Latin OCR Development Resources

Corpora & wordlists:

• Perseus
  • 240,000 attested words, 49,000 lemmata, manually-entered editions
• Bruce Robertson’s Rigaudon dictionary
  • 693,000 attested words, Latin from Uwe Springmann
• Open Greek and Latin Project
  • Machine-corrected versions of Lace OCR output
• Lace
  • Raw OCR output of Rigaudon (Gamera-based process)
• David Bamman: 11K Latin Texts
  • 3.9GB compressed, 11,261 texts, 1.38 billion tokens, OCR-derived
• CAMENA (Corpus Automatum Multiplex Electorum Neolatinitatis Auctorum)
  • 50M words
• Words program of the late William Whitaker
  • 1M word forms derived from 39,000 lemmata
• Spell checking lexicon for OpenOffice/LibreOffice constructed by Karl Zeiler
  • 129,000 word forms
• Latin words in the English Wiktionary
  • 655,434 word forms for 32,860 roots
• Springmann et al. canonical lexicon (forthcoming?)
  • 2M word forms, 70,000 lemmata
• Bibliotecha Latina IntraText
• CLTK corpora
  • The Latin Library - 108MB
  • LacusCurtius - 28MB (HTML)
  • Latin proper names - ~40K proper names
• Pleiades - 22,140 Latin place names
• Johann Ramminger’s Neulateinische Wortliste (Neolatin wordlist)
  • “The NLW now contains 19225 lemmas with 1928 Variants […] The Sigelliste of NLW contains 2675 authors and 7402 works abbreviations.”
• Bibliotecha Augustana
• Musisque Deoque and Poeti d’Italia
• Corpus Grammaticorum Latinorum (CGL)
• Corpus Thomisticum
• Croatiae auctores Latini (CroALa)
• Index Thomisticus Treebank
  • 263,425 tokens
• The PROIEL Treebank
  • 146,892 Latin words
• The Ancient Greek and Latin Dependency Treebanks
  • 53,143 Latin words
• Integrating Digital Papyrology (IDP)
• See also: Digital Critical Editions of Texts in Greek and Latin

Related projects & other efforts:

• OCR of Historical Printings of Latin Texts: Problems, Prospects, Progress, Springmann et al., 2014
  • Talk, Slides
• Improving OCR Accuracy for Classical Critical Editions, Boschetti et al., 2009
• Himeros.eu Latin OCR Trainings for Tesseract, Boschetti, 2008
  • Based on 5 pages of a 1475 incunabular edition of Augustine’s De civitate Dei
• A Document Recognition System for Early Modern Latin, Sravana & Crane, 2006
  • Based on Gamera and Perseus

Other related links:

• McGillivray, Barbara. Methods in Latin Computational Linguistics. Leiden: Brill, 2014.
• eMOP: Early Modern OCR Project
  • Special Characters, Unicode, and Early Modern English
  • PRImA Research Tools
• Towards a Network Model of the Coreness of Texts: An Experiment in Classifying Latin Texts Using the TTLab Latin Tagger, Mehler et al., 2014
• On OCR ground truths and OCR post-correction gold standards, tools and formats, Reynaert, 2014
• PoCoTo - an open source system for efficient interactive postcorrection of OCRed historical texts, Vobl et al., 2014
• A Fast Alignment Scheme for Automatic OCR Evaluation of Books, Yalniz & Manmatha, 2011
• OCR and the transformation of the Humanities, Crane, 2011
• Digitizing Latin Incunabula: Challenges, Methods, and Possibilities, Rydberg-Cox, 2009
• Automatic disambiguation of Latin abbreviations in early modern texts for humanities digital libraries, Rydberg-Cox, 2003

Non-Free Corpora

• Library of Latin Texts (LLT), Brepols
• Thesaurus linguae Latinae (TLL), De Gruyter
• PHI Latin Texts
• Bibliotheca Teubneriana Latina (BTL)
• LatinWordNet