Automatized knowledge extraction from paper documents
Guillaume Rousse and Eric de la Clergerie, INRIAA large amount of knowledge in systematics is still available from non-numerical documents only, such as books and photos, whereas most computer projects are hampered by data input problems. The ongoing Biotim project explores the means to reduce this gap, by providing automatized acquisition of structured data directly from raw paper documents, for text and image simultaneously. This poster presents global project objectives, and focuses on current state of the text analysis process.