{"id":"https://openalex.org/W4406840591","doi":"https://doi.org/10.48550/arxiv.2501.14046","title":"LLM-guided Instance-level Image Manipulation with Diffusion U-Net Cross-Attention Maps","display_name":"LLM-guided Instance-level Image Manipulation with Diffusion U-Net Cross-Attention Maps","publication_year":2025,"publication_date":"2025-01-23","ids":{"openalex":"https://openalex.org/W4406840591","doi":"https://doi.org/10.48550/arxiv.2501.14046"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2501.14046","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2501.14046","pdf_url":"https://arxiv.org/pdf/2501.14046","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"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2501.14046","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5108299847","display_name":"Andrey Palaev","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Palaev, Andrey","raw_affiliation_strings":["AIDecisions , SW20 ODS , London , UK"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"AIDecisions , SW20 ODS , London , UK","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5070522496","display_name":"Adil Khan","orcid":"https://orcid.org/0000-0003-2220-8518","observed_orcids":["https://orcid.org/0000-0003-2220-8518"]},"institutions":[{"id":"https://openalex.org/I191240316","display_name":"University of Hull","ror":"https://ror.org/04nkhwh30","country_code":"GB","type":"education","lineage":["https://openalex.org/I191240316"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Khan, Adil","raw_affiliation_strings":["School of Computer Science , University of Hull , HU6 7RX , Hull , UK"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science , University of Hull , HU6 7RX , Hull , UK","institution_ids":["https://openalex.org/I191240316"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5110955150","display_name":"Syed Intasar Hussain Kazmi","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I178535277","display_name":"University of the West of England","ror":"https://ror.org/02nwg5t34","country_code":"GB","type":"education","lineage":["https://openalex.org/I178535277"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Kazmi, Syed M. Ahsan","raw_affiliation_strings":["Department of Computer Science , Uni- versity of the West of England , BS16 1QY , Bristol , UK"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science , Uni- versity of the West of England , BS16 1QY , Bristol , UK","institution_ids":["https://openalex.org/I178535277"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":1.0,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":1.0,"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/T12357","display_name":"Digital Media Forensic Detection","score":0.0,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.0,"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/text-to-image-generation","display_name":"text-to-image generation","score":0.8500000238418579},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.8460000157356262},{"id":"https://openalex.org/keywords/cross-attention-maps","display_name":"cross-attention maps","score":0.9990000128746033},{"id":"https://openalex.org/keywords/open-vocabulary-object-detection","display_name":"open-vocabulary object detection","score":0.8389999866485596}],"concepts":[{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.5737206339836121},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5197827219963074},{"id":"https://openalex.org/C14166107","wikidata":"https://www.wikidata.org/wiki/Q253829","display_name":"Net (polyhedron)","level":2,"score":0.4962611794471741},{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.48939552903175354},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.44915902614593506},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.40346792340278625},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.23913082480430603},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.1731925904750824},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.10163655877113342},{"id":"https://openalex.org/C97355855","wikidata":"https://www.wikidata.org/wiki/Q11473","display_name":"Thermodynamics","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2501.14046","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2501.14046","pdf_url":"https://arxiv.org/pdf/2501.14046","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.2501.14046","doi":"https://doi.org/10.48550/arxiv.2501.14046","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2501.14046","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":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2501.14046","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2501.14046","pdf_url":"https://arxiv.org/pdf/2501.14046","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"},"sustainable_development_goals":[],"sustainable_development_goals_aurora":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4406840591.pdf","grobid_xml":"https://content.openalex.org/works/W4406840591.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W2772917594","https://openalex.org/W2036807459","https://openalex.org/W2058170566","https://openalex.org/W2755342338","https://openalex.org/W2166024367","https://openalex.org/W3116076068","https://openalex.org/W2229312674","https://openalex.org/W2951359407","https://openalex.org/W2079911747","https://openalex.org/W1969923398"],"abstract_inverted_index":{"The":[0],"advancement":[1],"of":[2,11,70],"text-to-image":[3],"synthesis":[4],"has":[5],"introduced":[6],"powerful":[7],"generative":[8],"models":[9],"capable":[10],"creating":[12],"realistic":[13],"images":[14,110],"from":[15],"textual":[16],"prompts.":[17],"However,":[18],"precise":[19,92,118],"control":[20,36],"over":[21],"image":[22,75],"attributes":[23],"remains":[24],"challenging,":[25],"especially":[26],"at":[27,120,138],"the":[28,83,88,121],"instance":[29,122],"level.":[30],"While":[31],"existing":[32],"methods":[33],"offer":[34],"some":[35],"through":[37],"fine-tuning":[38,125],"or":[39,97,126,132],"auxiliary":[40,127],"information,":[41],"they":[42],"often":[43],"face":[44],"limitations":[45],"in":[46,82,87,108],"flexibility":[47],"and":[48,67,85],"accuracy.":[49],"To":[50],"address":[51],"these":[52],"challenges,":[53],"we":[54],"propose":[55],"a":[56],"pipeline":[57],"leveraging":[58],"Large":[59],"Language":[60],"Models":[61],"(LLMs),":[62],"open-vocabulary":[63],"detectors,":[64],"cross-attention":[65,102],"maps":[66],"intermediate":[68],"activations":[69],"diffusion":[71],"U-Net":[72],"for":[73],"instance-level":[74],"manipulation.":[76],"Our":[77,115],"method":[78,116],"detects":[79],"objects":[80],"mentioned":[81],"prompt":[84],"present":[86],"generated":[89],"image,":[90],"enabling":[91],"manipulation":[93],"without":[94,124],"extensive":[95],"training":[96],"input":[98],"masks.":[99],"By":[100],"incorporating":[101],"maps,":[103],"our":[104],"approach":[105],"ensures":[106],"coherence":[107],"manipulated":[109],"while":[111],"controlling":[112],"object":[113],"positions.":[114],"enables":[117],"manipulations":[119],"level":[123],"information":[128],"such":[129],"as":[130],"masks":[131],"bounding":[133],"boxes.":[134],"Code":[135],"is":[136],"available":[137],"https://github.com/Palandr123/DiffusionU-NetLLM":[139]},"counts_by_year":[],"updated_date":"2026-09-30T09:27:27.532063","created_date":"2025-10-10T00:00:00"}