	Cornell University
locations[0].source.host_organization_lineage	https://openalex.org/I205783295
locations[0].license
locations[0].pdf_url	https://arxiv.org/pdf/2512.05922
locations[0].version	submittedVersion
locations[0].raw_type	text
locations[0].license_id
locations[0].is_accepted	False
locations[0].is_published	False
locations[0].raw_source_name
locations[0].landing_page_url	http://arxiv.org/abs/2512.05922
locations[1].id	doi:10.48550/arxiv.2512.05922
locations[1].is_oa	True
locations[1].source.id	https://openalex.org/S4306400194
locations[1].source.issn
locations[1].source.type	repository
locations[1].source.is_oa	True
locations[1].source.issn_l
locations[1].source.is_core	False
locations[1].source.is_in_doaj	False
locations[1].source.display_name	arXiv (Cornell University)
locations[1].source.host_organization	https://openalex.org/I205783295
locations[1].source.host_organization_name	Cornell University
locations[1].source.host_organization_lineage	https://openalex.org/I205783295
locations[1].license
locations[1].pdf_url
locations[1].version
locations[1].raw_type	article
locations[1].license_id
locations[1].is_accepted	False
locations[1].is_published
locations[1].raw_source_name
locations[1].landing_page_url	https://doi.org/10.48550/arxiv.2512.05922
indexed_in	arxiv, datacite
authorships[0].author.id	https://openalex.org/A5104006945
authorships[0].author.orcid
authorships[0].author.display_name	Khang Le
authorships[0].author_position	first
authorships[0].raw_author_name	Le, Khang
authorships[0].is_corresponding	False
authorships[1].author.id	https://openalex.org/A5041664690
authorships[1].author.orcid	https://orcid.org/0000-0002-5546-4422
authorships[1].author.display_name	Tuan A. Vu
authorships[1].author_position	middle
authorships[1].raw_author_name	Vu, Anh Mai
authorships[1].is_corresponding	False
authorships[2].author.id	https://openalex.org/A5030294531
authorships[2].author.orcid
authorships[2].author.display_name	Ha Thach
authorships[2].author_position	middle
authorships[2].raw_author_name	Thach, Ha
authorships[2].is_corresponding	False
authorships[3].author.id	https://openalex.org/A5101173242
authorships[3].author.orcid
authorships[3].author.display_name	Ngoc Bui Lam Quang
authorships[3].author_position	middle
authorships[3].raw_author_name	Quang, Ngoc Bui Lam
authorships[3].is_corresponding	False
authorships[4].author.id	https://openalex.org/A5103260009
authorships[4].author.orcid	https://orcid.org/0009-0003-9355-5434
authorships[4].author.display_name	Thanh-Huy Nguyen
authorships[4].author_position	middle
authorships[4].raw_author_name	Nguyen, Thanh-Huy
authorships[4].is_corresponding	False
authorships[5].author.id	https://openalex.org/A5048194853
authorships[5].author.orcid	https://orcid.org/0000-0003-0908-1188
authorships[5].author.display_name	Minh Hoang Le
authorships[5].author_position	middle
authorships[5].raw_author_name	Le, Minh H. N.
authorships[5].is_corresponding	False
authorships[6].author.id	https://openalex.org/A5063667378
authorships[6].author.orcid	https://orcid.org/0000-0002-6606-5822
authorships[6].author.display_name	Zhu Han
authorships[6].author_position	middle
authorships[6].raw_author_name	Han, Zhu
authorships[6].is_corresponding	False
authorships[7].author.id	https://openalex.org/A5042585597
authorships[7].author.orcid	https://orcid.org/0000-0001-7896-5740
authorships[7].author.display_name	Chandra Mohan
authorships[7].author_position	last
authorships[7].raw_author_name	Mohan, Chandra
authorships[7].is_corresponding	False
authorships[8].author.id	https://openalex.org/A5101985554
authorships[8].author.orcid	https://orcid.org/0000-0001-7280-2182
authorships[8].author.display_name	Hien Van Nguyen
authorships[8].author_position	middle
authorships[8].raw_author_name	Van Nguyen, Hien
authorships[8].is_corresponding	False
has_content.pdf	False
has_content.grobid_xml	False
is_paratext	False
open_access.is_oa	True
open_access.oa_url	https://arxiv.org/pdf/2512.05922
open_access.oa_status	green
open_access.any_repository_has_fulltext	False
created_date	2025-12-09T00:00:00
display_name	LPD: Learnable Prototypes with Diversity Regularization for Weakly Supervised Histopathology Segmentation
has_fulltext	False
is_retracted	False
updated_date	2025-12-10T02:49:46.989445
primary_topic
cited_by_count	0
locations_count	2
best_oa_location.id	pmh:oai:arXiv.org:2512.05922
best_oa_location.is_oa	True
best_oa_location.source.id	https://openalex.org/S4306400194
best_oa_location.source.issn
best_oa_location.source.type	repository
best_oa_location.source.is_oa	True
best_oa_location.source.issn_l
best_oa_location.source.is_core	False
best_oa_location.source.is_in_doaj	False
best_oa_location.source.display_name	arXiv (Cornell University)
best_oa_location.source.host_organization	https://openalex.org/I205783295
best_oa_location.source.host_organization_name	Cornell University
best_oa_location.source.host_organization_lineage	https://openalex.org/I205783295
best_oa_location.license
best_oa_location.pdf_url	https://arxiv.org/pdf/2512.05922
best_oa_location.version	submittedVersion
best_oa_location.raw_type	text
best_oa_location.license_id
best_oa_location.is_accepted	False
best_oa_location.is_published	False
best_oa_location.raw_source_name
best_oa_location.landing_page_url	http://arxiv.org/abs/2512.05922
primary_location.id	pmh:oai:arXiv.org:2512.05922
primary_location.is_oa	True
primary_location.source.id	https://openalex.org/S4306400194
primary_location.source.issn
primary_location.source.type	repository
primary_location.source.is_oa	True
primary_location.source.issn_l
primary_location.source.is_core	False
primary_location.source.is_in_doaj	False
primary_location.source.display_name	arXiv (Cornell University)
primary_location.source.host_organization	https://openalex.org/I205783295
primary_location.source.host_organization_name	Cornell University
primary_location.source.host_organization_lineage	https://openalex.org/I205783295
primary_location.license
primary_location.pdf_url	https://arxiv.org/pdf/2512.05922
primary_location.version	submittedVersion
primary_location.raw_type	text
primary_location.license_id
primary_location.is_accepted	False
primary_location.is_published	False
primary_location.raw_source_name
primary_location.landing_page_url	http://arxiv.org/abs/2512.05922
publication_date	2025-12-05
publication_year	2025
referenced_works_count	0
abstract_inverted_index.a	53, 62, 88
abstract_inverted_index.We	86
abstract_inverted_index.by	10, 19, 51
abstract_inverted_index.in	5, 61, 113
abstract_inverted_index.is	17
abstract_inverted_index.it	16
abstract_inverted_index.on	108
abstract_inverted_index.to	72, 96
abstract_inverted_index.Our	102
abstract_inverted_index.and	24, 41, 57, 74, 84, 115, 123, 126, 142
abstract_inverted_index.are	69
abstract_inverted_index.but	15
abstract_inverted_index.for	152
abstract_inverted_index.our	136
abstract_inverted_index.the	37
abstract_inverted_index.bank	56
abstract_inverted_index.each	146
abstract_inverted_index.from	12, 78
abstract_inverted_index.mIoU	114
abstract_inverted_index.maps	32, 119
abstract_inverted_index.miss	42
abstract_inverted_index.more	140
abstract_inverted_index.most	38
abstract_inverted_index.only	36
abstract_inverted_index.show	120
abstract_inverted_index.such	66
abstract_inverted_index.then	58
abstract_inverted_index.with	93, 133
abstract_inverted_index.areas	40
abstract_inverted_index.class	30, 44
abstract_inverted_index.cover	139
abstract_inverted_index.fewer	124
abstract_inverted_index.masks	60
abstract_inverted_index.prior	111
abstract_inverted_index.that,	131
abstract_inverted_index.their	82, 153
abstract_inverted_index.these	49
abstract_inverted_index.whose	33
abstract_inverted_index.works	47
abstract_inverted_index.(SOTA)	106
abstract_inverted_index.(WSSS)	4
abstract_inverted_index.Recent	46
abstract_inverted_index.Weakly	0
abstract_inverted_index.class,	147
abstract_inverted_index.mDice.	116
abstract_inverted_index.nearby	43
abstract_inverted_index.region	26
abstract_inverted_index.reveal	130
abstract_inverted_index.stage;	64
abstract_inverted_index.within	145
abstract_inverted_index.(global	28
abstract_inverted_index.address	48
abstract_inverted_index.costly,	70
abstract_inverted_index.diverse	141
abstract_inverted_index.enhance	97
abstract_inverted_index.further	129
abstract_inverted_index.labels,	14
abstract_inverted_index.methods	112
abstract_inverted_index.propose	87
abstract_inverted_index.reduces	7
abstract_inverted_index.regions	144
abstract_inverted_index.sharper	121
abstract_inverted_index.achieves	104
abstract_inverted_index.approach	103
abstract_inverted_index.compared	132
abstract_inverted_index.decouple	75
abstract_inverted_index.evidence	151
abstract_inverted_index.heatmaps	128
abstract_inverted_index.hindered	18
abstract_inverted_index.however,	65
abstract_inverted_index.labeling	9
abstract_inverted_index.learning	11
abstract_inverted_index.limiting	81
abstract_inverted_index.refining	59
abstract_inverted_index.semantic	2
abstract_inverted_index.separate	63
abstract_inverted_index.coverage.	101
abstract_inverted_index.discovery	77
abstract_inverted_index.diversity	94
abstract_inverted_index.framework	92
abstract_inverted_index.highlight	35
abstract_inverted_index.learnable	137
abstract_inverted_index.learning,	80
abstract_inverted_index.one-stage	90
abstract_inverted_index.pipelines	68
abstract_inverted_index.prototype	55, 76
abstract_inverted_index.providing	148
abstract_inverted_index.regions).	45
abstract_inverted_index.sensitive	71
abstract_inverted_index.shrinkage	27
abstract_inverted_index.two-stage	67
abstract_inverted_index.BCSS-WSSS,	109
abstract_inverted_index.activation	31, 127
abstract_inverted_index.boundaries	122
abstract_inverted_index.challenges	50
abstract_inverted_index.clustering	54
abstract_inverted_index.consistent	149
abstract_inverted_index.mislabels,	125
abstract_inverted_index.prototypes	138
abstract_inverted_index.supervised	1
abstract_inverted_index.CAM-induced	25
abstract_inverted_index.Qualitative	117
abstract_inverted_index.activations	34
abstract_inverted_index.distinctive	39
abstract_inverted_index.efficiency.	85
abstract_inverted_index.image-level	13
abstract_inverted_index.inter-class	20
abstract_inverted_index.intra-class	22, 99
abstract_inverted_index.performance	107
abstract_inverted_index.pixel-level	8
abstract_inverted_index.prototypes,	135
abstract_inverted_index.qualitative	150
abstract_inverted_index.constructing	52
abstract_inverted_index.homogeneity,	21
abstract_inverted_index.segmentation	3, 79, 118
abstract_inverted_index.cluster-free,	89
abstract_inverted_index.complementary	143
abstract_inverted_index.effectiveness	83
abstract_inverted_index.heterogeneity	100
abstract_inverted_index.morphological	98
abstract_inverted_index.outperforming	110
abstract_inverted_index.pooling-based	29
abstract_inverted_index.effectiveness.	154
abstract_inverted_index.heterogeneity,	23
abstract_inverted_index.histopathology	6
abstract_inverted_index.regularization	95
abstract_inverted_index.clustering-based	134
abstract_inverted_index.hyperparameters,	73
abstract_inverted_index.state-of-the-art	105
abstract_inverted_index.learnable-prototype	91
cited_by_percentile_year
countries_distinct_count	0
institutions_distinct_count	9
citation_normalized_percentile