Leveraging Representations from Intermediate Encoder-blocks for Synthetic Image Detection Article Swipe

PDF

Christos Koutlis , Symeon Papadopoulos ·

YOU? · · 2024 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2402.19091

The recently developed and publicly available synthetic image generation methods and services make it possible to create extremely realistic imagery on demand, raising great risks for the integrity and safety of online information. State-of-the-art Synthetic Image Detection (SID) research has led to strong evidence on the advantages of feature extraction from foundation models. However, such extracted features mostly encapsulate high-level visual semantics instead of fine-grained details, which are more important for the SID task. On the contrary, shallow layers encode low-level visual information. In this work, we leverage the image representations extracted by intermediate Transformer blocks of CLIP's image-encoder via a lightweight network that maps them to a learnable forgery-aware vector space capable of generalizing exceptionally well. We also employ a trainable module to incorporate the importance of each Transformer block to the final prediction. Our method is compared against the state-of-the-art by evaluating it on 20 test datasets and exhibits an average +10.6% absolute performance improvement. Notably, the best performing models require just a single epoch for training (~8 minutes). Code available at https://github.com/mever-team/rine.

Related Topics

Computer Science

Artificial Intelligence

Computer Vision

Concepts

Encoder Computer science Image (mathematics) Artificial intelligence Computer vision Operating system

Metadata

Type: preprint
Language: en
Landing Page: http://arxiv.org/abs/2402.19091
PDF: https://arxiv.org/pdf/2402.19091
OA Status: green
Cited By: 1
Related Works: 10
OpenAlex ID: https://openalex.org/W4400516589

All OpenAlex metadata

Raw OpenAlex JSON

OpenAlex ID: https://openalex.org/W4400516589

Canonical identifier for this work in OpenAlex
DOI: https://doi.org/10.48550/arxiv.2402.19091

Digital Object Identifier
Title: Leveraging Representations from Intermediate Encoder-blocks for Synthetic Image Detection

Work title
Type: preprint

OpenAlex work type
Language: en

Primary language
Publication year: 2024

Year of publication
Publication date: 2024-02-29

Full publication date if available
Authors: Christos Koutlis, Symeon Papadopoulos

List of authors in order
Landing page: https://arxiv.org/abs/2402.19091

Publisher landing page
PDF URL: https://arxiv.org/pdf/2402.19091

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/2402.19091

Direct OA link when available
Concepts: Encoder, Computer science, Image (mathematics), Artificial intelligence, Computer vision, Operating system

Top concepts (fields/topics) attached by OpenAlex
Cited by: 1

Total citation count in OpenAlex
Citations by year (recent): 2025: 1

Per-year citation counts (last 5 years)
Related works (count): 10

Other works algorithmically related by OpenAlex

Full payload

id	https://openalex.org/W4400516589
doi	https://doi.org/10.48550/arxiv.2402.19091
ids.doi	https://doi.org/10.48550/arxiv.2402.19091
ids.openalex	https://openalex.org/W4400516589
fwci
type	preprint
title	Leveraging Representations from Intermediate Encoder-blocks for Synthetic Image Detection
biblio.issue
biblio.volume
biblio.last_page
biblio.first_page
topics[0].id	https://openalex.org/T10775
topics[0].field.id	https://openalex.org/fields/17
topics[0].field.display_name	Computer Science
topics[0].score	0.9070000052452087
topics[0].domain.id	https://openalex.org/domains/3
topics[0].domain.display_name	Physical Sciences
topics[0].subfield.id	https://openalex.org/subfields/1707
topics[0].subfield.display_name	Computer Vision and Pattern Recognition
topics[0].display_name	Generative Adversarial Networks and Image Synthesis
is_xpac	False
apc_list
apc_paid
concepts[0].id	https://openalex.org/C118505674
concepts[0].level	2
concepts[0].score	0.6727032661437988
concepts[0].wikidata	https://www.wikidata.org/wiki/Q42586063
concepts[0].display_name	Encoder
concepts[1].id	https://openalex.org/C41008148
concepts[1].level	0
concepts[1].score	0.6004446148872375
concepts[1].wikidata	https://www.wikidata.org/wiki/Q21198
concepts[1].display_name	Computer science
concepts[2].id	https://openalex.org/C115961682
concepts[2].level	2
concepts[2].score	0.5919526815414429
concepts[2].wikidata	https://www.wikidata.org/wiki/Q860623
concepts[2].display_name	Image (mathematics)
concepts[3].id	https://openalex.org/C154945302
concepts[3].level	1
concepts[3].score	0.5195425152778625
concepts[3].wikidata	https://www.wikidata.org/wiki/Q11660
concepts[3].display_name	Artificial intelligence
concepts[4].id	https://openalex.org/C31972630
concepts[4].level	1
concepts[4].score	0.4439114034175873
concepts[4].wikidata	https://www.wikidata.org/wiki/Q844240
concepts[4].display_name	Computer vision
concepts[5].id	https://openalex.org/C111919701
concepts[5].level	1
concepts[5].score	0.08843263983726501
concepts[5].wikidata	https://www.wikidata.org/wiki/Q9135
concepts[5].display_name	Operating system
keywords[0].id	https://openalex.org/keywords/encoder
keywords[0].score	0.6727032661437988
keywords[0].display_name	Encoder
keywords[1].id	https://openalex.org/keywords/computer-science
keywords[1].score	0.6004446148872375
keywords[1].display_name	Computer science
keywords[2].id	https://openalex.org/keywords/image
keywords[2].score	0.5919526815414429
keywords[2].display_name	Image (mathematics)
keywords[3].id	https://openalex.org/keywords/artificial-intelligence
keywords[3].score	0.5195425152778625
keywords[3].display_name	Artificial intelligence
keywords[4].id	https://openalex.org/keywords/computer-vision
keywords[4].score	0.4439114034175873
keywords[4].display_name	Computer vision
keywords[5].id	https://openalex.org/keywords/operating-system
keywords[5].score	0.08843263983726501
keywords[5].display_name	Operating system
language	en
locations[0].id	pmh:oai:arXiv.org:2402.19091
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/2402.19091
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/2402.19091
locations[1].id	doi:10.48550/arxiv.2402.19091
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.2402.19091
indexed_in	arxiv, datacite
authorships[0].author.id	https://openalex.org/A5024874767
authorships[0].author.orcid	https://orcid.org/0000-0003-3682-408X
authorships[0].author.display_name	Christos Koutlis
authorships[0].author_position	first
authorships[0].raw_author_name	Koutlis, Christos
authorships[0].is_corresponding	False
authorships[1].author.id	https://openalex.org/A5013616365
authorships[1].author.orcid	https://orcid.org/0000-0002-5441-7341
authorships[1].author.display_name	Symeon Papadopoulos
authorships[1].author_position	last
authorships[1].raw_author_name	Papadopoulos, Symeon
authorships[1].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/2402.19091
open_access.oa_status	green
open_access.any_repository_has_fulltext	False
created_date	2024-07-11T00:00:00
display_name	Leveraging Representations from Intermediate Encoder-blocks for Synthetic Image Detection
has_fulltext	False
is_retracted	False
updated_date	2025-11-06T06:51:31.235846
primary_topic.id	https://openalex.org/T10775
primary_topic.field.id	https://openalex.org/fields/17
primary_topic.field.display_name	Computer Science
primary_topic.score	0.9070000052452087
primary_topic.domain.id	https://openalex.org/domains/3
primary_topic.domain.display_name	Physical Sciences
primary_topic.subfield.id	https://openalex.org/subfields/1707
primary_topic.subfield.display_name	Computer Vision and Pattern Recognition
primary_topic.display_name	Generative Adversarial Networks and Image Synthesis
related_works	https://openalex.org/W2058170566, https://openalex.org/W2755342338, https://openalex.org/W2772917594, https://openalex.org/W2775347418, https://openalex.org/W2166024367, https://openalex.org/W3116076068, https://openalex.org/W2229312674, https://openalex.org/W2951359407, https://openalex.org/W2079911747, https://openalex.org/W1969923398
cited_by_count	1
counts_by_year[0].year	2025
counts_by_year[0].cited_by_count	1
locations_count	2
best_oa_location.id	pmh:oai:arXiv.org:2402.19091
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/2402.19091
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/2402.19091
primary_location.id	pmh:oai:arXiv.org:2402.19091
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/2402.19091
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/2402.19091
publication_date	2024-02-29
publication_year	2024
referenced_works_count	0
abstract_inverted_index.a	100, 107, 120, 164
abstract_inverted_index.20	146
abstract_inverted_index.In	83
abstract_inverted_index.On	74
abstract_inverted_index.We	117
abstract_inverted_index.an	151
abstract_inverted_index.at	173
abstract_inverted_index.by	92, 142
abstract_inverted_index.is	137
abstract_inverted_index.it	13, 144
abstract_inverted_index.of	30, 47, 63, 96, 113, 127
abstract_inverted_index.on	20, 44, 145
abstract_inverted_index.to	15, 41, 106, 123, 131
abstract_inverted_index.we	86
abstract_inverted_index.(~8	169
abstract_inverted_index.Our	135
abstract_inverted_index.SID	72
abstract_inverted_index.The	0
abstract_inverted_index.and	3, 10, 28, 149
abstract_inverted_index.are	67
abstract_inverted_index.for	25, 70, 167
abstract_inverted_index.has	39
abstract_inverted_index.led	40
abstract_inverted_index.the	26, 45, 71, 75, 88, 125, 132, 140, 158
abstract_inverted_index.via	99
abstract_inverted_index.Code	171
abstract_inverted_index.also	118
abstract_inverted_index.best	159
abstract_inverted_index.each	128
abstract_inverted_index.from	50
abstract_inverted_index.just	163
abstract_inverted_index.make	12
abstract_inverted_index.maps	104
abstract_inverted_index.more	68
abstract_inverted_index.such	54
abstract_inverted_index.test	147
abstract_inverted_index.that	103
abstract_inverted_index.them	105
abstract_inverted_index.this	84
abstract_inverted_index.(SID)	37
abstract_inverted_index.Image	35
abstract_inverted_index.block	130
abstract_inverted_index.epoch	166
abstract_inverted_index.final	133
abstract_inverted_index.great	23
abstract_inverted_index.image	7, 89
abstract_inverted_index.risks	24
abstract_inverted_index.space	111
abstract_inverted_index.task.	73
abstract_inverted_index.well.	116
abstract_inverted_index.which	66
abstract_inverted_index.work,	85
abstract_inverted_index.+10.6%	153
abstract_inverted_index.CLIP's	97
abstract_inverted_index.blocks	95
abstract_inverted_index.create	16
abstract_inverted_index.employ	119
abstract_inverted_index.encode	79
abstract_inverted_index.layers	78
abstract_inverted_index.method	136
abstract_inverted_index.models	161
abstract_inverted_index.module	122
abstract_inverted_index.mostly	57
abstract_inverted_index.online	31
abstract_inverted_index.safety	29
abstract_inverted_index.single	165
abstract_inverted_index.strong	42
abstract_inverted_index.vector	110
abstract_inverted_index.visual	60, 81
abstract_inverted_index.against	139
abstract_inverted_index.average	152
abstract_inverted_index.capable	112
abstract_inverted_index.demand,	21
abstract_inverted_index.feature	48
abstract_inverted_index.imagery	19
abstract_inverted_index.instead	62
abstract_inverted_index.methods	9
abstract_inverted_index.models.	52
abstract_inverted_index.network	102
abstract_inverted_index.raising	22
abstract_inverted_index.require	162
abstract_inverted_index.shallow	77
abstract_inverted_index.However,	53
abstract_inverted_index.Notably,	157
abstract_inverted_index.absolute	154
abstract_inverted_index.compared	138
abstract_inverted_index.datasets	148
abstract_inverted_index.details,	65
abstract_inverted_index.evidence	43
abstract_inverted_index.exhibits	150
abstract_inverted_index.features	56
abstract_inverted_index.leverage	87
abstract_inverted_index.possible	14
abstract_inverted_index.publicly	4
abstract_inverted_index.recently	1
abstract_inverted_index.research	38
abstract_inverted_index.services	11
abstract_inverted_index.training	168
abstract_inverted_index.Detection	36
abstract_inverted_index.Synthetic	34
abstract_inverted_index.available	5, 172
abstract_inverted_index.contrary,	76
abstract_inverted_index.developed	2
abstract_inverted_index.extracted	55, 91
abstract_inverted_index.extremely	17
abstract_inverted_index.important	69
abstract_inverted_index.integrity	27
abstract_inverted_index.learnable	108
abstract_inverted_index.low-level	80
abstract_inverted_index.minutes).	170
abstract_inverted_index.realistic	18
abstract_inverted_index.semantics	61
abstract_inverted_index.synthetic	6
abstract_inverted_index.trainable	121
abstract_inverted_index.advantages	46
abstract_inverted_index.evaluating	143
abstract_inverted_index.extraction	49
abstract_inverted_index.foundation	51
abstract_inverted_index.generation	8
abstract_inverted_index.high-level	59
abstract_inverted_index.importance	126
abstract_inverted_index.performing	160
abstract_inverted_index.Transformer	94, 129
abstract_inverted_index.encapsulate	58
abstract_inverted_index.incorporate	124
abstract_inverted_index.lightweight	101
abstract_inverted_index.performance	155
abstract_inverted_index.prediction.	134
abstract_inverted_index.fine-grained	64
abstract_inverted_index.generalizing	114
abstract_inverted_index.improvement.	156
abstract_inverted_index.information.	32, 82
abstract_inverted_index.intermediate	93
abstract_inverted_index.exceptionally	115
abstract_inverted_index.forgery-aware	109
abstract_inverted_index.image-encoder	98
abstract_inverted_index.representations	90
abstract_inverted_index.State-of-the-art	33
abstract_inverted_index.state-of-the-art	141
abstract_inverted_index.https://github.com/mever-team/rine.	174
cited_by_percentile_year
countries_distinct_count	0
institutions_distinct_count	2
citation_normalized_percentile