Algorithms for the recognition of handwriting in real-time

Wright, P.T., 1989. Algorithms for the recognition of handwriting in real-time. PhD, Nottingham Trent University.

Preview

Text 10182985.pdf - Published version Download (30MB) | Preview

Abstract

This thesis details the work undertaken by the author from September 1984 to September 1988 into the field of dynamic script recognition. It reviews the various techniques developed since 1960 and it analyses the more popular approaches to processing the raw pen motion information. It also details the progress made in the nature of the user interface over the last 28 years.

The main emphasis of the work has been the development of algorithms capable of recognising, in a real-time user independent environment, lower case hand-printing. In particular, the design of the character shape databases provides for rapid searching and character matching and the techniques of feature reduction provide character matches to be found from previously original character encodings.

The most successful algorithm, based on the method of curve encoding by H.Freeman, forms a foundation towards the development of natural user text and data entry system. Extension of the character base is also possible with no alteration to the basic algorithm methodology. A technique of robust word segmentation has been designed that has enabled the design of a prototype cursive script recognition system. This is presently writer independent, running on a 68020 micro-processor. Initial results show a word level recognition rate of 95+ %. Development of natural editing functions provides a self contained text entry environment.

In the future, the algorithms will be ported to an 'electronic paper' environment and a user training phase will be designed as a front end to the recogniser.

Item Type:	Thesis
Creators:	Wright, P.T.
Date:	1989
ISBN:	9781369312836
Identifiers:	Number Type PQ10182985 Other
Rights:	This copy of the thesis has been supplied on condition that anyone who consults it is understood to recognise that its copyright rests with its author and that no quotation from the thesis and no information derived from it may be published without the authors prior written consent.
Divisions:	Schools > School of Science and Technology
Record created by:	Linda Sullivan
Date Added:	28 Aug 2020 09:29
Last Modified:	15 Jun 2023 09:37
URI:	https://irep.ntu.ac.uk/id/eprint/40556

Actions (login required)

Edit View

Views

Downloads