For more than twenty years, Optical Character Recognition (OCR) technology has been available for converting scans of printed and typewritten pages into machine-readable text. It was only with the introduction of deep learning technologies in the last decade that comparable results could also be achieved for handwritten documents. Transkribus is a leading Handwritten Text Recognition (HTR) platform developed by the “READ” project (Recognition and Enrichment of Archival Documents) within the European Commission’s Seventh Framework Program. We are planning to integrate Transkribus into Migrant Connections to complement the manual transcription of letters and diaries.
In an article for href, former GHI intern Janna Katharina Müller describes how HTR can also improve the transcription of historical prints for typefaces not well supported by traditional OCR solutions. Read the article…
OpenEdition suggests that you cite this post as follows:
migconn (February 12, 2022). Handwritten Text Recognition for Historical Prints. Migrant Connections. Retrieved July 20, 2026 from https://migconn.hypotheses.org/26