{"id":"https://openalex.org/W7164487953","doi":"https://doi.org/10.48550/arxiv.2606.12898","title":"Magnifying What Matters: Attention-Guided Adaptive Rendering for Visual Text Comprehension","display_name":"Magnifying What Matters: Attention-Guided Adaptive Rendering for Visual Text Comprehension","publication_year":2026,"publication_date":"2026-06-11","ids":{"openalex":"https://openalex.org/W7164487953","doi":"https://doi.org/10.48550/arxiv.2606.12898"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.12898","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.12898","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":"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"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2606.12898","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5138487453","display_name":"Shenglai Zeng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zeng, Shenglai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138522835","display_name":"Qirui Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Qirui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138506909","display_name":"Kai Guo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guo, Kai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138503398","display_name":"Xinnan Dai","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dai, Xinnan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138518475","display_name":"Xianxuan Long","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Long, Xianxuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5138506093","display_name":"Hui Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Hui","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":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"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":0.9524000287055969,"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":0.9524000287055969,"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/T10601","display_name":"Handwritten Text Recognition Techniques","score":0.006000000052154064,"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/T11605","display_name":"Visual Attention and Saliency Detection","score":0.0052999998442828655,"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/rendering","display_name":"Rendering (computer graphics)","score":0.7261999845504761},{"id":"https://openalex.org/keywords/preprocessor","display_name":"Preprocessor","score":0.6754999756813049},{"id":"https://openalex.org/keywords/visualization","display_name":"Visualization","score":0.4778999984264374},{"id":"https://openalex.org/keywords/bottleneck","display_name":"Bottleneck","score":0.4104999899864197},{"id":"https://openalex.org/keywords/comprehension","display_name":"Comprehension","score":0.4018999934196472},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.3458000123500824}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8148000240325928},{"id":"https://openalex.org/C205711294","wikidata":"https://www.wikidata.org/wiki/Q176953","display_name":"Rendering (computer graphics)","level":2,"score":0.7261999845504761},{"id":"https://openalex.org/C34736171","wikidata":"https://www.wikidata.org/wiki/Q918333","display_name":"Preprocessor","level":2,"score":0.6754999756813049},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5735999941825867},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.4778999984264374},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.4291999936103821},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.4104999899864197},{"id":"https://openalex.org/C511192102","wikidata":"https://www.wikidata.org/wiki/Q5156948","display_name":"Comprehension","level":2,"score":0.4018999934196472},{"id":"https://openalex.org/C121684516","wikidata":"https://www.wikidata.org/wiki/Q7600677","display_name":"Computer graphics (images)","level":1,"score":0.3896999955177307},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.3458000123500824},{"id":"https://openalex.org/C2777561058","wikidata":"https://www.wikidata.org/wiki/Q2652119","display_name":"Program comprehension","level":4,"score":0.32359999418258667},{"id":"https://openalex.org/C2778143727","wikidata":"https://www.wikidata.org/wiki/Q1820650","display_name":"Readability","level":2,"score":0.3190999925136566},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3046000003814697},{"id":"https://openalex.org/C2779227376","wikidata":"https://www.wikidata.org/wiki/Q6505497","display_name":"Layer (electronics)","level":2,"score":0.29159998893737793},{"id":"https://openalex.org/C172367668","wikidata":"https://www.wikidata.org/wiki/Q6504956","display_name":"Data visualization","level":3,"score":0.2865000069141388},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.2590000033378601},{"id":"https://openalex.org/C160633673","wikidata":"https://www.wikidata.org/wiki/Q355198","display_name":"Pixel","level":2,"score":0.25369998812675476}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.12898","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.12898","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.12898","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.12898","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":"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"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/4","score":0.4493520259857178,"display_name":"Quality Education"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Visual":[0],"Text":[1],"Comprehension":[2],"(VTC)":[3],"renders":[4],"text":[5],"into":[6],"images":[7],"for":[8],"a":[9,38,56,69,98,114,120,175],"vision-language":[10],"model":[11],"(VLM)":[12],"to":[13,25,126,136,182],"read,":[14],"sidestepping":[15],"LLM":[16],"context-window":[17],"limits":[18],"and":[19,35,43,80,139,159,163,186,192],"powering":[20],"applications":[21],"from":[22,84],"long-page":[23],"OCR":[24],"multi-page":[26,160],"memory":[27,161],"QA.":[28],"Yet":[29],"existing":[30],"VTC":[31,61,155],"pipelines":[32],"treat":[33],"rendering":[34],"layout":[36],"as":[37,174],"fixed,":[39],"content-agnostic":[40],"preprocessing":[41],"step":[42],"offer":[44],"little":[45],"mechanistic":[46],"understanding":[47],"of":[48,101],"how":[49],"VLMs":[50,67,173],"internally":[51],"process":[52],"visualized":[53],"text.":[54],"Through":[55],"focused":[57],"empirical":[58],"study":[59],"on":[60,93,105],"QA":[62],"tasks,":[63],"we":[64,108],"reveal":[65],"that":[66,118,168],"exhibit":[68],"localization-without-utilization":[70],"regime:":[71],"evidence-localizing":[72],"attention":[73,125],"emerges":[74],"sharply":[75],"in":[76],"the":[77,90,94,102,128,141,149],"middle-to-late":[78,123],"layers":[79],"is":[81],"largely":[82],"decoupled":[83],"answer":[85],"correctness,":[86],"yet":[87],"simply":[88],"enlarging":[89],"localized":[91],"spans":[92,145],"rendered":[95],"page":[96,142],"recovers":[97],"large":[99],"fraction":[100],"failures.":[103],"Building":[104],"these":[106],"observations,":[107],"propose":[109],"AGAR":[110,169],"(Attention-Guided":[111],"Adaptive":[112],"Rendering),":[113],"training-free,":[115],"model-agnostic":[116],"method":[117],"leverages":[119],"VLM's":[121],"own":[122],"layer":[124],"identify":[127],"top-K":[129],"important":[130],"visual":[131],"patches,":[132],"maps":[133],"them":[134],"back":[135],"word":[137],"spans,":[138],"re-renders":[140],"with":[143,179],"those":[144],"enlarged":[146],"before":[147],"re-inferring":[148],"answer.":[150],"Extensive":[151],"experiments":[152],"across":[153],"nine":[154],"benchmarks":[156],"(short-form,":[157],"long-context,":[158],"QA)":[162],"four":[164],"VLM":[165,180],"backbones":[166],"show":[167],"(i)consistently":[170],"improves":[171],"off-the-shelf":[172],"plug-and-play":[176],"enhancement,":[177],"(ii)composes":[178],"post-training":[181],"yield":[183],"further":[184],"gains,":[185],"(iii)remains":[187],"robust":[188],"under":[189],"both":[190],"visual-":[191],"text-side":[193],"input":[194],"degradation.":[195]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-06-13T00:00:00"}