{"id":"https://openalex.org/W7158481166","doi":"https://doi.org/10.48550/arxiv.2604.26772","title":"TAP into the Patch Tokens: Leveraging Vision Foundation Model Features for AI-Generated Image Detection","display_name":"TAP into the Patch Tokens: Leveraging Vision Foundation Model Features for AI-Generated Image Detection","publication_year":2026,"publication_date":"2026-04-29","ids":{"openalex":"https://openalex.org/W7158481166","doi":"https://doi.org/10.48550/arxiv.2604.26772"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.26772","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.26772","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2604.26772","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5001042076","display_name":"Ahmed Abdullah","orcid":"https://orcid.org/0000-0003-0941-0404"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Abdullah, Ahmed","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5051530011","display_name":"Nikolas Ebert","orcid":"https://orcid.org/0000-0002-2971-1012"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ebert, Nikolas","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5020732874","display_name":"Oliver Wasenm\u00fcller","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wasenm\u00fcller, Oliver","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12357","display_name":"Digital Media Forensic Detection","score":0.7509999871253967,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T12357","display_name":"Digital Media Forensic Detection","score":0.7509999871253967,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.07010000199079514,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.048900000751018524,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/leverage","display_name":"Leverage (statistics)","score":0.5565000176429749},{"id":"https://openalex.org/keywords/pooling","display_name":"Pooling","score":0.5160999894142151},{"id":"https://openalex.org/keywords/object-detection","display_name":"Object detection","score":0.49570000171661377},{"id":"https://openalex.org/keywords/robustness","display_name":"Robustness (evolution)","score":0.4767000079154968},{"id":"https://openalex.org/keywords/classifier","display_name":"Classifier (UML)","score":0.41920000314712524},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.38609999418258667},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.38600000739097595},{"id":"https://openalex.org/keywords/feature-extraction","display_name":"Feature extraction","score":0.37610000371932983},{"id":"https://openalex.org/keywords/generative-model","display_name":"Generative model","score":0.36890000104904175}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7332000136375427},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.7264999747276306},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.5565000176429749},{"id":"https://openalex.org/C70437156","wikidata":"https://www.wikidata.org/wiki/Q7228652","display_name":"Pooling","level":2,"score":0.5160999894142151},{"id":"https://openalex.org/C2776151529","wikidata":"https://www.wikidata.org/wiki/Q3045304","display_name":"Object detection","level":3,"score":0.49570000171661377},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4771000146865845},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.4767000079154968},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.4544999897480011},{"id":"https://openalex.org/C95623464","wikidata":"https://www.wikidata.org/wiki/Q1096149","display_name":"Classifier (UML)","level":2,"score":0.41920000314712524},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.38609999418258667},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.38600000739097595},{"id":"https://openalex.org/C52622490","wikidata":"https://www.wikidata.org/wiki/Q1026626","display_name":"Feature extraction","level":2,"score":0.37610000371932983},{"id":"https://openalex.org/C167966045","wikidata":"https://www.wikidata.org/wiki/Q5532625","display_name":"Generative model","level":3,"score":0.36890000104904175},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.3472000062465668},{"id":"https://openalex.org/C2987933465","wikidata":"https://www.wikidata.org/wiki/Q141130","display_name":"Image manipulation","level":3,"score":0.33959999680519104},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.3393999934196472},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.3384999930858612},{"id":"https://openalex.org/C9417928","wikidata":"https://www.wikidata.org/wiki/Q1070689","display_name":"Image processing","level":3,"score":0.3100000023841858},{"id":"https://openalex.org/C46686674","wikidata":"https://www.wikidata.org/wiki/Q466303","display_name":"Boosting (machine learning)","level":2,"score":0.29840001463890076},{"id":"https://openalex.org/C2780156472","wikidata":"https://www.wikidata.org/wiki/Q2355550","display_name":"Pedestrian detection","level":3,"score":0.2897000014781952},{"id":"https://openalex.org/C5339829","wikidata":"https://www.wikidata.org/wiki/Q1425977","display_name":"Machine vision","level":2,"score":0.2851000130176544},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.2842999994754791},{"id":"https://openalex.org/C86251818","wikidata":"https://www.wikidata.org/wiki/Q816754","display_name":"Benchmarking","level":2,"score":0.26930001378059387},{"id":"https://openalex.org/C75294576","wikidata":"https://www.wikidata.org/wiki/Q5165192","display_name":"Contextual image classification","level":3,"score":0.26750001311302185},{"id":"https://openalex.org/C59404180","wikidata":"https://www.wikidata.org/wiki/Q17013334","display_name":"Feature learning","level":2,"score":0.25360000133514404}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.26772","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.26772","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2604.26772","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.26772","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/16","display_name":"Peace, Justice and strong institutions","score":0.48304879665374756}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Recent":[0],"methods":[1,28],"demonstrate":[2],"that":[3,110],"large-scale":[4],"pretrained":[5],"models,":[6],"such":[7],"as":[8,23],"CLIP":[9,117],"vision":[10,46],"transformers,":[11],"effectively":[12],"detect":[13],"AI-generated":[14,30,190],"images":[15],"(AIGIs)":[16],"from":[17],"unseen":[18],"generative":[19],"models":[20,48],"when":[21],"used":[22],"feature":[24],"extractors.":[25],"Many":[26],"state-of-the-art":[27,181],"for":[29,64,101,186],"image":[31,69],"detection":[32,66,176,188],"build":[33],"upon":[34],"the":[35,111,115,128,133,145,166],"original":[36,116],"CLIP-ViT":[37],"to":[38],"enhance":[39],"this":[40,75],"generalization.":[41],"Since":[42],"CLIP's":[43],"release,":[44],"numerous":[45],"foundation":[47],"(VFMs)":[49],"have":[50],"emerged,":[51],"incorporating":[52],"architectural":[53],"improvements":[54],"and":[55,67,92,105,108,191],"different":[56],"training":[57],"paradigms.":[58],"Despite":[59],"these":[60],"advances,":[61],"their":[62,98],"potential":[63],"AIGI":[65,175],"AI":[68],"forensics":[70],"remains":[71],"largely":[72],"unexplored.":[73],"In":[74],"work,":[76],"we":[77,139],"present":[78],"a":[79,136,141,159,179],"comprehensive":[80],"benchmark":[81],"across":[82,173],"multiple":[83],"VFM":[84],"families,":[85],"covering":[86],"diverse":[87],"pretraining":[88],"objectives,":[89],"input":[90],"resolutions,":[91],"model":[93,113],"scales.":[94],"We":[95],"systematically":[96],"evaluate":[97],"out-of-the-box":[99],"performance":[100,171],"detecting":[102],"fully-generated":[103],"AI-images":[104],"AI-inpainted":[106],"images,":[107],"discover":[109],"best":[112],"outperforms":[114],"by":[118,148],"more":[119],"than":[120],"12%":[121],"in":[122,127],"accuracy,":[123],"beating":[124],"established":[125],"approaches":[126],"process.":[129],"To":[130],"fully":[131],"leverage":[132],"features":[134],"of":[135,144,189],"modern":[137],"VFM,":[138],"propose":[140],"simple":[142],"redesign":[143],"classifier":[146],"head":[147],"utilizing":[149],"tunable":[150],"attention":[151],"pooling":[152],"(TAP),":[153],"which":[154],"aggregates":[155],"output":[156],"tokens":[157],"into":[158],"refined":[160],"global":[161],"representation.":[162],"Integrating":[163],"TAP":[164],"with":[165],"latest":[167],"VFMs":[168],"yields":[169],"substantial":[170],"gains":[172],"several":[174],"benchmarks,":[177],"establishing":[178],"new":[180],"on":[182],"two":[183],"challenging":[184],"benchmarks":[185],"in-the-wild":[187],"-inpainted":[192],"images.":[193]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-01T00:00:00"}