Abstract

In earlier papers [6, 8], we presented an HMM-based OCR system referred to as the BBN BYBLOS OCR system in this paper, incorporating the BBN BYBLOS continuous speech recognition system. The BYBLOS OCR system uses a character model trained on a corpus of text images, a lexicon and a grammar. A brief review of the BYBLOS OCR system is provided in the following section.

Date

1999

Authors

Issam Bazzi, Premkumar Natarajan, Richard Schwartz, Andras Kornai, Zhidong Lu, John Makhoul

Journal

Proceedings Symposium on Document Image Understanding Technology

Pages

149