A knowledge-based segmentation algorithm for enhanced recognition of handwritten courtesy amounts

A knowledge-based segmentation algorithm to enhance recognition of courtesy amounts on bank checks is proposed in this paper, This algorithm uses mult iple contextual cues to enhance segmentation and recognition. The system de scribed extracts context from the handwritten numerals and uses a syntax pa rser based on a deterministic finite automaton to provide adequate feedback to enhance recognition. Further feedback is provided by a simple legal amo unt decoder that determines word count and recognizes several key words (e. g. thousand and hundred). This provides an additional semantic constraint o n the dollar section. The segmentation analysis module presented is capable of handling a number of commonly used styles for courtesy amount represent ation, Both handwritten and machine written courtesy and legal amounts were utilized to test the efficacy of the preprocessor for the check recognitio n system described in this paper. The substitution error was reduced by 30- 40% depending on the input check mix. (C) 1999 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.