{"id":"https://openalex.org/W7131389402","doi":"https://doi.org/10.48550/arxiv.2602.20972","title":"Are Multimodal Large Language Models Good Annotators for Image Tagging?","display_name":"Are Multimodal Large Language Models Good Annotators for Image Tagging?","publication_year":2026,"publication_date":"2026-02-24","ids":{"openalex":"https://openalex.org/W7131389402","doi":"https://doi.org/10.48550/arxiv.2602.20972"},"language":null,"primary_location":{"id":"pmh:doi:10.48550/arxiv.2602.20972","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"type":"article","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":null,"any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5126845919","display_name":"Ming-Kun Xie","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xie, Ming-Kun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102118583","display_name":"Jia\u2010Hao Xiao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiao, Jia-Hao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126818480","display_name":"Zhiqiang Kou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kou, Zhiqiang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5086917578","display_name":"Zhongnian Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Zhongnian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126819534","display_name":"Gang Niu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Niu, Gang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5126791334","display_name":"Masashi Sugiyama","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sugiyama, Masashi","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":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.21831637,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.19900000095367432,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.19900000095367432,"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/T11550","display_name":"Text and Document Classification Technologies","score":0.12309999763965607,"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"}},{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.0649000033736229,"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/annotation","display_name":"Annotation","score":0.8537999987602234},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.6410999894142151},{"id":"https://openalex.org/keywords/image","display_name":"Image (mathematics)","score":0.5519000291824341},{"id":"https://openalex.org/keywords/quality","display_name":"Quality (philosophy)","score":0.5228000283241272},{"id":"https://openalex.org/keywords/semantic-gap","display_name":"Semantic gap","score":0.43149998784065247},{"id":"https://openalex.org/keywords/downstream","display_name":"Downstream (manufacturing)","score":0.4311000108718872},{"id":"https://openalex.org/keywords/training-set","display_name":"Training set","score":0.42250001430511475}],"concepts":[{"id":"https://openalex.org/C2776321320","wikidata":"https://www.wikidata.org/wiki/Q857525","display_name":"Annotation","level":2,"score":0.8537999987602234},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.839900016784668},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.6410999894142151},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5938000082969666},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.5519000291824341},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.5228000283241272},{"id":"https://openalex.org/C86034646","wikidata":"https://www.wikidata.org/wiki/Q474311","display_name":"Semantic gap","level":4,"score":0.43149998784065247},{"id":"https://openalex.org/C2776207758","wikidata":"https://www.wikidata.org/wiki/Q5303302","display_name":"Downstream (manufacturing)","level":2,"score":0.4311000108718872},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.42250001430511475},{"id":"https://openalex.org/C199579030","wikidata":"https://www.wikidata.org/wiki/Q2851778","display_name":"Automatic image annotation","level":4,"score":0.4219000041484833},{"id":"https://openalex.org/C1667742","wikidata":"https://www.wikidata.org/wiki/Q10927554","display_name":"Image retrieval","level":3,"score":0.4154999852180481},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.4016999900341034},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.3831999897956848},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.36469998955726624},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.33379998803138733},{"id":"https://openalex.org/C9417928","wikidata":"https://www.wikidata.org/wiki/Q1070689","display_name":"Image processing","level":3,"score":0.30869999527931213},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3046000003814697},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.30390000343322754},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.27489998936653137},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.266400009393692},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.259799987077713},{"id":"https://openalex.org/C55020928","wikidata":"https://www.wikidata.org/wiki/Q3813865","display_name":"Image quality","level":3,"score":0.25029999017715454}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:doi:10.48550/arxiv.2602.20972","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},{"id":"doi:10.48550/arxiv.2602.20972","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.20972","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:doi:10.48550/arxiv.2602.20972","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"sustainable_development_goals":[{"display_name":"Decent work and economic growth","score":0.753263533115387,"id":"https://metadata.un.org/sdg/8"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Image":[0],"tagging,":[1,136],"a":[2,75,131,162],"fundamental":[3],"vision":[4],"task,":[5],"traditionally":[6],"relies":[7],"on":[8,121],"human-annotated":[9],"datasets":[10],"to":[11,30,35,44,54,63,83,102,111,139,159,223],"train":[12],"multi-label":[13],"classifiers,":[14],"which":[15,97,137,154,182],"incurs":[16],"significant":[17],"labor":[18],"and":[19,50,53,145,179,193,210],"costs.":[20],"While":[21],"Multimodal":[22],"Large":[23],"Language":[24],"Models":[25],"(MLLMs)":[26],"offer":[27],"promising":[28],"potential":[29],"automate":[31],"annotation,":[32],"their":[33],"capability":[34],"replace":[36,64],"human":[37,51,90,114,146,211],"annotators":[38],"remains":[39],"underexplored.":[40],"This":[41],"paper":[42],"aims":[43,138],"analyze":[45],"the":[46,89,141,185,191,196,206,226],"gap":[47,142,207],"between":[48,143,208],"MLLM-generated":[49,144,209],"annotations":[52,71],"propose":[55,129],"an":[56],"effective":[57],"solution":[58],"that":[59,166,202],"enables":[60],"MLLM-based":[61],"annotation":[62,81,106,177],"manual":[65,103],"labeling.":[66],"Our":[67],"analysis":[68],"of":[69,88,113,188,225],"MLLM":[70],"reveals":[72],"that,":[73],"under":[74],"conservative":[76],"estimate,":[77],"MLLMs":[78],"can":[79],"reduce":[80],"cost":[82],"as":[84,86,168,172],"low":[85],"one-thousandth":[87],"cost,":[91],"mainly":[92],"accounting":[93],"for":[94,134],"GPU":[95],"usage,":[96],"is":[98],"nearly":[99],"negligible":[100],"compared":[101],"efforts.":[104],"Their":[105],"quality":[107],"reaches":[108],"about":[109,221],"50\\%":[110],"80\\%":[112,224],"performance,":[115,217],"while":[116,174],"achieving":[117],"over":[118],"90\\%":[119],"performance":[120],"downstream":[122,215],"training":[123,216],"tasks.Motivated":[124],"by":[125],"these":[126],"findings,":[127],"we":[128],"TagLLM,":[130],"novel":[132],"framework":[133],"image":[135],"narrow":[140],"annotations.":[147],"TagLLM":[148,203],"comprises":[149],"two":[150],"components:":[151],"Candidates":[152],"generation,":[153],"employs":[155],"structured":[156],"group-wise":[157],"prompting":[158],"efficiently":[160],"produce":[161],"compact":[163],"candidate":[164,197],"set":[165],"covers":[167],"many":[169],"true":[170],"labels":[171],"possible":[173],"reducing":[175],"subsequent":[176],"workload;":[178],"label":[180],"disambiguation,":[181],"interactively":[183],"calibrates":[184],"semantic":[186],"concept":[187],"categories":[189],"in":[190,214],"prompts":[192],"effectively":[194],"refines":[195],"labels.":[198],"Extensive":[199],"experiments":[200],"show":[201],"substantially":[204],"narrows":[205],"annotations,":[212],"especially":[213],"where":[218],"it":[219],"closes":[220],"60\\%":[222],"difference.":[227]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-02-26T00:00:00"}