Computational Linguistics for Metadata Building (CLiMB)

Description: CLiMB (Computational Linguistics for Metadata Building) aims to improve subject access to image collections through the use of computational linguistic techniques. CLiMB-2, under the leadership of Judith Klavans, is funded by the Mellon Foundation from 2005-2007. CLiMB-1, also funded by the Mellon Foundation, was based at Columbia University under Judith Klavans' direction, from 2002 to 2004.

Implementation description: Two art history survey texts were scanned and encoded in TEI Lite for use in testing the feasibility of using computational linguistic techniques to associate texts and images. A list of image (plates) in the text is produced, then for each image, one or more paragraphs is identified that can be a source of descriptive (subject matter) metadata for indexing the image.


Rebecca Passonneau
Center for Computational Learning Systems (CCLS) Interchurch Center 475 Riverside Drive MC 7717 Columbia University New York, NY 10115
Tel: (212) 854-5132
Fax: (212) 854-5589

Last recorded change to this page: 2007-09-20  •  For corrections or updates, contact