{"id":"https://openalex.org/W4415536730","doi":"https://doi.org/10.1145/3746027.3754582","title":"Enhancing Diffusion Model Stability for Image Restoration via Gradient Management","display_name":"Enhancing Diffusion Model Stability for Image Restoration via Gradient Management","publication_year":2025,"publication_date":"2025-10-25","ids":{"openalex":"https://openalex.org/W4415536730","doi":"https://doi.org/10.1145/3746027.3754582"},"language":"en","primary_location":{"id":"doi:10.1145/3746027.3754582","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3746027.3754582","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":"preprint","indexed_in":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2507.06656","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5028144381","display_name":"Hongjie Wu","orcid":"https://orcid.org/0009-0007-3203-1521"},"institutions":[{"id":"https://openalex.org/I4210125143","display_name":"Chengdu University","ror":"https://ror.org/034z67559","country_code":"CN","type":"education","lineage":["https://openalex.org/I4210125143"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Hongjie Wu","raw_affiliation_strings":["College of Computer Science, Sichuan University, Chengdu, Sichuan, China and Engineering Research Center of Machine Learning and Industry Intelligence, Ministry of Education, Chengdu, Sichuan, China"],"raw_orcid":"https://orcid.org/0009-0007-3203-1521","affiliations":[{"raw_affiliation_string":"College of Computer Science, Sichuan University, Chengdu, Sichuan, China and Engineering Research Center of Machine Learning and Industry Intelligence, Ministry of Education, Chengdu, Sichuan, China","institution_ids":["https://openalex.org/I4210125143"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5111169711","display_name":"Mingqin Zhang","orcid":null},"institutions":[{"id":"https://openalex.org/I4210125143","display_name":"Chengdu University","ror":"https://ror.org/034z67559","country_code":"CN","type":"education","lineage":["https://openalex.org/I4210125143"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Mingqin Zhang","raw_affiliation_strings":["College of Computer Science, Sichuan University, Chengdu, Sichuan, China and Engineering Research Center of Machine Learning and Industry Intelligence, Ministry of Education, Chengdu, Sichuan, China"],"raw_orcid":"https://orcid.org/0009-0008-7273-9302","affiliations":[{"raw_affiliation_string":"College of Computer Science, Sichuan University, Chengdu, Sichuan, China and Engineering Research Center of Machine Learning and Industry Intelligence, Ministry of Education, Chengdu, Sichuan, China","institution_ids":["https://openalex.org/I4210125143"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5048742182","display_name":"Linchao He","orcid":"https://orcid.org/0000-0002-0562-3026"},"institutions":[{"id":"https://openalex.org/I24185976","display_name":"Sichuan University","ror":"https://ror.org/011ashp19","country_code":"CN","type":"education","lineage":["https://openalex.org/I24185976"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Linchao He","raw_affiliation_strings":["National Key Laboratory of Fundamental Science on Synthetic Vision, Sichuan University, Chengdu, Sichuan, China"],"raw_orcid":"https://orcid.org/0000-0002-0562-3026","affiliations":[{"raw_affiliation_string":"National Key Laboratory of Fundamental Science on Synthetic Vision, Sichuan University, Chengdu, Sichuan, China","institution_ids":["https://openalex.org/I24185976"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5003270633","display_name":"Jizhe Zhou","orcid":"https://orcid.org/0000-0002-2447-1806"},"institutions":[{"id":"https://openalex.org/I4210125143","display_name":"Chengdu University","ror":"https://ror.org/034z67559","country_code":"CN","type":"education","lineage":["https://openalex.org/I4210125143"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ji-Zhe Zhou","raw_affiliation_strings":["College of Computer Science, Sichuan University, Chengdu, Sichuan, China and Engineering Research Center of Machine Learning and Industry Intelligence, Ministry of Education, Chengdu, Sichuan, China"],"raw_orcid":"https://orcid.org/0000-0002-2447-1806","affiliations":[{"raw_affiliation_string":"College of Computer Science, Sichuan University, Chengdu, Sichuan, China and Engineering Research Center of Machine Learning and Industry Intelligence, Ministry of Education, Chengdu, Sichuan, China","institution_ids":["https://openalex.org/I4210125143"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5073535763","display_name":"Jiancheng Lv","orcid":"https://orcid.org/0000-0001-6551-3884"},"institutions":[{"id":"https://openalex.org/I4210125143","display_name":"Chengdu University","ror":"https://ror.org/034z67559","country_code":"CN","type":"education","lineage":["https://openalex.org/I4210125143"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiancheng Lv","raw_affiliation_strings":["College of Computer Science, Sichuan University, Chengdu, Sichuan, China and Engineering Research Center of Machine Learning and Industry Intelligence, Ministry of Education, Chengdu, Sichuan, China"],"raw_orcid":"https://orcid.org/0000-0001-6551-3884","affiliations":[{"raw_affiliation_string":"College of Computer Science, Sichuan University, Chengdu, Sichuan, China and Engineering Research Center of Machine Learning and Industry Intelligence, Ministry of Education, Chengdu, Sichuan, China","institution_ids":["https://openalex.org/I4210125143"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5028144381"],"corresponding_institution_ids":["https://openalex.org/I4210125143"],"apc_list":null,"apc_paid":null,"fwci":2.2665,"has_fulltext":false,"cited_by_count":2,"citation_normalized_percentile":{"value":0.90270544,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":91,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"10768","last_page":"10777"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10688","display_name":"Image and Signal Denoising Methods","score":0.9997000098228455,"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/T10688","display_name":"Image and Signal Denoising Methods","score":0.9997000098228455,"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/T11105","display_name":"Advanced Image Processing Techniques","score":0.9994999766349792,"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/T10052","display_name":"Medical Image Segmentation Techniques","score":0.9914000034332275,"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/image-restoration","display_name":"Image restoration","score":0.722100019454956},{"id":"https://openalex.org/keywords/smoothing","display_name":"Smoothing","score":0.5275999903678894},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.48980000615119934},{"id":"https://openalex.org/keywords/gradient-descent","display_name":"Gradient descent","score":0.4253999888896942},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.42289999127388},{"id":"https://openalex.org/keywords/deblurring","display_name":"Deblurring","score":0.4106999933719635},{"id":"https://openalex.org/keywords/stability","display_name":"Stability (learning theory)","score":0.4066999852657318},{"id":"https://openalex.org/keywords/bayesian-probability","display_name":"Bayesian probability","score":0.3855000138282776},{"id":"https://openalex.org/keywords/bayesian-inference","display_name":"Bayesian inference","score":0.3815000057220459},{"id":"https://openalex.org/keywords/diffusion","display_name":"Diffusion","score":0.3700000047683716}],"concepts":[{"id":"https://openalex.org/C106430172","wikidata":"https://www.wikidata.org/wiki/Q6002272","display_name":"Image restoration","level":4,"score":0.722100019454956},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5724999904632568},{"id":"https://openalex.org/C3770464","wikidata":"https://www.wikidata.org/wiki/Q775963","display_name":"Smoothing","level":2,"score":0.5275999903678894},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.48980000615119934},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.44449999928474426},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.4253999888896942},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.42289999127388},{"id":"https://openalex.org/C2777693668","wikidata":"https://www.wikidata.org/wiki/Q25053743","display_name":"Deblurring","level":5,"score":0.4106999933719635},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.4066999852657318},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.39320001006126404},{"id":"https://openalex.org/C107673813","wikidata":"https://www.wikidata.org/wiki/Q812534","display_name":"Bayesian probability","level":2,"score":0.3855000138282776},{"id":"https://openalex.org/C160234255","wikidata":"https://www.wikidata.org/wiki/Q812535","display_name":"Bayesian inference","level":3,"score":0.3815000057220459},{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.3700000047683716},{"id":"https://openalex.org/C126042441","wikidata":"https://www.wikidata.org/wiki/Q1324888","display_name":"Frame (networking)","level":2,"score":0.352400004863739},{"id":"https://openalex.org/C95923904","wikidata":"https://www.wikidata.org/wiki/Q6760420","display_name":"Marginal likelihood","level":3,"score":0.34769999980926514},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.3343000113964081},{"id":"https://openalex.org/C167966045","wikidata":"https://www.wikidata.org/wiki/Q5532625","display_name":"Generative model","level":3,"score":0.33079999685287476},{"id":"https://openalex.org/C47446073","wikidata":"https://www.wikidata.org/wiki/Q5165890","display_name":"Control theory (sociology)","level":3,"score":0.3230000138282776},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.31610000133514404},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.30889999866485596},{"id":"https://openalex.org/C12713177","wikidata":"https://www.wikidata.org/wiki/Q1900281","display_name":"Perspective (graphical)","level":2,"score":0.305400013923645},{"id":"https://openalex.org/C52421305","wikidata":"https://www.wikidata.org/wiki/Q1151499","display_name":"Particle filter","level":3,"score":0.29919999837875366},{"id":"https://openalex.org/C81184566","wikidata":"https://www.wikidata.org/wiki/Q1191895","display_name":"Conjugate gradient method","level":2,"score":0.2985000014305115},{"id":"https://openalex.org/C174576160","wikidata":"https://www.wikidata.org/wiki/Q1183700","display_name":"Deconvolution","level":2,"score":0.296999990940094},{"id":"https://openalex.org/C163294075","wikidata":"https://www.wikidata.org/wiki/Q581861","display_name":"Noise reduction","level":2,"score":0.2921999990940094},{"id":"https://openalex.org/C68710425","wikidata":"https://www.wikidata.org/wiki/Q5275442","display_name":"Diffusion process","level":3,"score":0.28540000319480896},{"id":"https://openalex.org/C114289077","wikidata":"https://www.wikidata.org/wiki/Q3284399","display_name":"Statistical model","level":2,"score":0.28519999980926514},{"id":"https://openalex.org/C2779304628","wikidata":"https://www.wikidata.org/wiki/Q3503480","display_name":"Face (sociological concept)","level":2,"score":0.2842999994754791},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.2827000021934509},{"id":"https://openalex.org/C143587482","wikidata":"https://www.wikidata.org/wiki/Q1543216","display_name":"Iterative and incremental development","level":2,"score":0.2827000021934509},{"id":"https://openalex.org/C9417928","wikidata":"https://www.wikidata.org/wiki/Q1070689","display_name":"Image processing","level":3,"score":0.2773999869823456},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.27219998836517334},{"id":"https://openalex.org/C176217482","wikidata":"https://www.wikidata.org/wiki/Q860554","display_name":"Metric (unit)","level":2,"score":0.26420000195503235}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1145/3746027.3754582","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3746027.3754582","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:2507.06656","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2507.06656","pdf_url":"https://arxiv.org/pdf/2507.06656","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:2507.06656","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2507.06656","pdf_url":"https://arxiv.org/pdf/2507.06656","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":[{"id":"https://openalex.org/G6362919342","display_name":null,"funder_award_id":"1082204112364","funder_id":"https://openalex.org/F4320335787","funder_display_name":"Fundamental Research Funds for the Central Universities"},{"id":"https://openalex.org/G6978877745","display_name":null,"funder_award_id":"62427820","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320335787","display_name":"Fundamental Research Funds for the Central Universities","ror":null}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":20,"referenced_works":["https://openalex.org/W1973286131","https://openalex.org/W1998419211","https://openalex.org/W2015810892","https://openalex.org/W2117539524","https://openalex.org/W2133665775","https://openalex.org/W2508457857","https://openalex.org/W2962770929","https://openalex.org/W2962785568","https://openalex.org/W2963814976","https://openalex.org/W2964125708","https://openalex.org/W2992838650","https://openalex.org/W3167568784","https://openalex.org/W3207918547","https://openalex.org/W3212516020","https://openalex.org/W4308950926","https://openalex.org/W4312497550","https://openalex.org/W4312933868","https://openalex.org/W4386071831","https://openalex.org/W4390872583","https://openalex.org/W4400524387"],"related_works":[],"abstract_inverted_index":{"Diffusion":[0,106],"models":[1],"have":[2],"shown":[3],"remarkable":[4],"promise":[5],"for":[6],"image":[7],"restoration":[8,18,95,146],"by":[9],"leveraging":[10],"powerful":[11],"priors.":[12],"Prominent":[13],"methods":[14],"typically":[15],"frame":[16],"the":[17,37,44,54,70,80,90,139],"problem":[19],"within":[20],"a":[21,28,32,108,119],"Bayesian":[22],"inference":[23],"framework,":[24],"which":[25],"iteratively":[26],"combines":[27],"denoising":[29],"step":[30],"with":[31],"likelihood":[33,73,81,121,140],"guidance":[34],"step.":[35],"However,":[36],"interactions":[38],"between":[39,69],"these":[40,59,87,99],"two":[41,115],"components":[42,60],"in":[43,79,138,159],"generation":[45,153],"process":[46,92],"remain":[47],"underexplored.":[48],"In":[49],"this":[50],"paper,":[51],"we":[52,66,101],"analyze":[53],"underlying":[55],"gradient":[56,74,82,110,126],"dynamics":[57],"of":[58],"and":[61,72,93,128,162],"identify":[62],"significant":[63],"instabilities.":[64],"Specifically,":[65],"demonstrate":[67,148],"conflicts":[68],"prior":[71],"directions,":[75],"alongside":[76],"temporal":[77],"fluctuations":[78,137],"itself.":[83],"We":[84],"show":[85],"that":[86,149],"instabilities":[88],"disrupt":[89],"generative":[91],"compromise":[94],"performance.":[96],"To":[97],"address":[98],"issues,":[100],"propose":[102],"Stabilized":[103],"Progressive":[104],"Gradient":[105],"(SPGD),":[107],"novel":[109],"management":[111],"technique.":[112],"SPGD":[113,150],"integrates":[114],"synergistic":[116],"components:":[117],"(1)":[118],"progressive":[120],"warm-up":[122],"strategy":[123],"to":[124,135,156],"mitigate":[125],"conflicts;":[127],"(2)":[129],"adaptive":[130],"directional":[131],"momentum":[132],"(ADM)":[133],"smoothing":[134],"reduce":[136],"gradient.":[141],"Extensive":[142],"experiments":[143],"across":[144],"diverse":[145],"tasks":[147],"significantly":[151],"enhances":[152],"stability,":[154],"leading":[155],"state-of-the-art":[157],"performance":[158],"quantitative":[160],"metrics":[161],"visually":[163],"superior":[164],"results.":[165],"Code":[166],"is":[167],"available":[168],"at":[169],"https://github.com/74587887/SPGD.":[170]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":1}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-25T00:00:00"}