THOTH – Transcribing historical objects with tabulated handwriting

1. Abstract

We present a new computer vision and machine learning tool for extracting data from historical sources. This tool has application in the humanities and the social sciences, public archives and libraries. It is being developed as a collaboration between Cambridge-based researchers in history and computer sciences. It develops on existing machine learning tech for manuscript recognition to machine read historical table structures and to more rapidly extract these into modern relational databases.

Oliver Dunn (, Cambridge University, United Kingdom and Alexis Litvine (, Cambridge University, United Kingdom

Theme: Lux by Bootswatch.