archiscribe

Use facsimile images and OCR layout information from Archive.org to quickly transcribe OCR ground truth with a user-friendly webinterface.

Currently hard-coded to sample from 19th century German books typeset in Fraktur, but that can easily be changed.

Based on a python prototype Created as part of the OCR Workshop at the BBAW in Berlin, 28/29th. September 2017, ported to Go for better performance and concurrency.