<?xml version="1.0" standalone="yes"?>
<Paper uid="H94-1085">
  <Title>USE OF LEXICAL AND SYNTACTIC TECHNIQUES IN RECOGNIZING HANDWRITTEN TEXT</Title>
  <Section position="3" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1. INTRODUCTION
</SectionTitle>
    <Paragraph position="0"> This paper focuses on the use of human language models in performing handwriting recognition. Systems that recognize handwriting are referred to as off-line or on-line systems, depending on whether ordinary handwriting on paper is scanned and digitized or a special stylus and a pressuresensitive tablet are used. The central component of a hand-written text recognizer is a word recognizer (Wit) which takes as input, a word signal and a lexicon. Its output consists of an ordered list of the best n words in the lexicon which match the word signal. Due to wide variability in writing, WRs often do not return the correct word as the top choice and get worse as the lexicon size increases. Furthermore, the correct word may not even be present in the top n choices. This is illustrated in Figure 1 which shows the output of an actual word recognizer (offiine) on isolated word images.</Paragraph>
    <Paragraph position="1"> imately 200 words on the average. In the second stage, the word-image is segmented into several components; physical features of each component lead to a set of character choices for each segment thus resulting in a set of candidate words.</Paragraph>
    <Paragraph position="2"> All candidate words which are in the lexicon are returned as the direct recognition output of the Wit. In case none of the words are found in the lexicon (,~ 62% of the time), string matching (the third stage) is performed.</Paragraph>
    <Paragraph position="3"> Since the training phase (of the language module) requires the processing of several thousand sentences, the computationally expensive procedure of digitizing followed by recognition is avoided by employing a program which simulates the output of an actual WR. Based on the intermediate results of the actual word recognizer, we have computed statistics which model the behavi'our of the second stage 1 . These include substitution, splitting and merging statistics. Given an input (ASCII) word, and the above statistics, candidate (corrupted) words are generated based on simulating and propogating each of the above three types of errors at each character position. The string matching algorithm used in the simulator is the same as that used in the actual WR.</Paragraph>
    <Paragraph position="4"> Figure 2 illustrates the entire model for recognizing handwritten text. The ultimate goal of language models is to provide feedback to the word recognizer as indicated by the dashed lines in Figure 2. There are two types of feedback provided: (i) feedback information to the Wit post-processor in terms of eliminating syntactic categories from contention, or (ii) feedback to word recognition e.g., if syntactic analysis has determined that a particular token must be alphabetic only (as opposed to mixed alphanumeric), this information could be incorporated in a second &amp;quot;reading&amp;quot; of the word image. This necessitates the use of linguistic constraints (which employ phrase and sentence-level context) to achieve a performance level comparable to that of humans \[1, 2\]. We present two techniques, (i) lexical analysis using collocations, and (ii) syntactic (n-gram) analysis using part-of-speech (POS) tags, both designed to improve the WR rate.</Paragraph>
  </Section>
class="xml-element"></Paper>