‘ "1*

= - . =11112‘gmr
1' 1 ' “.3.“
hi [139

.nn
w
v

-;:r..:‘
:gr“.
. J:-
M
s
1‘"

an"
inn
*3)"
1.13

. )1

1 mi

“£5.11

v sum"

.-
gu-
1.7%.? , ~

., u.
W”

W .
‘flifim
“J
..

1-1
“5.:

‘ ‘3‘
"3‘21

‘ , .. .
. Zapm 1
. .4 -
.. ‘-
,3».._.¢

......... "7;, _ “t

m»

a... .

u .
um"

oufl’

. .2

..
.. '. *
'z‘t'lql
" ME ‘3

 

‘ I

I‘.
3. 5);:
‘ 1

:33”
1' “‘3;

\

7.3.1.!

v"- 3-}
.n‘ '
“E3?

1;" 5?; h

51 1332.31: :
'.§ .0?
1‘“

 

.- .E
.3 , - mam-
1‘ H ‘ E. “ "a a , ‘ ‘
v n‘?” . ‘ . [I ' K. ‘ ‘1’. ' 21!. ‘3 . ‘ ‘ a“ ‘4'. .4
1 -' 11.. ‘ . «"J-vresh was}.
J: a. Y} H , . fl (0%. 95:;&’—1-‘L:fi
‘ fl . r, "’E: ".11.” ' d1 \‘ 8:5.
“:‘1‘1‘1‘ 7 . ‘3 ‘ v , ‘ {NR ‘kifi’fi' 0 IX :7" u
3 1 : ,,. .2. ..,.-.a . «’1‘
‘J g: Hagan?" $3
3-? “ (9,:
v ,gzgiv‘léyzfi
$91,121:!
3 0 ta.)
await"
‘ z.
2 a
‘ ' ‘2 | 1 :q . . ‘ ~
. 4:195? ‘ k ‘ 5* l “" A z "5‘ " m4 . : iii ~ L
.. , .. .. . .. , ‘ A
.vL’t‘flgli‘ i 5%” t . l a ‘l' .\ 4" ‘ . t- 43”“ f 2; ‘1 'h ' “an“ ‘ £9
'11.; i. 1 ' ‘ - z! - mgr .
{asnis :, , . , ' ‘
1 Huhp _ g; ‘

A , lg
aih'ufllih: 5:,

THE 513

5L

 

Date

llllllllllHllllllllllllllllllllll‘llllllllllllllllllllllllll

3 1293 013893

 

This is to certify that the

thesis entitled

A Comparison of Hidden Markov Model Features
for the Recognition of Cursive Handwriting

presented by

Scott Connell

has been accepted towards fulfillment
of the requirements for

MS degree in Computer Science

 

 

va

Major professor

 

5/3/7C

 

0-7639

MS U i: an Affirmative Action/Equal Opportunity Institution

 

 

 

 

LIBRARY
Mlchlgan State
Unlverelty

 

 

 

PLACE II RETURN BOX to remove thle checkout from you record.
TO AVOID FINES return on or betore dete due.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU to An Nfirmetive Action/EM Opporumlty lnetltulon
Wane-m

A Comparison of Hidden Markov Model Features
for the Recognition of Cursive Handwriting

By

Scott Connell

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Computer Science Department

1996

ABSTRACT

A Comparison of Hidden Markov Model Features
for the Recognition of Cursive Handwriting

By

Scott Connell

Due to the difficulty of character segmentation in cursive handwriting recognition,
much recent research has turned to segmentation free approaches of word recognition.
While techniques of feature extraction for presegmented characters have been thor-
oughly explored in the literature, an evaluation of features for use with segmentation
during recognition techniques remains sparse. The main purpose of this thesis is to
provide a comparison of a number of feature extraction techniques applied to the
domain of legal amount recognition in bank checks. An experimental system using
Hidden Markov Models and a. horizontally sliding window is described. Results are
presented for the recognition of the entire legal field using a variety of features. Of
the experiments presented here, the best results were obtained by concatenating the
feature vectors from the present, previous, and next windows and using principal

component analysis to reduce the dimensionality of this resulting vector.