{"id":"https://openalex.org/W4415539971","doi":"https://doi.org/10.1145/3746027.3755203","title":"<scp>CapRecover</scp> : A Cross-Modality Feature Inversion Attack Framework on Vision Language Models","display_name":"<scp>CapRecover</scp> : A Cross-Modality Feature Inversion Attack Framework on Vision Language Models","publication_year":2025,"publication_date":"2025-10-25","ids":{"openalex":"https://openalex.org/W4415539971","doi":"https://doi.org/10.1145/3746027.3755203"},"language":"en","primary_location":{"id":"doi:10.1145/3746027.3755203","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3746027.3755203","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 33rd ACM International Conference on Multimedia","raw_type":"proceedings-article"},"type":"article","indexed_in":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2507.22828","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5120132212","display_name":"Kedong Xiu","orcid":null},"institutions":[{"id":"https://openalex.org/I57206974","display_name":"New York University","ror":"https://ror.org/0190ak572","country_code":"US","type":"education","lineage":["https://openalex.org/I57206974"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Kedong Xiu","raw_affiliation_strings":["New York University, New York, New York, USA"],"affiliations":[{"raw_affiliation_string":"New York University, New York, New York, USA","institution_ids":["https://openalex.org/I57206974"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5013517247","display_name":"Sai Qian Zhang","orcid":"https://orcid.org/0000-0002-4815-9235"},"institutions":[{"id":"https://openalex.org/I57206974","display_name":"New York University","ror":"https://ror.org/0190ak572","country_code":"US","type":"education","lineage":["https://openalex.org/I57206974"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Sai Qian Zhang","raw_affiliation_strings":["New York University, New York, New York, USA"],"affiliations":[{"raw_affiliation_string":"New York University, New York, New York, USA","institution_ids":["https://openalex.org/I57206974"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5120132212"],"corresponding_institution_ids":["https://openalex.org/I57206974"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.32553951,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"3808","last_page":"3816"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9983000159263611,"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":0.9983000159263611,"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":0.9929999709129333,"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":0.9926999807357788,"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/encoder","display_name":"Encoder","score":0.6104000210762024},{"id":"https://openalex.org/keywords/semantic-feature","display_name":"Semantic feature","score":0.5185999870300293},{"id":"https://openalex.org/keywords/semantics","display_name":"Semantics (computer science)","score":0.4668000042438507},{"id":"https://openalex.org/keywords/encode","display_name":"ENCODE","score":0.4519999921321869},{"id":"https://openalex.org/keywords/inversion","display_name":"Inversion (geology)","score":0.4498000144958496},{"id":"https://openalex.org/keywords/semantic-computing","display_name":"Semantic computing","score":0.43160000443458557},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.42419999837875366},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.35989999771118164},{"id":"https://openalex.org/keywords/semantic-data-model","display_name":"Semantic data model","score":0.3549000024795532}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8029999732971191},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.6104000210762024},{"id":"https://openalex.org/C2781122975","wikidata":"https://www.wikidata.org/wiki/Q16928266","display_name":"Semantic feature","level":2,"score":0.5185999870300293},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.4668000042438507},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.46650001406669617},{"id":"https://openalex.org/C66746571","wikidata":"https://www.wikidata.org/wiki/Q1134833","display_name":"ENCODE","level":3,"score":0.4519999921321869},{"id":"https://openalex.org/C1893757","wikidata":"https://www.wikidata.org/wiki/Q3653001","display_name":"Inversion (geology)","level":3,"score":0.4498000144958496},{"id":"https://openalex.org/C511149849","wikidata":"https://www.wikidata.org/wiki/Q7449051","display_name":"Semantic computing","level":3,"score":0.43160000443458557},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.42419999837875366},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3621000051498413},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.35989999771118164},{"id":"https://openalex.org/C90312973","wikidata":"https://www.wikidata.org/wiki/Q7449052","display_name":"Semantic data model","level":2,"score":0.3549000024795532},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.34689998626708984},{"id":"https://openalex.org/C202708506","wikidata":"https://www.wikidata.org/wiki/Q7449050","display_name":"Semantic compression","level":5,"score":0.3456000089645386},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.3319999873638153},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.32260000705718994},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.31949999928474426},{"id":"https://openalex.org/C130318100","wikidata":"https://www.wikidata.org/wiki/Q2268914","display_name":"Semantic similarity","level":2,"score":0.3167000114917755},{"id":"https://openalex.org/C52622490","wikidata":"https://www.wikidata.org/wiki/Q1026626","display_name":"Feature extraction","level":2,"score":0.3147999942302704},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.3124000132083893},{"id":"https://openalex.org/C169258074","wikidata":"https://www.wikidata.org/wiki/Q245748","display_name":"Random forest","level":2,"score":0.29829999804496765},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.28529998660087585},{"id":"https://openalex.org/C204806902","wikidata":"https://www.wikidata.org/wiki/Q2333581","display_name":"Semantic security","level":5,"score":0.2824999988079071},{"id":"https://openalex.org/C67277372","wikidata":"https://www.wikidata.org/wiki/Q7449085","display_name":"Semantic role labeling","level":3,"score":0.2736999988555908},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.2718000113964081},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.2606000006198883},{"id":"https://openalex.org/C2779227376","wikidata":"https://www.wikidata.org/wiki/Q6505497","display_name":"Layer (electronics)","level":2,"score":0.2605000138282776}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1145/3746027.3755203","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3746027.3755203","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 33rd ACM International Conference on Multimedia","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2507.22828","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2507.22828","pdf_url":"https://arxiv.org/pdf/2507.22828","source":{"id":"https://openalex.org/S4393918464","display_name":"ArXiv.org","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2507.22828","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2507.22828","pdf_url":"https://arxiv.org/pdf/2507.22828","source":{"id":"https://openalex.org/S4393918464","display_name":"ArXiv.org","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":7,"referenced_works":["https://openalex.org/W68733909","https://openalex.org/W2108598243","https://openalex.org/W2194775991","https://openalex.org/W2963163009","https://openalex.org/W2982083293","https://openalex.org/W2989885118","https://openalex.org/W4390874053"],"related_works":[],"abstract_inverted_index":{"As":[0],"Vision-Language":[1],"Models":[2],"(VLMs)":[3],"are":[4],"increasingly":[5],"deployed":[6],"in":[7,46,92,163],"split-DNN":[8],"configurations--with":[9],"visual":[10],"encoders":[11],"(e.g.,":[12],"ResNet,":[13],"ViT)":[14],"operating":[15],"on":[16,83,104,113],"user":[17],"devices":[18],"and":[19,86,106,159],"sending":[20],"intermediate":[21,42,75,154],"features":[22,43,76,112,155],"to":[23,37,99,119,135,153],"the":[24,161,164],"cloud--there":[25],"is":[26,182],"a":[27,59,144],"growing":[28],"privacy":[29],"risk":[30],"from":[31,40,74,110],"semantic":[32,54,66,93,132,140,174],"information":[33,133],"leakage.":[34],"Existing":[35],"approaches":[36],"reconstructing":[38],"images":[39],"these":[41],"often":[44],"result":[45],"blurry,":[47],"semantically":[48],"ambiguous":[49],"images.":[50],"To":[51,138],"directly":[52,73],"address":[53],"leakage,":[55,141],"we":[56,142],"propose":[57],"CapRecover,":[58],"cross-modality":[60],"inversion":[61],"framework":[62],"that":[63,125,170],"recovers":[64],"high-level":[65],"content,":[67],"such":[68],"as":[69],"labels":[70],"or":[71],"captions,":[72],"without":[77,176],"image":[78],"reconstruction.":[79],"We":[80],"evaluate":[81],"CapRecover":[82,96],"multiple":[84],"datasets":[85],"victim":[87],"models,":[88],"demonstrating":[89],"strong":[90],"performance":[91],"recovery.":[94],"Specifically,":[95],"achieves":[97],"up":[98,118],"92.71%":[100],"Top-1":[101],"label":[102],"accuracy":[103],"CIFAR-10":[105],"generates":[107],"fluent":[108],"captions":[109],"ResNet50":[111],"COCO2017":[114],"with":[115],"ROUGE-L":[116],"scores":[117],"0.52.":[120],"Our":[121,180],"analysis":[122],"further":[123],"reveals":[124],"deeper":[126],"convolutional":[127],"layers":[128],"encode":[129],"significantly":[130],"more":[131],"compared":[134],"shallow":[136],"layers.":[137],"mitigate":[139],"introduce":[143],"simple":[145],"yet":[146],"effective":[147],"protection":[148],"method:":[149],"adding":[150],"random":[151],"noise":[152,162],"at":[156,184],"each":[157],"layer":[158],"removing":[160],"next":[165],"layer.":[166],"Experimental":[167],"results":[168],"show":[169],"this":[171],"approach":[172],"prevents":[173],"leakage":[175],"additional":[177],"training":[178],"costs.":[179],"code":[181],"available":[183],"https://jus1mple.github.io/Image2CaptionAttack.":[185]},"counts_by_year":[],"updated_date":"2026-03-07T16:01:11.037858","created_date":"2025-10-25T00:00:00"}