Artificial Intelligence Registers the Special Collections
With a new digital tool, we will make cataloging our many hidden collections much faster.
Since the University Library of Southern Denmark was founded, we have continuously incorporated larger collections of literature; and often taken over the entire stock of books from older libraries. Among these, we have discovered many literary treasures, and we work hard to make them accessible. But over time, the number of unregistered works has piled up, and it would take us many decades to hand-catalog everything.
Artificial intelligence may be our salvation. Michael Monefeldt has developed the AI tool Katalogica; a web application that reads images of title pages and sorts the digitized text by title, author, publisher, etc. In this way, the program creates an entry in our catalogue in a fraction of the time it would take using traditional methods. Katalogica even reads the gothic letters in our oldest library books, and there is no need for advanced scanning equipment – Katalogica can interpret an image produced by a phone camera.
OCR (Optical Character Recognition) is not flawless, and the AI-generated entries must be viewed over. Therefore, Katalogica is equipped with an intuitive user interface that quickens proofreading before the entries are made available in the catalogue of the library.