Thursday 16th February
Speaker: Ngoni Munyaradzi
Theme: Citizen cyberscience around the world
Research questions: Can we use volunteered thinking to help us transcribe the Bushman historical text? How does this compare to automated machine learning techniques? How effective are cell phones as a platform for these operations?
The Bushman text under research is from a private collection of cultural tales. A technical challenge that has arisen is how to segment the images of Bushman text. A segmentation application has been created, but perfect segmentation is impossible because the text is written in free form handwriting and there is no consistency between texts (e.g. some writing is tightly packed, other texts can be widely spaced).
Research aims/impact:
- Provide a generic solution applicable to other historical collections
- Preservation of Bushman Historical text
- Make text searchable
- Indexing text
- Reprint of text in books
- Text-to-Speech potential for future?
This project is not part of the Endangered Languages Project.
Extreme Citizen Science blog 