In this paper, we present a method for the recognition of handwritten words extracted from real incoming mail documents. The word recognition process is based on three different sources of information: outputs of a character classifier, contextual information extracted from word shapes and some a priori knowledge. Reported results demonstrate the benefit of those additional information on the word recognition rates. This approach is evaluated on a database of 5000 words examples. © 2004 IEEE.
Mendeley saves you time finding and organizing research
Choose a citation style from the tabs below