Browse Prior Art Database

Method That Eliminates Maverick Prototypes in Online Handwriting Recognition

IP.com Disclosure Number: IPCOM000120387D
Original Publication Date: 1991-Apr-01
Included in the Prior Art Database: 2005-Apr-02
Document File: 2 page(s) / 72K

Publishing Venue

IBM

Related People

Fujisaki, T: AUTHOR [+3]

Abstract

Many online, handwriting recognition systems match an unknown character against prototype characters (*). The accuracy of these recognizers depends on the quality of the prototypes. The speed of recognition depends on the number of prototypes. The problem is to design prototype establishment procedures that optimize recognition accuracy and speed.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Method That Eliminates Maverick Prototypes in Online Handwriting
Recognition

      Many online, handwriting recognition systems match an
unknown character against prototype characters (*).  The accuracy of
these recognizers depends on the quality of the prototypes.  The
speed of recognition depends on the number of prototypes.  The
problem is to design prototype establishment procedures that optimize
recognition accuracy and speed.

      Desired characteristics of prototypes are:
(1)  The prototypes should have sufficient coverage, that is, there
should be at least one prototype for each distinct way of writing a
character (or recognition strategies that handle variation).  With
online handwriting, this includes variations in the number, order,
direction, and shape of the strokes of a character.  A stroke is the
writing from pen-down to pen-up.
(2)  Prototypes should have reasonable separation in prototype space.
In particular, a prototype character should be different from
prototypes of other alphabet characters; for example, prototypes for
0 and O should be different.
(3)  A prototype should be a good representation of a way of writing
a character.  This might be done by averaging a number of similarly
written characters or by choosing a suitable training character to
represent a distinct variation.

      Mavericks should be avoided.  A maverick is a poor
representation of a character, that is, a character that is not
written as intended, not captured accurately due to hardware
problems, or distorted in some fashion due to other problems.

      Our recognition system collects training characters from a
user's writing samples through a training scenar...