{"id":"https://openalex.org/W7204863567","doi":"https://doi.org/10.48550/arxiv.2608.30727","title":"RailGen: Improving Railway Intrusion Detection via Agent-Guided Small-Scale Foreign Object Generation","display_name":"RailGen: Improving Railway Intrusion Detection via Agent-Guided Small-Scale Foreign Object Generation","publication_year":2026,"publication_date":"2026-08-31","ids":{"openalex":"https://openalex.org/W7204863567","doi":"https://doi.org/10.48550/arxiv.2608.30727"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2608.30727","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.30727","pdf_url":"https://arxiv.org/pdf/2608.30727","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":null,"license_id":null,"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/2608.30727","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5148871688","display_name":"Quan Hao","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hao, Quan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5047200044","display_name":"Ziyang Tao","orcid":"https://orcid.org/0009-0001-4320-8683","observed_orcids":["https://orcid.org/0009-0001-4320-8683"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tao, Ziyang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148879765","display_name":"Chenxi Zhang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Chenxi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148892755","display_name":"Yudong Wang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Yudong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148925013","display_name":"Rui Shi","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shi, Rui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5148916795","display_name":"Liguo Zhang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Liguo","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/T10036","display_name":"Advanced Neural Network Applications","score":0.9779999852180481,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.9779999852180481,"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/T11606","display_name":"Infrastructure Maintenance and Monitoring","score":0.00989999994635582,"subfield":{"id":"https://openalex.org/subfields/2205","display_name":"Civil and Structural Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10331","display_name":"Video Surveillance and Tracking Methods","score":0.003700000001117587,"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/small-object-detection","display_name":"small object detection","score":0.7960000038146973},{"id":"https://openalex.org/keywords/long-tailed-data","display_name":"long-tailed data","score":0.8539999723434448},{"id":"https://openalex.org/keywords/synthetic-data-generation","display_name":"synthetic data generation","score":0.7599999904632568},{"id":"https://openalex.org/keywords/multimodal-image-generation","display_name":"multimodal image generation","score":0.9660000205039978},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.6140000224113464},{"id":"https://openalex.org/keywords/focal-modulation","display_name":"focal modulation","score":0.5440000295639038},{"id":"https://openalex.org/keywords/focal-loss","display_name":"focal loss","score":0.9929999709129333},{"id":"https://openalex.org/keywords/dense-image-matching","display_name":"dense matching","score":0.5249999761581421},{"id":"https://openalex.org/keywords/railway-safety","display_name":"railway safety","score":0.382999986410141}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7271999716758728},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6388999819755554},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.61080002784729},{"id":"https://openalex.org/C35525427","wikidata":"https://www.wikidata.org/wiki/Q745881","display_name":"Intrusion detection system","level":2,"score":0.5821999907493591},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.5655999779701233},{"id":"https://openalex.org/C2776151529","wikidata":"https://www.wikidata.org/wiki/Q3045304","display_name":"Object detection","level":3,"score":0.5365999937057495},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.5236999988555908},{"id":"https://openalex.org/C2781238097","wikidata":"https://www.wikidata.org/wiki/Q175026","display_name":"Object (grammar)","level":2,"score":0.47200000286102295},{"id":"https://openalex.org/C52622490","wikidata":"https://www.wikidata.org/wiki/Q1026626","display_name":"Feature extraction","level":2,"score":0.450300008058548},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.4343000054359436},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.43160000443458557},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.4115000069141388},{"id":"https://openalex.org/C160633673","wikidata":"https://www.wikidata.org/wiki/Q355198","display_name":"Pixel","level":2,"score":0.4052000045776367},{"id":"https://openalex.org/C141353440","wikidata":"https://www.wikidata.org/wiki/Q182221","display_name":"Fuse (electrical)","level":2,"score":0.39980000257492065},{"id":"https://openalex.org/C160920958","wikidata":"https://www.wikidata.org/wiki/Q7662746","display_name":"Synthetic data","level":2,"score":0.3815999925136566},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.38109999895095825},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.37720000743865967},{"id":"https://openalex.org/C2779304628","wikidata":"https://www.wikidata.org/wiki/Q3503480","display_name":"Face (sociological concept)","level":2,"score":0.37389999628067017},{"id":"https://openalex.org/C167966045","wikidata":"https://www.wikidata.org/wiki/Q5532625","display_name":"Generative model","level":3,"score":0.35760000348091125},{"id":"https://openalex.org/C62354387","wikidata":"https://www.wikidata.org/wiki/Q875399","display_name":"Boundary (topology)","level":2,"score":0.33880001306533813},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.33329999446868896},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.31450000405311584},{"id":"https://openalex.org/C126422989","wikidata":"https://www.wikidata.org/wiki/Q93586","display_name":"Feature detection (computer vision)","level":4,"score":0.31380000710487366},{"id":"https://openalex.org/C83665646","wikidata":"https://www.wikidata.org/wiki/Q42139305","display_name":"Feature vector","level":2,"score":0.3091999888420105},{"id":"https://openalex.org/C2777735758","wikidata":"https://www.wikidata.org/wiki/Q817765","display_name":"Path (computing)","level":2,"score":0.3075000047683716},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.29840001463890076},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.2768000066280365},{"id":"https://openalex.org/C158251709","wikidata":"https://www.wikidata.org/wiki/Q354025","display_name":"Intrusion","level":2,"score":0.2711000144481659},{"id":"https://openalex.org/C64876066","wikidata":"https://www.wikidata.org/wiki/Q5141226","display_name":"Cognitive neuroscience of visual object recognition","level":3,"score":0.2700999975204468},{"id":"https://openalex.org/C2778572836","wikidata":"https://www.wikidata.org/wiki/Q380933","display_name":"Space (punctuation)","level":2,"score":0.2689000070095062},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.2653000056743622},{"id":"https://openalex.org/C193536780","wikidata":"https://www.wikidata.org/wiki/Q1513153","display_name":"Edge detection","level":4,"score":0.25519999861717224}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.30727","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.30727","pdf_url":"https://arxiv.org/pdf/2608.30727","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2608.30727","doi":"https://doi.org/10.48550/arxiv.2608.30727","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.30727","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2608.30727","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.30727","pdf_url":"https://arxiv.org/pdf/2608.30727","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/9","score":0.5378999710083008,"display_name":"Industry, innovation and infrastructure"}],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4414102130","https://openalex.org/W7204865247","https://openalex.org/W7104509107","https://openalex.org/W2961276435","https://openalex.org/W4211154234","https://openalex.org/W4417334752","https://openalex.org/W7116299965","https://openalex.org/W7126247969","https://openalex.org/W4409820058","https://openalex.org/W4414497476"],"abstract_inverted_index":{"Small-object":[0],"detection":[1,38,121,228],"under":[2],"long-tailed":[3,226],"data":[4],"distributions":[5],"is":[6],"a":[7,36,59,120],"fundamental":[8],"yet":[9],"challenging":[10,185],"problem":[11],"in":[12,154,197,229],"multimedia.":[13],"Railway":[14],"Foreign":[15],"Object":[16],"Detection":[17],"(RFOD)":[18],"epitomizes":[19],"this":[20,114],"challenge":[21],"with":[22,127,134,183],"easily":[23],"confused":[24],"small":[25,53],"intrusions":[26],"and":[27,52,85,107,141,178,195,199,202,213],"scarce":[28],"samples.":[29],"To":[30],"address":[31],"these":[32,184],"issues,":[33],"we":[34,116],"propose":[35,118],"generative-augmented":[37],"paradigm":[39,188,218],"that":[40,98,161],"leverages":[41],"multimodal":[42,60,222],"image":[43,61],"generation":[44,62],"to":[45,75,124,145,176],"enrich":[46],"the":[47,101,109,170,190],"feature":[48,102,111],"space":[49],"of":[50,104],"rare":[51],"objects.":[54],"We":[55],"first":[56],"construct":[57],"RailGen,":[58],"agent":[63],"based":[64],"on":[65,180],"large":[66],"models.":[67],"Under":[68],"semantic":[69],"constraints,":[70],"RailGen":[71,162],"automatically":[72],"invokes":[73],"tools":[74],"generate":[76,164],"railway":[77,156],"scenes,":[78],"calibrate":[79],"intrusion":[80,90],"positions,":[81],"extract":[82],"foreign":[83,167],"objects,":[84,168],"fuse":[86],"them":[87],"into":[88],"realistic":[89],"effects.":[91],"This":[92],"process":[93],"produces":[94],"high-quality":[95,165],"synthetic":[96],"samples":[97],"effectively":[99],"densify":[100],"representations":[103],"tail":[105],"classes":[106],"complete":[108],"small-object":[110,139,227],"space.":[112],"Within":[113],"paradigm,":[115],"further":[117],"FocalDEIM,":[119],"framework":[122],"designed":[123],"enhance":[125],"training":[126],"generated":[128],"data.":[129],"FocalDEIM":[130],"improves":[131],"dense":[132],"matching":[133],"Focal":[135,143],"Modulation":[136],"for":[137,225],"better":[138],"discrimination":[140],"adopts":[142],"Loss":[144],"emphasize":[146],"hard":[147],"samples,":[148,186],"thereby":[149],"alleviating":[150],"blurred":[151],"inter-class":[152],"boundaries":[153],"complex":[155],"scenes.":[157],"Experimental":[158],"results":[159],"demonstrate":[160],"can":[163],"small-scale":[166],"reducing":[169],"object":[171],"pixel":[172],"area":[173],"by":[174,193],"up":[175],"58x":[177],"13.85x":[179],"average.":[181],"Equipped":[182],"our":[187],"surpasses":[189],"baseline":[191],"DEIM":[192],"5.6%":[194],"7.5%":[196],"mAP@50":[198],"mAP@(50-95),":[200],"respectively,":[201],"outperforms":[203],"existing":[204],"state-of-the-art":[205],"methods.":[206],"Ablation":[207],"studies":[208],"verify":[209],"RailGen's":[210],"feature-space":[211],"enrichment":[212],"FocalDEIM's":[214],"boundary":[215],"discrimination.":[216],"The":[217],"provides":[219],"an":[220],"effective":[221],"generative":[223],"solution":[224],"safety-critical":[230],"applications.":[231]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-09-02T00:00:00"}