{"id":"https://openalex.org/W4410114623","doi":"https://doi.org/10.1109/tcsvt.2025.3567138","title":"DiViCo: Disentangled Visual Token Compression for Efficient Large Vision-Language Model","display_name":"DiViCo: Disentangled Visual Token Compression for Efficient Large Vision-Language Model","publication_year":2025,"publication_date":"2025-05-06","ids":{"openalex":"https://openalex.org/W4410114623","doi":"https://doi.org/10.1109/tcsvt.2025.3567138"},"language":"en","primary_location":{"id":"doi:10.1109/tcsvt.2025.3567138","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tcsvt.2025.3567138","pdf_url":null,"source":{"id":"https://openalex.org/S115173108","display_name":"IEEE Transactions on Circuits and Systems for Video Technology","issn_l":"1051-8215","issn":["1051-8215","1558-2205"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Circuits and Systems for Video Technology","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5022927606","display_name":"Xin Wang","orcid":"https://orcid.org/0000-0002-0351-2939"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xin Wang","raw_affiliation_strings":["Department of Computer Science and Technology, Beijing National Research Center for Information Science and Technology, Tsinghua University, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0002-0351-2939","affiliations":[{"raw_affiliation_string":"Department of Computer Science and Technology, Beijing National Research Center for Information Science and Technology, Tsinghua University, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5074869600","display_name":"Zirui Pan","orcid":"https://orcid.org/0000-0002-6795-0620"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zirui Pan","raw_affiliation_strings":["Department of Computer Science and Technology, Beijing National Research Center for Information Science and Technology, Tsinghua University, Beijing, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science and Technology, Beijing National Research Center for Information Science and Technology, Tsinghua University, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100420416","display_name":"Hong Chen","orcid":"https://orcid.org/0000-0002-0943-2286"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hong Chen","raw_affiliation_strings":["Department of Computer Science and Technology, Beijing National Research Center for Information Science and Technology, Tsinghua University, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0002-0943-2286","affiliations":[{"raw_affiliation_string":"Department of Computer Science and Technology, Beijing National Research Center for Information Science and Technology, Tsinghua University, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100339293","display_name":"Wenwu Zhu","orcid":"https://orcid.org/0000-0003-2236-9290"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wenwu Zhu","raw_affiliation_strings":["Department of Computer Science and Technology, Beijing National Research Center for Information Science and Technology, Tsinghua University, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0003-2236-9290","affiliations":[{"raw_affiliation_string":"Department of Computer Science and Technology, Beijing National Research Center for Information Science and Technology, Tsinghua University, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I99065089"],"apc_list":null,"apc_paid":null,"fwci":1.5311,"has_fulltext":false,"cited_by_count":2,"citation_normalized_percentile":{"value":0.82070005,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":95,"max":96},"biblio":{"volume":"36","issue":"2","first_page":"1392","last_page":"1405"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10627","display_name":"Advanced Image and Video Retrieval Techniques","score":0.9879999756813049,"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/T10627","display_name":"Advanced Image and Video Retrieval Techniques","score":0.9879999756813049,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.963100016117096,"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/T10824","display_name":"Image Retrieval and Classification Techniques","score":0.9581999778747559,"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/computer-science","display_name":"Computer science","score":0.7594002485275269},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.6232344508171082},{"id":"https://openalex.org/keywords/compression","display_name":"Compression (physics)","score":0.4647117853164673},{"id":"https://openalex.org/keywords/computer-vision","display_name":"Computer vision","score":0.4591081142425537},{"id":"https://openalex.org/keywords/data-compression","display_name":"Data compression","score":0.455719918012619},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.44156789779663086},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.32614314556121826},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.13831225037574768}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7594002485275269},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.6232344508171082},{"id":"https://openalex.org/C180016635","wikidata":"https://www.wikidata.org/wiki/Q2712821","display_name":"Compression (physics)","level":2,"score":0.4647117853164673},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.4591081142425537},{"id":"https://openalex.org/C78548338","wikidata":"https://www.wikidata.org/wiki/Q2493","display_name":"Data compression","level":2,"score":0.455719918012619},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.44156789779663086},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.32614314556121826},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.13831225037574768},{"id":"https://openalex.org/C192562407","wikidata":"https://www.wikidata.org/wiki/Q228736","display_name":"Materials science","level":0,"score":0.0},{"id":"https://openalex.org/C159985019","wikidata":"https://www.wikidata.org/wiki/Q181790","display_name":"Composite material","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tcsvt.2025.3567138","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tcsvt.2025.3567138","pdf_url":null,"source":{"id":"https://openalex.org/S115173108","display_name":"IEEE Transactions on Circuits and Systems for Video Technology","issn_l":"1051-8215","issn":["1051-8215","1558-2205"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Circuits and Systems for Video Technology","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[{"score":0.5199999809265137,"display_name":"Reduced inequalities","id":"https://metadata.un.org/sdg/10"}],"awards":[{"id":"https://openalex.org/G3022012397","display_name":null,"funder_award_id":"BNR2023TD03006","funder_id":"https://openalex.org/F4320329777","funder_display_name":"Beijing National Research Center For Information Science And Technology"},{"id":"https://openalex.org/G8567821897","display_name":null,"funder_award_id":"62222209","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320329777","display_name":"Beijing National Research Center For Information Science And Technology","ror":null}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":28,"referenced_works":["https://openalex.org/W2066334462","https://openalex.org/W2099111195","https://openalex.org/W2560730294","https://openalex.org/W2606982687","https://openalex.org/W2765716052","https://openalex.org/W2892797978","https://openalex.org/W2963518342","https://openalex.org/W2963622213","https://openalex.org/W2979382951","https://openalex.org/W3015959599","https://openalex.org/W4213448193","https://openalex.org/W4246233237","https://openalex.org/W4311415873","https://openalex.org/W4312238419","https://openalex.org/W4312933868","https://openalex.org/W4319300002","https://openalex.org/W4385245566","https://openalex.org/W4385572090","https://openalex.org/W4387449305","https://openalex.org/W4389523832","https://openalex.org/W4392172801","https://openalex.org/W4400188660","https://openalex.org/W4402713111","https://openalex.org/W4402727764","https://openalex.org/W4403002096","https://openalex.org/W4403081466","https://openalex.org/W4403939369","https://openalex.org/W4411403346"],"related_works":["https://openalex.org/W4388335561","https://openalex.org/W2970530566","https://openalex.org/W4288261899","https://openalex.org/W4307309205","https://openalex.org/W2967478618","https://openalex.org/W4385009901","https://openalex.org/W4385572700","https://openalex.org/W2997152889","https://openalex.org/W2612632602","https://openalex.org/W2321805087"],"abstract_inverted_index":{"Large":[0],"Vision-Language":[1],"Models":[2],"have":[3],"drawn":[4],"much":[5],"attention":[6,52,137],"and":[7,83,118,147,159,167,227],"become":[8],"increasingly":[9],"applicable":[10],"in":[11,80,88,98,156],"complicated":[12],"multimodal":[13],"tasks":[14],"such":[15],"as":[16],"visual":[17,45,63,81,116,131,185],"question":[18],"answering,":[19],"video":[20],"grounding,":[21],"etc.":[22],"However,":[23],"it":[24],"still":[25],"suffers":[26],"from":[27,173],"inefficiency":[28],"problem":[29],"during":[30],"the":[31,36,41,60,73,77,115,128,141,148,153,165,170,174,180,183,192,195,200,235],"inference":[32],"stage":[33],"due":[34],"to":[35,57,70,91,134,223],"computational":[37],"overhead":[38],"brought":[39],"by":[40,151],"large":[42],"number":[43],"of":[44,182,194,202,234],"tokens.":[46],"Existing":[47],"works":[48],"either":[49],"utilize":[50],"an":[51],"score":[53],"(or":[54],"visual-text":[55],"relevance)":[56],"filter":[58],"out":[59],"less":[61],"significant":[62],"tokens,":[64,74],"or":[65],"insert":[66],"learnable":[67],"projection":[68],"layers":[69],"directly":[71],"compress":[72],"which":[75,190],"neglects":[76],"informative":[78],"details":[79],"signals":[82],"introduces":[84],"information":[85,150],"loss,":[86],"resulting":[87],"poor":[89],"generalizability":[90,193],"test":[92],"data.":[93],"To":[94],"solve":[95],"these":[96,144],"problems,":[97],"this":[99],"paper":[100],"we":[101,125,163],"propose":[102],"a":[103,157],"novel":[104],"Disentangled":[105],"Visual":[106],"Token":[107],"Compression":[108],"module,":[109],"i.e.,":[110],"DiViCo,":[111],"that":[112],"effectively":[113],"compresses":[114],"tokens":[117,132,146,155,186],"maintains":[119],"good":[120],"performance":[121],"simultaneously.":[122],"In":[123],"concrete,":[124],"first":[126],"select":[127],"top":[129],"\u03c4%":[130],"according":[133],"their":[135],"average":[136],"scores,":[138],"then":[139],"predict":[140],"gap":[142,172],"between":[143],"selected":[145],"original":[149],"employing":[152],"chosen":[154],"disentangled":[158],"variational":[160],"manner.":[161],"Specifically,":[162],"model":[164],"mean":[166],"variance,":[168],"sampling":[169],"predicted":[171],"Gaussian":[175],"prior.":[176],"We":[177],"further":[178],"keep":[179],"informativeness":[181],"compressed":[184],"via":[187],"KL":[188],"divergence,":[189],"ensures":[191],"model.":[196],"Extensive":[197],"experiments":[198],"demonstrate":[199],"advantage":[201],"our":[203],"proposed":[204],"DiViCo":[205,220],"module":[206],"against":[207],"several":[208],"state-of-the-art":[209],"baselines":[210],"over":[211],"various":[212],"real-world":[213],"datasets.":[214],"Most":[215],"notably,":[216],"LLaVA-v1.5-7b":[217,238],"equipped":[218],"with":[219],"is":[221],"able":[222],"reduce":[224],"67.7%":[225],"FLOPs":[226],"save":[228],"51.7%":[229],"time":[230],"while":[231],"maintaining":[232],"95.6%":[233],"accuracy":[236],"for":[237],"without":[239],"any":[240],"compression.":[241]},"counts_by_year":[{"year":2025,"cited_by_count":2}],"updated_date":"2026-02-12T06:13:03.167441","created_date":"2025-10-10T00:00:00"}