{"id":"https://openalex.org/W4405172865","doi":"https://doi.org/10.1109/iccv51701.2025.02189","title":"CompCap: Improving Multimodal Large Language Models with Composite Captions","display_name":"CompCap: Improving Multimodal Large Language Models with Composite Captions","publication_year":2025,"publication_date":"2025-10-19","ids":{"openalex":"https://openalex.org/W4405172865","doi":"https://doi.org/10.1109/iccv51701.2025.02189"},"language":"en","primary_location":{"id":"doi:10.1109/iccv51701.2025.02189","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iccv51701.2025.02189","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/CVF International Conference on Computer Vision (ICCV)","raw_type":"proceedings-article"},"type":"preprint","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2412.05243","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Xiaohui Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiaohui Chen","raw_affiliation_strings":["Meta"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5065117717","display_name":"Satya Narayan Shukla","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Satya Narayan Shukla","raw_affiliation_strings":["Meta"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5110646972","display_name":"Mahmoud Azab","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mahmoud Azab","raw_affiliation_strings":["Meta"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Aashu Singh","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Aashu Singh","raw_affiliation_strings":["Meta"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101774659","display_name":"Qifan Wang","orcid":"https://orcid.org/0000-0002-7570-5756"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Qifan Wang","raw_affiliation_strings":["Meta"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5053754155","display_name":"David Dawei Yang","orcid":"https://orcid.org/0000-0002-2830-7261"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"David Yang","raw_affiliation_strings":["Meta"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5006255387","display_name":"Shengyun Peng","orcid":"https://orcid.org/0000-0003-3063-2052"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"ShengYun Peng","raw_affiliation_strings":["Meta"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5028988765","display_name":"Hanchao Yu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hanchao Yu","raw_affiliation_strings":["Meta"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5013074509","display_name":"Yan Shen","orcid":"https://orcid.org/0000-0001-9287-1206"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shen Yan","raw_affiliation_strings":["Meta"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100765489","display_name":"Xuewen Zhang","orcid":"https://orcid.org/0000-0001-7373-3764"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xuewen Zhang","raw_affiliation_strings":["Meta"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5041131763","display_name":"Baosheng He","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Baosheng He","raw_affiliation_strings":["Meta"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":11,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.00073164,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"23582","last_page":"23592"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.9993000030517578,"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"}},"topics":[{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.9993000030517578,"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/T10028","display_name":"Topic Modeling","score":0.9943000078201294,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9743000268936157,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/composite-number","display_name":"Composite number","score":0.6631669998168945},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.5784902572631836},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.5378066897392273},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.41451066732406616},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.06504610180854797}],"concepts":[{"id":"https://openalex.org/C104779481","wikidata":"https://www.wikidata.org/wiki/Q50707","display_name":"Composite number","level":2,"score":0.6631669998168945},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5784902572631836},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.5378066897392273},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.41451066732406616},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.06504610180854797}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.1109/iccv51701.2025.02189","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iccv51701.2025.02189","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/CVF International Conference on Computer Vision (ICCV)","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2412.05243","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2412.05243","pdf_url":"https://arxiv.org/pdf/2412.05243","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":null},{"id":"doi:10.48550/arxiv.2412.05243","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2412.05243","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2412.05243","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2412.05243","pdf_url":"https://arxiv.org/pdf/2412.05243","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":null},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W3204019825"],"abstract_inverted_index":{"How":[0],"well":[1],"can":[2],"Multimodal":[3],"Large":[4,130],"Language":[5,131],"Models":[6,132],"(MLLMs)":[7],"understand":[8],"composite":[9],"images?":[10],"Composite":[11,122],"images":[12,53],"(CIs)":[13],"are":[14,39,89,111],"synthetic":[15],"visuals":[16],"created":[17],"by":[18,34,166],"merging":[19],"multiple":[20],"visual":[21],"elements,":[22],"such":[23],"as":[24],"charts,":[25],"posters,":[26],"or":[27,73],"screenshots,":[28],"rather":[29],"than":[30],"being":[31],"captured":[32],"directly":[33],"a":[35,125,150],"camera.":[36],"While":[37],"CIs":[38,88,139],"prevalent":[40],"in":[41,64,96],"real-world":[42],"applications,":[43],"recent":[44],"MLLM":[45],"developments":[46],"have":[47],"primarily":[48],"focused":[49],"on":[50,78],"interpreting":[51],"natural":[52],"(NIs).":[54],"Our":[55],"research":[56],"reveals":[57],"that":[58,83,128,179],"current":[59],"MLLMs":[60,169],"face":[61],"significant":[62],"challenges":[63],"accurately":[65],"understanding":[66,184],"CIs,":[67,186],"often":[68],"struggling":[69],"to":[70,137],"extract":[71],"information":[72],"perform":[74],"complex":[75],"reasoning":[76],"based":[77],"these":[79],"images.":[80],"We":[81,160],"find":[82],"existing":[84],"training":[85],"data":[86],"for":[87,92,107,114],"mostly":[90],"formatted":[91],"question-answer":[93],"tasks":[94],"(e.g.,":[95],"datasets":[97],"like":[98],"ChartQA":[99],"and":[100,134,142,174,193],"ScienceQA),":[101],"while":[102],"high-quality":[103],"image-caption":[104,154],"datasets,":[105],"critical":[106],"robust":[108],"vision-language":[109],"alignment,":[110],"only":[112],"available":[113],"NIs.":[115],"To":[116],"bridge":[117],"this":[118],"gap,":[119],"we":[120,147],"introduce":[121],"Captions":[123],"(CompCap),":[124],"flexible":[126],"framework":[127],"leverages":[129],"(LLMs)":[133],"automation":[135],"tools":[136],"synthesize":[138],"with":[140],"accurate":[141],"detailed":[143],"captions.":[144],"Using":[145],"CompCap,":[146],"curate":[148],"CompCap-118K,":[149],"dataset":[151],"containing":[152],"118K":[153],"pairs":[155],"across":[156,195],"six":[157],"CI":[158],"types.":[159],"validate":[161],"the":[162],"effectiveness":[163],"of":[164,170,185,190],"CompCap-118K":[165,180],"supervised":[167],"fine-tuning":[168],"three":[171],"sizes:":[172],"xGen-MM-inst.-4B":[173],"LLaVA-NeXT-Vicuna-7B/13B.":[175],"Empirical":[176],"results":[177],"show":[178],"significantly":[181],"enhances":[182],"MLLMs'":[183],"yielding":[187],"average":[188],"gains":[189],"1.7%,":[191],"2.0%,":[192],"2.9%":[194],"eleven":[196],"benchmarks,":[197],"respectively.":[198]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-10T00:00:00"}