15 years of e-rara: New search options and enhanced full-text recognition
New ways of accessing content by places, people and topics, as well as advanced full-text recognition, make it easier to access digitised print material.
The external page e-rara platform is celebrating its 15th anniversary and has undergone significant development since its inception with continuous growth in content alongside constant functional and technical enhancement. Since 2025, the homepage has featured entry points for the entity types 'external page Persons', 'external page Places' and 'external page Subjects', with the integration of advanced OCR tools facilitating full-text recognition.
‘Genfer See’, ‘Lake Geneva’ or ‘Lacus Lemanus’ – all results with one search on e-rara
The platform's functionality has evolved to encompass NER and NEL, sophisticated algorithms that facilitate the identification and linking of named entities – persons, subjects, and places – within full texts, thereby integrating these entities with the GND authority database. This enhancement ensures that all existing authority records are accessible for exploration on e-rara, enhancing the platform's capacity for comprehensive research. A search for 'Genfersee' in the place search entry yields not only identical results, but also text passages where Lake Geneva occurs in other languages or spellings.
The person search can be used to search for a specific person in the same way, with the list of search results filterable according to person and corporate body types, professions, places of activity or dates of life.
For instance, a search for 'Pestalozzi' can be filtered to the profession 'cartographer' to find the Zurich engineer external page Heinrich Pestalozzi (1790-1857) and not the much more common pedagogue external page Johann Heinrich Pestalozzi (1746-1827).
Improved full-text recognition for e-rara
The basis for NER and NEL is high-quality full-text recognition in digitised prints. New OCR options have been available in e-rara for this purpose since 2024, with the previously used solution Abbyy Finereader being joined by the open-source software Tesseract. Thanks to specifically trained language and font models, Tesseract also achieves good results for older prints.
Depending on the trained model used, Tesseract also recognises special characters such as ſ, ꝛ, or uͤ.
The new indexing and OCR options at e-rara are technically based on the Textlab module in the backend, which facilitates the integration of various OCR and NER/NEL solutions. As previously, full texts can be searched directly on e-rara or downloaded as plain text, AltoXML or PDF. Further information on NER/NEL and OCR on e-rara can be found on the external page corresponding info page on e-rara.
external page e-rara is the platform for digitised prints from Swiss libraries.