Proofreading a text
All OCR transcriptions must be proofread by human users to verify that the text on the scanned page matches the OCR-generated text.
Help:Beginner's guide to reliability - Wikisource, the free online library
OCR and other tools for Wikisource (Wikimania 2021)
- Video: OCR and other tools for Wikisource
- An overview of recent developments with tools for Wikisource, including the new OCR tool. Wikisource is a project that makes use of several technicalities to alleviate the user's job in reading texts. Optical character recognition (OCR) in this sense is central for the project, and is currently the object of several improvements. The Wikisource community follows these improvements with attention, as any other tool that can be integrated in the multi-language project. This presentation is an opportunity to have a glimpse on the Wishlist selection process done by Community Tech, and the new tools developed for Wikisource, with a particular focus on OCR.
A presentation at Wikimania annual conference 2021, with speakers Ruthven, Sam Wilson and Natalia Rodriguez.
An overview of recent developments with tools for Wikisource, including the new OCR tool.
Wikisource is a project that makes use of several technicalities to alleviate the user's job in reading texts. Optical character recognition (OCR) in this sense is central for the project, and is currently the object of several improvements. The Wikisource community follows these improvements with attention, as any other tool that can be integrated in the multi-language project. This presentation is an opportunity to have a glimpse on the Wishlist selection process done by Community Tech, and the new tools developped for Wikisource, with a particular focus on OCR.
Useful links:
- OCR Tool (web interface; usual interaction is via Page editing on Wikisources)
- WS Export
- IA Upload
- Community Tech team
- Tesseract OCR
Session Outcomes