Browse Prior Art Database

Block Letter Handprint Recognition Algorithm

IP.com Disclosure Number: IPCOM000044703D
Original Publication Date: 1984-Mar-01
Included in the Prior Art Database: 2005-Feb-06
Document File: 2 page(s) / 13K

Publishing Venue

IBM

Related People

Reed, MA: AUTHOR

Abstract

This article describes new technology for reliable and computationally expedient block letter recognition as coded characters in a touch-sensitive screen overlay or electronic tablet with minimal training to a specific individual style.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 2

Block Letter Handprint Recognition Algorithm

This article describes new technology for reliable and computationally expedient block letter recognition as coded characters in a touch-sensitive screen overlay or electronic tablet with minimal training to a specific individual style.

Block letters are hand drawn upper-case alphabetic and numeric characters. They are entered on a touch-sensitive screen overlay or an electronic tablet by means of a stylus. The location of the stylus on the surface at any point in time is sensed by the electronics of the tablet or overlay, and is available to the user in the form of integer (x,y) coordinates. The sampling rate of the device determines how often an (x,y) location is measured. The resolution of the device indicates how precisely each x or y value can be measured. Assume that the input medium has sufficient resolution (accurate to at least .003 inch, for example) and high enough sampling rate (100-200 points per second) to generate a digitization of the hand-drawn image to any desired degree of detail.

No training is required, i.e., it is not necessary for the operator to enter a sample block letter alphabet prior to using the system for the first time. Rather, a set of prestored character shapes representing typical variations in individual style is stored for each letter. The use of training is not precluded by the system, but is an optional means for potentially increasing accuracy by particularizing the recognition logic for an individual.

A recognition line is presented to the user in which block letter entries are made. The recognition line consists of a series of equally spaced rectangular boxes, into which one letter per box is entered. Hence, there are no segmentation related problems.

By considering the x and y displacement between adjacent points generated by t...