Text Recognition and Collocations and Domain Codes
Tony G. Rose, Lindsay J. Evett · Meeting of the Association for Computational Linguistics · 1993
Text recognition systems require the use of contextual information in order to maximise the accuracy of their output. However, the acquisition of such knowledge for a realistically sized vocabulary presents a major problem. This paper describes methods for extracting contextual knowledge from text corpora, and demonstrates its contribution to the performance of handwriting recognition systems.