{"id":"https://openalex.org/W4414928217","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1664","title":"microCLIP: Unsupervised CLIP Adaptation via Coarse-Fine Token Fusion for Fine-Grained Image Classification","display_name":"microCLIP: Unsupervised CLIP Adaptation via Coarse-Fine Token Fusion for Fine-Grained Image Classification","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W4414928217","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1664"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2026.findings-acl.1664","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1664","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1664","pdf_url":null,"source":{"id":"https://openalex.org/S4363605144","display_name":"Findings of the Association for Computational Linguistics: ACL (Findings of ACL)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://doi.org/10.18653/v1/2026.findings-acl.1664","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5052661681","display_name":"Sathira Silva","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sathira Silva","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Eman Ali","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Eman Ali","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Chetan Arora","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chetan Arora","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5032830353","display_name":"Muhammad Haris Khan","orcid":"https://orcid.org/0000-0001-9746-276X","observed_orcids":["https://orcid.org/0000-0001-9746-276X"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Muhammad Haris Khan","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":8.5469,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.92685659,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":93,"max":96},"biblio":{"volume":null,"issue":null,"first_page":"33277","last_page":"33294"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.7026000022888184,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.7026000022888184,"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":0.2687999904155731,"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":0.024000000208616257,"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/fine-grained-image-classification","display_name":"fine-grained image classification","score":0.9120000004768372},{"id":"https://openalex.org/keywords/clip","display_name":"CLIP","score":0.8880000114440918},{"id":"https://openalex.org/keywords/pseudo-labeling","display_name":"pseudo-labeling","score":0.9110000133514404},{"id":"https://openalex.org/keywords/self-training","display_name":"self-training","score":0.7919999957084656},{"id":"https://openalex.org/keywords/zero-shot-transfer","display_name":"zero-shot transfer","score":0.5070000290870667},{"id":"https://openalex.org/keywords/patch-embedding","display_name":"patch embedding","score":0.4740000069141388}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7842000126838684},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.723800003528595},{"id":"https://openalex.org/C95623464","wikidata":"https://www.wikidata.org/wiki/Q1096149","display_name":"Classifier (UML)","level":2,"score":0.6449000239372253},{"id":"https://openalex.org/C70437156","wikidata":"https://www.wikidata.org/wiki/Q7228652","display_name":"Pooling","level":2,"score":0.5494999885559082},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.5311999917030334},{"id":"https://openalex.org/C75294576","wikidata":"https://www.wikidata.org/wiki/Q5165192","display_name":"Contextual image classification","level":3,"score":0.5152000188827515},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.5062000155448914},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.45320001244544983},{"id":"https://openalex.org/C177769412","wikidata":"https://www.wikidata.org/wiki/Q278090","display_name":"Prior probability","level":3,"score":0.4377000033855438},{"id":"https://openalex.org/C8038995","wikidata":"https://www.wikidata.org/wiki/Q1152135","display_name":"Unsupervised learning","level":2,"score":0.35690000653266907},{"id":"https://openalex.org/C139807058","wikidata":"https://www.wikidata.org/wiki/Q352374","display_name":"Adaptation (eye)","level":2,"score":0.35269999504089355},{"id":"https://openalex.org/C34872919","wikidata":"https://www.wikidata.org/wiki/Q7092302","display_name":"One-class classification","level":3,"score":0.30649998784065247},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.2728999853134155},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.25049999356269836}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.18653/v1/2026.findings-acl.1664","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1664","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1664","pdf_url":null,"source":{"id":"https://openalex.org/S4363605144","display_name":"Findings of the Association for Computational Linguistics: ACL (Findings of ACL)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2510.02270","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2510.02270","pdf_url":"https://arxiv.org/pdf/2510.02270","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,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2510.02270","doi":"https://doi.org/10.48550/arxiv.2510.02270","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2510.02270","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,"listed_in":[],"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"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-acl.1664","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1664","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1664","pdf_url":null,"source":{"id":"https://openalex.org/S4363605144","display_name":"Findings of the Association for Computational Linguistics: ACL (Findings of ACL)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4414697183","https://openalex.org/W4417473525","https://openalex.org/W7168967114","https://openalex.org/W4412381901","https://openalex.org/W4403620814","https://openalex.org/W7160424083","https://openalex.org/W7131849894","https://openalex.org/W6948418764","https://openalex.org/W4416026986","https://openalex.org/W4391632036"],"abstract_inverted_index":{"Unsupervised":[0],"adaptation":[1],"of":[2],"CLIP-based":[3],"vision-language":[4],"models":[5],"(VLMs)":[6],"for":[7,106,130],"fine-grained":[8,33,39,74,169,182],"image":[9],"classification":[10,34],"requires":[11],"sensitivity":[12],"to":[13,160],"microscopic":[14],"local":[15],"cues.":[16,75],"While":[17],"CLIP":[18,50],"exhibits":[19],"strong":[20],"zero-shot":[21],"transfer,":[22],"its":[23,30,77],"reliance":[24],"on":[25,32],"coarse":[26],"global":[27,103],"features":[28],"restricts":[29],"performance":[31],"tasks.":[35],"Prior":[36],"efforts":[37],"inject":[38],"knowledge":[40],"by":[41],"aligning":[42],"large":[43],"language":[44],"model":[45],"(LLM)":[46],"descriptions":[47,139],"with":[48,101,142,156],"the":[49,102],"$\\texttt{[CLS]}$":[51,104],"token;":[52],"however,":[53],"this":[54],"approach":[55],"overlooks":[56],"spatial":[57],"precision.":[58],"We":[59,144],"propose":[60],"$\\textbf{microCLIP}$,":[61],"a":[62,85,91,114,118,126,133,174],"self-training":[63],"framework":[64],"that":[65],"jointly":[66],"refines":[67],"CLIP's":[68],"visual":[69],"and":[70,98,132,140],"textual":[71],"representations":[72],"using":[73],"At":[76],"core":[78],"is":[79,191],"Saliency-Oriented":[80],"Attention":[81],"Pooling":[82],"(SOAP)":[83],"within":[84],"lightweight":[86],"TokenFusion":[87],"module,":[88],"which":[89,150],"builds":[90],"saliency-guided":[92],"$\\texttt{[FG]}$":[93],"token":[94,105],"from":[95,137],"patch":[96],"embeddings":[97],"fuses":[99],"it":[100],"coarse-fine":[107],"alignment.":[108],"To":[109],"stabilize":[110],"adaptation,":[111],"we":[112],"introduce":[113],"two-headed":[115],"LLM-derived":[116],"classifier:":[117],"frozen":[119],"classifier":[120,135],"that,":[121],"via":[122],"multi-view":[123],"alignment,":[124],"provides":[125],"stable":[127],"text-based":[128],"prior":[129],"pseudo-labeling,":[131],"learnable":[134],"initialized":[136],"LLM":[138],"fine-tuned":[141],"TokenFusion.":[143],"further":[145],"develop":[146],"Dynamic":[147],"Knowledge":[148],"Aggregation,":[149],"convexly":[151],"combines":[152],"fixed":[153],"LLM/CLIP":[154],"priors":[155],"TokenFusion's":[157],"evolving":[158],"logits":[159],"iteratively":[161],"refine":[162],"pseudo-labels.":[163],"Together,":[164],"these":[165],"components":[166],"uncover":[167],"latent":[168],"signals":[170],"in":[171],"CLIP,":[172],"yielding":[173],"consistent":[175],"$2.90\\%$":[176],"average":[177],"accuracy":[178],"gain":[179],"across":[180],"13":[181],"benchmarks":[183],"while":[184],"requiring":[185],"only":[186],"light":[187],"adaptation.":[188],"Our":[189],"code":[190],"available":[192],"at":[193],"https://github.com/sathiiii/microCLIP.":[194]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-10-26T11:51:55.731654","created_date":"2025-10-10T00:00:00"}