{"id":"https://openalex.org/W7225336773","doi":"https://doi.org/10.21203/rs.3.rs-11348739/v1","title":"Handwritten German Administrative Forms (HGAF): A Dataset for Handwriting Recognition with Fine-Tuned and Zero-Shot Baselines","display_name":"Handwritten German Administrative Forms (HGAF): A Dataset for Handwriting Recognition with Fine-Tuned and Zero-Shot Baselines","publication_year":2026,"publication_date":"2026-11-05","ids":{"openalex":"https://openalex.org/W7225336773","doi":"https://doi.org/10.21203/rs.3.rs-11348739/v1"},"language":"en","primary_location":{"id":"doi:10.21203/rs.3.rs-11348739/v1","doi":"https://doi.org/10.21203/rs.3.rs-11348739/v1","is_oa":true,"landing_page_url":"https://doi.org/10.21203/rs.3.rs-11348739/v1","pdf_url":"https://www.researchsquare.com/article/rs-11348739/latest.pdf","source":{"id":"https://openalex.org/S4306525896","display_name":"Research Square","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"Research Square","raw_type":"posted-content"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://www.researchsquare.com/article/rs-11348739/latest.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5118969044","display_name":"Tobias Kalmbach","orcid":"https://orcid.org/0009-0005-7447-2195","observed_orcids":["https://orcid.org/0009-0005-7447-2195"]},"institutions":[{"id":"https://openalex.org/I70886390","display_name":"Karlsruhe University of Applied Sciences","ror":"https://ror.org/01c0m1t63","country_code":"DE","type":"education","lineage":["https://openalex.org/I70886390"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Tobias Kalmbach","raw_affiliation_strings":["Karlsruhe University of Applied Sciences"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Karlsruhe University of Applied Sciences","institution_ids":["https://openalex.org/I70886390"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5007820528","display_name":"Andreas Wagner","orcid":"https://orcid.org/0000-0001-5772-1261","observed_orcids":["https://orcid.org/0000-0001-5772-1261"]},"institutions":[{"id":"https://openalex.org/I70886390","display_name":"Karlsruhe University of Applied Sciences","ror":"https://ror.org/01c0m1t63","country_code":"DE","type":"education","lineage":["https://openalex.org/I70886390"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Andreas Wagner","raw_affiliation_strings":["Karlsruhe University of Applied Sciences"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Karlsruhe University of Applied Sciences","institution_ids":["https://openalex.org/I70886390"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I70886390"],"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/T10601","display_name":"Handwritten Text Recognition Techniques","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/T10601","display_name":"Handwritten Text Recognition Techniques","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/T12707","display_name":"Vehicle License Plate Recognition","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/T11398","display_name":"Hand Gesture Recognition Systems","score":0.0,"subfield":{"id":"https://openalex.org/subfields/1709","display_name":"Human-Computer Interaction"},"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/handwriting-recognition","display_name":"handwriting recognition","score":0.7170000076293945},{"id":"https://openalex.org/keywords/shape-recognition","display_name":"shape recognition","score":0.3400000035762787},{"id":"https://openalex.org/keywords/vision-language-models","display_name":"vision-language models","score":0.6809999942779541},{"id":"https://openalex.org/keywords/zero-shot-learning","display_name":"zero-shot learning","score":0.8209999799728394},{"id":"https://openalex.org/keywords/fine-tuning","display_name":"fine-tuning","score":0.9089999794960022},{"id":"https://openalex.org/keywords/trocr","display_name":"TrOCR","score":0.7990000247955322},{"id":"https://openalex.org/keywords/character-error-rate","display_name":"character error rate","score":0.781000018119812},{"id":"https://openalex.org/keywords/structured-information-extraction","display_name":"structured data extraction","score":0.6930000185966492}],"concepts":[{"id":"https://openalex.org/C2779386606","wikidata":"https://www.wikidata.org/wiki/Q2393642","display_name":"Handwriting","level":2,"score":0.796500027179718},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7583000063896179},{"id":"https://openalex.org/C154775046","wikidata":"https://www.wikidata.org/wiki/Q188","display_name":"German","level":2,"score":0.748199999332428},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.6862999796867371},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.636900007724762},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5266000032424927},{"id":"https://openalex.org/C2780861071","wikidata":"https://www.wikidata.org/wiki/Q1062934","display_name":"Character (mathematics)","level":2,"score":0.47200000286102295},{"id":"https://openalex.org/C112640561","wikidata":"https://www.wikidata.org/wiki/Q2440634","display_name":"Handwriting recognition","level":3,"score":0.4578999876976013},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.40610000491142273},{"id":"https://openalex.org/C40969351","wikidata":"https://www.wikidata.org/wiki/Q3516228","display_name":"Word error rate","level":2,"score":0.38839998841285706},{"id":"https://openalex.org/C2776299755","wikidata":"https://www.wikidata.org/wiki/Q432449","display_name":"Carry (investment)","level":2,"score":0.3725999891757965},{"id":"https://openalex.org/C26983874","wikidata":"https://www.wikidata.org/wiki/Q263864","display_name":"Tag cloud","level":3,"score":0.34209999442100525},{"id":"https://openalex.org/C79974875","wikidata":"https://www.wikidata.org/wiki/Q483639","display_name":"Cloud computing","level":2,"score":0.2946999967098236},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.28200000524520874},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.27720001339912415},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.2671000063419342},{"id":"https://openalex.org/C67905146","wikidata":"https://www.wikidata.org/wiki/Q5287646","display_name":"Document processing","level":2,"score":0.262800008058548},{"id":"https://openalex.org/C44868376","wikidata":"https://www.wikidata.org/wiki/Q3099089","display_name":"Intelligent character recognition","level":4,"score":0.2596000134944916},{"id":"https://openalex.org/C2987247673","wikidata":"https://www.wikidata.org/wiki/Q167555","display_name":"Character recognition","level":3,"score":0.2542000114917755},{"id":"https://openalex.org/C546480517","wikidata":"https://www.wikidata.org/wiki/Q167555","display_name":"Optical character recognition","level":3,"score":0.2517000138759613}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.21203/rs.3.rs-11348739/v1","doi":"https://doi.org/10.21203/rs.3.rs-11348739/v1","is_oa":true,"landing_page_url":"https://doi.org/10.21203/rs.3.rs-11348739/v1","pdf_url":"https://www.researchsquare.com/article/rs-11348739/latest.pdf","source":{"id":"https://openalex.org/S4306525896","display_name":"Research Square","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"Research Square","raw_type":"posted-content"}],"best_oa_location":{"id":"doi:10.21203/rs.3.rs-11348739/v1","doi":"https://doi.org/10.21203/rs.3.rs-11348739/v1","is_oa":true,"landing_page_url":"https://doi.org/10.21203/rs.3.rs-11348739/v1","pdf_url":"https://www.researchsquare.com/article/rs-11348739/latest.pdf","source":{"id":"https://openalex.org/S4306525896","display_name":"Research Square","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"Research Square","raw_type":"posted-content"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7225336773.pdf","grobid_xml":"https://content.openalex.org/works/W7225336773.grobid-xml"},"referenced_works_count":27,"referenced_works":["https://openalex.org/W1982907973","https://openalex.org/W1993155451","https://openalex.org/W2104806940","https://openalex.org/W2152928267","https://openalex.org/W2906791574","https://openalex.org/W2947697927","https://openalex.org/W2971751059","https://openalex.org/W3178709966","https://openalex.org/W3181798399","https://openalex.org/W4240568764","https://openalex.org/W4285105124","https://openalex.org/W4285125114","https://openalex.org/W4285167662","https://openalex.org/W4312750536","https://openalex.org/W4382202677","https://openalex.org/W4385819993","https://openalex.org/W4386262694","https://openalex.org/W4390670543","https://openalex.org/W4405399726","https://openalex.org/W4406031111","https://openalex.org/W4410826117","https://openalex.org/W4413156663","https://openalex.org/W4413256343","https://openalex.org/W4414311642","https://openalex.org/W4417470555","https://openalex.org/W6944211801","https://openalex.org/W7154724451"],"related_works":["https://openalex.org/W6950612055","https://openalex.org/W7171933600","https://openalex.org/W7204208786","https://openalex.org/W2973679659","https://openalex.org/W7204471617","https://openalex.org/W2966671722","https://openalex.org/W2395466929","https://openalex.org/W7204059083","https://openalex.org/W2065986632","https://openalex.org/W7155087914"],"abstract_inverted_index":{"Abstract":[0],"Administrative":[1,60],"processes":[2],"remain":[3],"largely":[4],"paper-based,and":[5],"handwritten":[6],"form":[7,45,215],"contents":[8],"must":[9],"be":[10,211],"transferred":[11],"into":[12],"structured":[13],"data.":[14],"Whereforms":[15],"carry":[16],"personal":[17,96],"information,":[18],"the":[19,128,146,154,172,187,208],"proprietary":[20],"cloud":[21],"services":[22],"that":[23],"leadpublished":[24],"benchmarks":[25],"are":[26],"not":[27,178],"suitable,":[28],"so":[29],"only":[30,129],"on-premise":[31,157],"models":[32,120,135],"remain.":[33],"Currently,":[34],"nopublic":[35],"corpus":[36,93],"supports":[37],"this":[38,168],"setting:":[39],"handwriting":[40],"datasets":[41,46],"hold":[42,47],"continuous":[43],"prose,and":[44],"printed":[48],"text.":[49],"Therefore,":[50],"a":[51,109,112,194,198],"dataset":[52],"of":[53,163],"291":[54],"German":[55],"administrative":[56],"forms,":[57],"named":[58],"HandwrittenGerman":[59],"Forms":[61],"(HGAF),":[62],"was":[63],"collected":[64],"from":[65],"95":[66],"writers.":[67],"Each":[68],"writercompleted":[69],"three":[70,118],"layouts,":[71],"A,":[72,127],"B":[73,151],"and":[74,79,98,105,117,138,152,206],"C,":[75,153],"with":[76,197],"identical":[77],"fields":[78],"identicalsynthetic":[80],"contents.":[81],"The":[82,92],"visual":[83],"structure":[84],"therefore":[85],"varies,":[86],"while":[87],"writer":[88],"andcontent":[89],"stay":[90],"constant.":[91],"carries":[94],"no":[95],"information":[97],"can":[99],"bereleased.":[100],"Two":[101],"approaches":[102],"were":[103],"implemented":[104],"scored":[106],"per":[107],"field:":[108],"specialistpipeline":[110],"around":[111],"fine-tuned":[113],"line-level":[114],"TrOCR":[115],"model,":[116],"open-weightvision-language":[119],"(VLM)":[121],"applied":[122],"without":[123,165],"fine-tuning.":[124],"On":[125,149],"layout":[126,130],"used":[131],"for":[132,145],"fine-tuning,":[133],"twozero-shot":[134],"reach":[136],"4.5%":[137],"4.7%":[139],"character":[140],"error":[141],"rate":[142],"(CER),":[143],"against6.8%":[144],"specialist":[147],"pipeline.":[148],"layouts":[150],"gap":[155],"widens.No":[156],"configuration":[158],"transcribed":[159],"more":[160],"than":[161],"7.7%":[162],"forms":[164],"anerror.":[166],"In":[167],"scenario,":[169],"VLMs":[170],"reached":[171],"best":[173],"performance.":[174],"Although":[175],"they":[176,180],"do":[177],"requirefine-tuning,":[179],"may":[181],"still":[182],"require":[183],"configurationadjustments":[184],"to":[185,210,213],"fulfil":[186],"HTR":[188],"task":[189],"properly.":[190],"We":[191],"recommend":[192],"prompting":[193],"generalinstruction-following":[195],"model":[196],"fixed":[199],"output":[200],"schema.":[201],"This":[202],"removes":[203],"theextraction":[204],"stage":[205],"allows":[207],"system":[209],"adapted":[212],"different":[214],"layoutswith":[216],"little":[217],"effort.":[218]},"counts_by_year":[],"updated_date":"2026-11-09T13:22:03.289127","created_date":"2026-11-06T00:00:00"}