Object-level Correlation for Few-Shot Segmentation Article Swipe

PDF

Chunlin Wen , Yu Zhang , Jie Fan , Hongyuan Zhu , Xiu-Shen Wei , Yijun Wang , Zhiqiang Kou , Shuzhou Sun ·

YOU? · · 2025 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2509.07917

Few-shot semantic segmentation (FSS) aims to segment objects of novel categories in the query images given only a few annotated support samples. Existing methods primarily build the image-level correlation between the support target object and the entire query image. However, this correlation contains the hard pixel noise, \textit{i.e.}, irrelevant background objects, that is intractable to trace and suppress, leading to the overfitting of the background. To address the limitation of this correlation, we imitate the biological vision process to identify novel objects in the object-level information. Target identification in the general objects is more valid than in the entire image, especially in the low-data regime. Inspired by this, we design an Object-level Correlation Network (OCNet) by establishing the object-level correlation between the support target object and query general objects, which is mainly composed of the General Object Mining Module (GOMM) and Correlation Construction Module (CCM). Specifically, GOMM constructs the query general object feature by learning saliency and high-level similarity cues, where the general objects include the irrelevant background objects and the target foreground object. Then, CCM establishes the object-level correlation by allocating the target prototypes to match the general object feature. The generated object-level correlation can mine the query target feature and suppress the hard pixel noise for the final prediction. Extensive experiments on PASCAL-${5}^{i}$ and COCO-${20}^{i}$ show that our model achieves the state-of-the-art performance.

Related Topics

Concepts

No concepts available.

Metadata

Type: preprint
Language: en
Landing Page: http://arxiv.org/abs/2509.07917
PDF: https://arxiv.org/pdf/2509.07917
OA Status: green
OpenAlex ID: https://openalex.org/W4416068630

All OpenAlex metadata

Raw OpenAlex JSON

OpenAlex ID: https://openalex.org/W4416068630

Canonical identifier for this work in OpenAlex
DOI: https://doi.org/10.48550/arxiv.2509.07917

Digital Object Identifier
Title: Object-level Correlation for Few-Shot Segmentation

Work title
Type: preprint

OpenAlex work type
Language: en

Primary language
Publication year: 2025

Year of publication
Publication date: 2025-09-09

Full publication date if available
Authors: Chunlin Wen, Yu Zhang, Jie Fan, Hongyuan Zhu, Xiu-Shen Wei, Yijun Wang, Zhiqiang Kou, Shuzhou Sun

List of authors in order
Landing page: https://arxiv.org/abs/2509.07917

Publisher landing page
PDF URL: https://arxiv.org/pdf/2509.07917

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/2509.07917

Direct OA link when available
Cited by: 0

Total citation count in OpenAlex

Full payload

id	https://openalex.org/W4416068630
doi	https://doi.org/10.48550/arxiv.2509.07917
ids.doi	https://doi.org/10.48550/arxiv.2509.07917
ids.openalex	https://openalex.org/W4416068630
fwci
type	preprint
title	Object-level Correlation for Few-Shot Segmentation
biblio.issue
biblio.volume
biblio.last_page
biblio.first_page
is_xpac	False
apc_list
apc_paid
language	en
locations[0].id	pmh:oai:arXiv.org:2509.07917
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/2509.07917
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/2509.07917
locations[1].id	doi:10.48550/arxiv.2509.07917
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.2509.07917
indexed_in	arxiv, datacite
authorships[0].author.id	https://openalex.org/A5101319199
authorships[0].author.orcid	https://orcid.org/0009-0001-5329-196X
authorships[0].author.display_name	Chunlin Wen
authorships[0].author_position	first
authorships[0].raw_author_name	Wen, Chunlin
authorships[0].is_corresponding	False
authorships[1].author.id	https://openalex.org/A5100678774
authorships[1].author.orcid	https://orcid.org/0009-0005-7730-8817
authorships[1].author.display_name	Yu Zhang
authorships[1].author_position	middle
authorships[1].raw_author_name	Zhang, Yu
authorships[1].is_corresponding	False
authorships[2].author.id	https://openalex.org/A5045040746
authorships[2].author.orcid	https://orcid.org/0000-0003-4330-6053
authorships[2].author.display_name	Jie Fan
authorships[2].author_position	middle
authorships[2].raw_author_name	Fan, Jie
authorships[2].is_corresponding	False
authorships[3].author.id	https://openalex.org/A5102483130
authorships[3].author.orcid
authorships[3].author.display_name	Hongyuan Zhu
authorships[3].author_position	middle
authorships[3].raw_author_name	Zhu, Hongyuan
authorships[3].is_corresponding	False
authorships[4].author.id	https://openalex.org/A5066964304
authorships[4].author.orcid	https://orcid.org/0000-0002-8200-1845
authorships[4].author.display_name	Xiu-Shen Wei
authorships[4].author_position	middle
authorships[4].raw_author_name	Wei, Xiu-Shen
authorships[4].is_corresponding	False
authorships[5].author.id	https://openalex.org/A5100713695
authorships[5].author.orcid	https://orcid.org/0000-0002-8161-2150
authorships[5].author.display_name	Yijun Wang
authorships[5].author_position	middle
authorships[5].raw_author_name	Wang, Yijun
authorships[5].is_corresponding	False
authorships[6].author.id	https://openalex.org/A5089186967
authorships[6].author.orcid	https://orcid.org/0000-0002-3335-3959
authorships[6].author.display_name	Zhiqiang Kou
authorships[6].author_position	middle
authorships[6].raw_author_name	Kou, Zhiqiang
authorships[6].is_corresponding	False
authorships[7].author.id	https://openalex.org/A5065503504
authorships[7].author.orcid	https://orcid.org/0000-0002-6104-8046
authorships[7].author.display_name	Shuzhou Sun
authorships[7].author_position	last
authorships[7].raw_author_name	Sun, Shuzhou
authorships[7].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/2509.07917
open_access.oa_status	green
open_access.any_repository_has_fulltext	False
created_date	2025-10-10T00:00:00
display_name	Object-level Correlation for Few-Shot Segmentation
has_fulltext	False
is_retracted	False
updated_date	2025-11-28T10:04:09.505741
primary_topic
cited_by_count	0
locations_count	2
best_oa_location.id	pmh:oai:arXiv.org:2509.07917
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/2509.07917
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/2509.07917
primary_location.id	pmh:oai:arXiv.org:2509.07917
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/2509.07917
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/2509.07917
publication_date	2025-09-09
publication_year	2025
referenced_works_count	0
abstract_inverted_index.a	17
abstract_inverted_index.To	65
abstract_inverted_index.an	110
abstract_inverted_index.by	106, 115, 153, 180
abstract_inverted_index.in	11, 82, 88, 96, 101
abstract_inverted_index.is	52, 92, 130
abstract_inverted_index.of	8, 62, 69, 133
abstract_inverted_index.on	213
abstract_inverted_index.to	5, 54, 59, 78, 185
abstract_inverted_index.we	72, 108
abstract_inverted_index.CCM	175
abstract_inverted_index.The	191
abstract_inverted_index.and	34, 56, 125, 140, 156, 169, 201, 215
abstract_inverted_index.can	195
abstract_inverted_index.few	18
abstract_inverted_index.for	207
abstract_inverted_index.our	219
abstract_inverted_index.the	12, 26, 30, 35, 43, 60, 63, 67, 74, 83, 89, 97, 102, 117, 121, 134, 148, 161, 165, 170, 177, 182, 187, 197, 203, 208, 222
abstract_inverted_index.GOMM	146
abstract_inverted_index.aims	4
abstract_inverted_index.hard	44, 204
abstract_inverted_index.mine	196
abstract_inverted_index.more	93
abstract_inverted_index.only	16
abstract_inverted_index.show	217
abstract_inverted_index.than	95
abstract_inverted_index.that	51, 218
abstract_inverted_index.this	40, 70
abstract_inverted_index.(FSS)	3
abstract_inverted_index.Then,	174
abstract_inverted_index.build	25
abstract_inverted_index.cues,	159
abstract_inverted_index.final	209
abstract_inverted_index.given	15
abstract_inverted_index.match	186
abstract_inverted_index.model	220
abstract_inverted_index.noise	206
abstract_inverted_index.novel	9, 80
abstract_inverted_index.pixel	45, 205
abstract_inverted_index.query	13, 37, 126, 149, 198
abstract_inverted_index.this,	107
abstract_inverted_index.trace	55
abstract_inverted_index.valid	94
abstract_inverted_index.where	160
abstract_inverted_index.which	129
abstract_inverted_index.(CCM).	144
abstract_inverted_index.(GOMM)	139
abstract_inverted_index.Mining	137
abstract_inverted_index.Module	138, 143
abstract_inverted_index.Object	136
abstract_inverted_index.Target	86
abstract_inverted_index.design	109
abstract_inverted_index.entire	36, 98
abstract_inverted_index.image,	99
abstract_inverted_index.image.	38
abstract_inverted_index.images	14
abstract_inverted_index.mainly	131
abstract_inverted_index.noise,	46
abstract_inverted_index.object	33, 124, 151, 189
abstract_inverted_index.target	32, 123, 171, 183, 199
abstract_inverted_index.vision	76
abstract_inverted_index.(OCNet)	114
abstract_inverted_index.General	135
abstract_inverted_index.Network	113
abstract_inverted_index.address	66
abstract_inverted_index.between	29, 120
abstract_inverted_index.feature	152, 200
abstract_inverted_index.general	90, 127, 150, 162, 188
abstract_inverted_index.imitate	73
abstract_inverted_index.include	164
abstract_inverted_index.leading	58
abstract_inverted_index.methods	23
abstract_inverted_index.object.	173
abstract_inverted_index.objects	7, 81, 91, 163, 168
abstract_inverted_index.process	77
abstract_inverted_index.regime.	104
abstract_inverted_index.segment	6
abstract_inverted_index.support	20, 31, 122
abstract_inverted_index.Existing	22
abstract_inverted_index.Few-shot	0
abstract_inverted_index.However,	39
abstract_inverted_index.Inspired	105
abstract_inverted_index.achieves	221
abstract_inverted_index.composed	132
abstract_inverted_index.contains	42
abstract_inverted_index.feature.	190
abstract_inverted_index.identify	79
abstract_inverted_index.learning	154
abstract_inverted_index.low-data	103
abstract_inverted_index.objects,	50, 128
abstract_inverted_index.saliency	155
abstract_inverted_index.samples.	21
abstract_inverted_index.semantic	1
abstract_inverted_index.suppress	202
abstract_inverted_index.Extensive	211
abstract_inverted_index.annotated	19
abstract_inverted_index.generated	192
abstract_inverted_index.primarily	24
abstract_inverted_index.suppress,	57
abstract_inverted_index.allocating	181
abstract_inverted_index.background	49, 167
abstract_inverted_index.biological	75
abstract_inverted_index.categories	10
abstract_inverted_index.constructs	147
abstract_inverted_index.especially	100
abstract_inverted_index.foreground	172
abstract_inverted_index.high-level	157
abstract_inverted_index.irrelevant	48, 166
abstract_inverted_index.limitation	68
abstract_inverted_index.prototypes	184
abstract_inverted_index.similarity	158
abstract_inverted_index.Correlation	112, 141
abstract_inverted_index.background.	64
abstract_inverted_index.correlation	28, 41, 119, 179, 194
abstract_inverted_index.establishes	176
abstract_inverted_index.experiments	212
abstract_inverted_index.image-level	27
abstract_inverted_index.intractable	53
abstract_inverted_index.overfitting	61
abstract_inverted_index.prediction.	210
abstract_inverted_index.Construction	142
abstract_inverted_index.Object-level	111
abstract_inverted_index.correlation,	71
abstract_inverted_index.establishing	116
abstract_inverted_index.information.	85
abstract_inverted_index.object-level	84, 118, 178, 193
abstract_inverted_index.performance.	224
abstract_inverted_index.segmentation	2
abstract_inverted_index.Specifically,	145
abstract_inverted_index.\textit{i.e.},	47
abstract_inverted_index.identification	87
abstract_inverted_index.COCO-${20}^{i}$	216
abstract_inverted_index.PASCAL-${5}^{i}$	214
abstract_inverted_index.state-of-the-art	223
cited_by_percentile_year
countries_distinct_count	0
institutions_distinct_count	8
citation_normalized_percentile