Archives: Enable text searching and data mining

Services for refining archive material by a distributed workforce

Our service combines automatic text recognition and human intelligence for accurate text recognition, structural analysis and keyword assignment. The service is especially useful for fixing the Optical Character Recognition (OCR) results of old archive material. Corrected names of people and places enable accurate full-body text searches. Customers include national archives, libraries and media houses who want to make their massive archives far more useful.

---

Fix old newspaper archives by playing games at www.digitalkoot.fi

The National Library of Finland has a large on-line archive of old newspaper. In order to enable full-body text search, the text recognition results need to be fixed. Due to the very high amount of manual work required, the only practical way is to crowdsource the work from a volunteers. In order to make the concept engaging, people participate by playing two games.

First month results:
25000 visitors
2 million completed tasks
101220 minutes of voluntary work