Skip to Content
Find More Like This
Return to Search

Script identification from images using cluster-based templates

United States Patent

5,844,991
December 1, 1998
View the Complete Patent at the US Patent & Trademark Office
Los Alamos National Laboratory - Visit the Technology Transfer Division Website
A computer-implemented method identifies a script used to create a document. A set of training documents for each script to be identified is scanned into the computer to store a series of exemplary images representing each script. Pixels forming the exemplary images are electronically processed to define a set of textual symbols corresponding to the exemplary images. Each textual symbol is assigned to a cluster of textual symbols that most closely represents the textual symbol. The cluster of textual symbols is processed to form a representative electronic template for each cluster. A document having a script to be identified is scanned into the computer to form one or more document images representing the script to be identified. Pixels forming the document images are electronically processed to define a set of document textual symbols corresponding to the document images. The set of document textual symbols is compared to the electronic templates to identify the script.
Hochberg; Judith G. (Los Alamos, NM), Kelly; Patrick M. (Los Alamos, NM), Thomas; Timothy R. (Santa Fe, NM)
The Regents of the University of California (Oakland, CA)
08/ 511,338
August 7, 1995
BACKGROUND OF THE INVENTION This invention relates to document identification, and, more particularly, to the identification of the script used in document preparation. This invention was made with government support under Contract No. W-7405-ENG-36 awarded by the U.S. Department of Energy. The government has certain rights in the invention.