After a period of 3.5 years, we have successfully completed the volunteer project Goetgevonden!.
This was possible thanks to a dedicated group of volunteers who manually corrected as many as 50,000 pages of scans with handwriting recognition, allowing us to improve our HTR model. As a result, we can now automatically read and search the remaining 500,000-plus scans.
This project was conducted through the VeleHanden.nl platform, where volunteers faced historical documents and corrected text recognition. The most active 10 volunteers entered an average of as many as 1,722 scans per person, with the front-runner entering as many as 9,450 scans. This extraordinary level of public involvement goes to the heart of our project: the goal of making large amounts of resources digitally accessible through public participation.
Recently, the closing meeting of the volunteer project was held through Zoom. The most active volunteers moved on to the next phase of the project, called Structureer de Besluiten! In this new project, they will tag the various structural elements of the Resolutions to further improve automatic retrieval. Stay up to date with our project through our website and on Twitter/X.