{"id":"https://openalex.org/W4414634804","doi":"https://doi.org/10.1109/iccv51701.2025.01422","title":"From Reflection to Perfection: Scaling Inference-Time Optimization for Text-to-Image Diffusion Models via Reflection Tuning","display_name":"From Reflection to Perfection: Scaling Inference-Time Optimization for Text-to-Image Diffusion Models via Reflection Tuning","publication_year":2025,"publication_date":"2025-10-19","ids":{"openalex":"https://openalex.org/W4414634804","doi":"https://doi.org/10.1109/iccv51701.2025.01422"},"language":"en","primary_location":{"id":"doi:10.1109/iccv51701.2025.01422","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iccv51701.2025.01422","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/CVF International Conference on Computer Vision (ICCV)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2504.16080","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100310256","display_name":"Le Zhuo","orcid":null},"institutions":[{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Le Zhuo","raw_affiliation_strings":["CUHK MMLab"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"CUHK MMLab","institution_ids":["https://openalex.org/I889458895"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5041967855","display_name":"Liangbing Zhao","orcid":null},"institutions":[{"id":"https://openalex.org/I4210099236","display_name":"Kootenay Association for Science & Technology","ror":"https://ror.org/011pv9p44","country_code":"CA","type":"nonprofit","lineage":["https://openalex.org/I4210099236"]},{"id":"https://openalex.org/I71920554","display_name":"King Abdullah University of Science and Technology","ror":"https://ror.org/01q3tbs38","country_code":"SA","type":"education","lineage":["https://openalex.org/I71920554"]}],"countries":["CA","SA"],"is_corresponding":false,"raw_author_name":"Liangbing Zhao","raw_affiliation_strings":["KAUST"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"KAUST","institution_ids":["https://openalex.org/I4210099236","https://openalex.org/I71920554"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5043910691","display_name":"Sayak Paul","orcid":"https://orcid.org/0000-0003-0217-0778"},"institutions":[{"id":"https://openalex.org/I4210136376","display_name":"FACE Foundation","ror":"https://ror.org/04121g745","country_code":"US","type":"nonprofit","lineage":["https://openalex.org/I4210136376"]},{"id":"https://openalex.org/I4387154989","display_name":"Hugging Face","ror":"https://ror.org/02grspc61","country_code":"US","type":"company","lineage":["https://openalex.org/I4387154989"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Sayak Paul","raw_affiliation_strings":["Hugging Face"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Hugging Face","institution_ids":["https://openalex.org/I4210136376","https://openalex.org/I4387154989"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Yue Liao","orcid":null},"institutions":[{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Yue Liao","raw_affiliation_strings":["CUHK MMLab"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"CUHK MMLab","institution_ids":["https://openalex.org/I889458895"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5086183847","display_name":"Renrui Zhang","orcid":null},"institutions":[{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Renrui Zhang","raw_affiliation_strings":["CUHK MMLab"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"CUHK MMLab","institution_ids":["https://openalex.org/I889458895"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100646572","display_name":"Xin Yi","orcid":"https://orcid.org/0000-0003-4240-7039"},"institutions":[{"id":"https://openalex.org/I4391012619","display_name":"Shanghai Artificial Intelligence Laboratory","ror":"https://ror.org/03wkvpx79","country_code":null,"type":"facility","lineage":["https://openalex.org/I4391012619"]}],"countries":[],"is_corresponding":false,"raw_author_name":"Yi Xin","raw_affiliation_strings":["Shanghai AI Lab"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Shanghai AI Lab","institution_ids":["https://openalex.org/I4391012619"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5028272399","display_name":"Peng Gao","orcid":"https://orcid.org/0000-0003-2230-3937"},"institutions":[{"id":"https://openalex.org/I4391012619","display_name":"Shanghai Artificial Intelligence Laboratory","ror":"https://ror.org/03wkvpx79","country_code":null,"type":"facility","lineage":["https://openalex.org/I4391012619"]}],"countries":[],"is_corresponding":false,"raw_author_name":"Peng Gao","raw_affiliation_strings":["Shanghai AI Lab"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Shanghai AI Lab","institution_ids":["https://openalex.org/I4391012619"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5085089542","display_name":"Mohamed Elhoseiny","orcid":"https://orcid.org/0000-0001-9659-1551"},"institutions":[{"id":"https://openalex.org/I4210099236","display_name":"Kootenay Association for Science & Technology","ror":"https://ror.org/011pv9p44","country_code":"CA","type":"nonprofit","lineage":["https://openalex.org/I4210099236"]},{"id":"https://openalex.org/I71920554","display_name":"King Abdullah University of Science and Technology","ror":"https://ror.org/01q3tbs38","country_code":"SA","type":"education","lineage":["https://openalex.org/I71920554"]}],"countries":["CA","SA"],"is_corresponding":false,"raw_author_name":"Mohamed Elhoseiny","raw_affiliation_strings":["KAUST"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"KAUST","institution_ids":["https://openalex.org/I4210099236","https://openalex.org/I71920554"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100732450","display_name":"Hongsheng Li","orcid":"https://orcid.org/0000-0002-2664-7975"},"institutions":[{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Hongsheng Li","raw_affiliation_strings":["CUHK MMLab"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"CUHK MMLab","institution_ids":["https://openalex.org/I889458895"]}]}],"institutions":[],"countries_distinct_count":4,"institutions_distinct_count":6,"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":"15329","last_page":"15339"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10417","display_name":"Lung Cancer Treatments and Mutations","score":0.4198000133037567,"subfield":{"id":"https://openalex.org/subfields/2740","display_name":"Pulmonary and Respiratory Medicine"},"field":{"id":"https://openalex.org/fields/27","display_name":"Medicine"},"domain":{"id":"https://openalex.org/domains/4","display_name":"Health Sciences"}},"topics":[{"id":"https://openalex.org/T10417","display_name":"Lung Cancer Treatments and Mutations","score":0.4198000133037567,"subfield":{"id":"https://openalex.org/subfields/2740","display_name":"Pulmonary and Respiratory Medicine"},"field":{"id":"https://openalex.org/fields/27","display_name":"Medicine"},"domain":{"id":"https://openalex.org/domains/4","display_name":"Health Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/scaling","display_name":"Scaling","score":0.8644000291824341},{"id":"https://openalex.org/keywords/reflection","display_name":"Reflection (computer programming)","score":0.742900013923645},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.6917999982833862},{"id":"https://openalex.org/keywords/construct","display_name":"Construct (python library)","score":0.6434999704360962},{"id":"https://openalex.org/keywords/diffusion","display_name":"Diffusion","score":0.5479000210762024},{"id":"https://openalex.org/keywords/quality","display_name":"Quality (philosophy)","score":0.3833000063896179}],"concepts":[{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.8644000291824341},{"id":"https://openalex.org/C65682993","wikidata":"https://www.wikidata.org/wiki/Q1056451","display_name":"Reflection (computer programming)","level":2,"score":0.742900013923645},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7006999850273132},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.6917999982833862},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.6434999704360962},{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.5479000210762024},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.45590001344680786},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4480000138282776},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.3833000063896179},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2896000146865845},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.28540000319480896},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.27790001034736633},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.2757999897003174},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.27000001072883606}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.1109/iccv51701.2025.01422","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iccv51701.2025.01422","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/CVF International Conference on Computer Vision (ICCV)","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2504.16080","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2504.16080","pdf_url":"https://arxiv.org/pdf/2504.16080","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2504.16080","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2504.16080","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"pmh:oai:arXiv.org:2504.16080","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2504.16080","pdf_url":"https://arxiv.org/pdf/2504.16080","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G5885573117","display_name":null,"funder_award_id":"2023B1515130008,XW","funder_id":"https://openalex.org/F4806748151","funder_display_name":"Centre for Perceptual and Interactive Intelligence"}],"funders":[{"id":"https://openalex.org/F4806748151","display_name":"Centre for Perceptual and Interactive Intelligence","ror":null}],"has_content":{"grobid_xml":false,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4414634804.pdf"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Recent":[0],"text-to-image":[1],"diffusion":[2,44,128],"models":[3,45],"achieve":[4],"impressive":[5],"visual":[6],"quality":[7],"through":[8],"extensive":[9],"scaling":[10,59,63,70,149],"of":[11],"training":[12],"data":[13],"and":[14,24,50,75,89,114,154],"model":[15],"parameters,":[16],"yet":[17],"they":[18],"often":[19],"struggle":[20],"with":[21],"complex":[22],"scenes":[23],"fine-grained":[25],"details.":[26],"Inspired":[27],"by":[28,131],"the":[29],"self-reflection":[30],"capabilities":[31],"emergent":[32],"in":[33],"large":[34],"language":[35],"models,":[36],"we":[37,97,121],"propose":[38],"ReflectionFlow,":[39],"an":[40,115],"inference-time":[41,58],"framework":[42],"enabling":[43],"to":[46,64,86],"iteratively":[47,87],"reflect":[48],"upon":[49],"refine":[51],"their":[52],"outputs.":[53],"ReflectionFlow":[54,144],"introduces":[55],"three":[56],"complementary":[57],"axes:":[60],"(1)":[61],"noise-level":[62,148],"optimize":[65],"latent":[66],"initialization;":[67],"(2)":[68],"prompt-level":[69],"for":[71],"precise":[72],"semantic":[73],"guidance;":[74],"most":[76],"notably,":[77],"(3)":[78],"reflection-level":[79,95],"scaling,":[80,96],"which":[81],"explicitly":[82],"provides":[83],"actionable":[84],"reflections":[85],"assess":[88],"correct":[90],"previous":[91],"generations.":[92],"To":[93],"facilitate":[94],"construct":[98],"GenRef,":[99],"a":[100,109,111,137,152],"large-scale":[101],"dataset":[102],"comprising":[103],"1":[104],"million":[105],"triplets,":[106],"each":[107],"containing":[108],"reflection,":[110],"flawed":[112],"image,":[113],"enhanced":[116],"image.":[117],"Leveraging":[118],"this":[119],"dataset,":[120],"efficiently":[122],"perform":[123],"reflection":[124],"tuning":[125],"on":[126,161],"state-of-the-art":[127],"transformer,":[129],"FLUX.1-dev,":[130],"jointly":[132],"modeling":[133],"multimodal":[134],"inputs":[135],"within":[136],"unified":[138],"framework.":[139],"Experimental":[140],"results":[141],"show":[142],"that":[143],"significantly":[145],"outperforms":[146],"naive":[147],"methods,":[150],"offering":[151],"scalable":[153],"compute-efficient":[155],"solution":[156],"toward":[157],"higher-quality":[158],"image":[159],"synthesis":[160],"challenging":[162],"tasks.":[163]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}