Transforming Text into Discovery: OCR Enrichment of Digital Collections in the University of Galway Library

University ofGalway.ie University ofGalway.ie Transforming Text into Discovery OCR Enrichment of Digital Collections in the University of Galway Library Oksana Dereza [email protected] 7 October 2025 Meet the Authors Open & Digital Research Team Marie-Louise Rouget Digital Archivist Cillian Joy Head of Open and Digital Research Chidi Egwu Full-stack Developer Oksana Dereza Digital Library Developer University ofGalway.ie University of Galway Digital Collections & Archives https://digital.library.universityofgalway.ie/ University ofGalway.ie Digital Workflow Digitisation Preservation Metadata Enrichment Access for humans Access for computers OCR pipeline Test Data •Imirce project: https://imirce.universityofgalway.ie/ •Letters and life stories documenting Irish emigration from the 1680s to 1990s •More than 5,000 letters and life stories released to date •Mostly typed transcriptions of handwritten letters by historian Kerby A. Miller and his research assistants, sometimes with handwritten annotations •Mostly in English, occasionally in Irish •10 English and 2 code-mixed pages selected for OCR software comparison University ofGalway.ie •Paper defects •Low contrast •Double-scanning artifacts •Irregular layout •Occasional handwritten letters •Handwritten annotations of typewritten text •Mixed language (English + Irish) •Spoken language, spelling errors Challenges University ofGalway.ie •System: must work on Windows •Procurement: can’t be a monthly subscription •Language: must work with at least English and Irish •Script: must be able to handle printed and typewritten text (ideally, also handwritten English) •Output Format: ALTO XML or easily convertible to ALTO XML (e.g. JSON) •Interoperability: must be interoperable with our DAMS and storage solutions Constraints Software Comparison Full-text item search University ofGalway.ie Machine Transcripts University ofGalway.ie University ofGalway.ie Handwritten Text University ofGalway.ie Irish Language University ofGalway.ie Other collections Journals by Political Prisoners from Long Kesh prison camp from the 1970s Threshold Literary Journal Next steps Data Enrichment •Include OCR as a necessary step in the publication of new collections •Enrich audio and video files with transcriptions •Extract information from text (personal names, placenames, dates, keywords etc.) → oAutomatically populate some metadata fields oEnhance advanced search oProvide visual markup University ofGalway.ie •Represent metadata as linked data •Automate metadata creation to assist archivists and streamline workflows •Structure collections as datasets with scheduled updates •Develop APIs to query metadata and retrieve datasets (images, extracted text, annotations, metadata) •Provide examples of how to use APIs & datasets •Integrate with European cultural heritage platforms, such as Europeana & EOSC Collections as Data University ofGalway.ie University ofGalway.ie Questions?