Browse Prior Art Database

Method for Defining Standard Stroke Numbers in Online Handwritten Character Recognition

IP.com Disclosure Number: IPCOM000108780D
Original Publication Date: 1992-Jun-01
Included in the Prior Art Database: 2005-Mar-22
Document File: 2 page(s) / 57K

Publishing Venue

IBM

Related People

Kitamura, K: AUTHOR

Abstract

Disclosed is a method for defining standard stroke numbers, especially for Kanji characters, in online handwritten character recognition.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 72% of the total text.

Method for Defining Standard Stroke Numbers in Online Handwritten Character Recognition

       Disclosed is a method for defining standard stroke
numbers, especially for Kanji characters, in online handwritten
character recognition.

      Since there are numerous categories of Kanji, reduction of the
number of candidates is a necessary pre-process for online
recognition.  The number of strokes in a written character is
generally used in the reduction pre-process, since it is simple and
effective.  For example, if the number of strokes in an input
character is N, the list of candidates is reduced to the categories
whose standard numbers of strokes are between N-M and N+M.  The
standard numbers are required because the number of strokes in a
written character may vary according to the writer and his/her
writing style.

      Existing methods take the standard numbers of strokes in Kanji
characters from Kanji dictionaries that define these numbers for
characters in block style.  Therefore, the numbers tend to be larger
than for characters in ordinary handwritten style.  To reduce the
number of candidates, it is more appropriate to use the numbers of
strokes in ordinary handwritten style.

      The disclosed method defines the standard numbers of strokes in
Kanji characters by using the distributions of the numbers of strokes
in handwritten character samples.  For example, the standard numbers
of strokes are defined as the average or most frequent numbe...