I
nte
rna
t
io
na
l
J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
p
ute
r
E
ng
in
ee
ring
(
I
J
E
CE
)
Vo
l.
10
,
No
.
1
,
Feb
r
u
ar
y
2020
,
p
p
.
778
~
785
I
SS
N:
2088
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v
10
i
1
.
p
p
7
7
8
-
785
778
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m/in
d
ex
.
p
h
p
/I
JE
C
E
Applica
tion
o
n
c
h
a
ra
cter
r
ecog
nitio
n
s
y
ste
m
on
r
o
a
d
s
ig
n
for
v
isua
lly
i
m
pa
i
red
:
c
a
se
s
tu
dy
a
ppr
o
a
ch and
f
utu
re
J
a
ej
o
o
n K
i
m
S
c
h
o
o
l
o
f
Co
m
p
u
te
r
a
n
d
Co
m
m
u
n
ic
a
ti
o
n
,
D
a
e
g
u
Un
i
v
e
r
sit
y
,
S
o
u
t
h
Ko
r
e
a
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
A
p
r
7
,
2
0
1
9
R
ev
i
s
ed
Sep
18
,
20
19
A
cc
ep
ted
Sep
27
,
2
0
19
M
a
n
y
v
isu
a
ll
y
im
p
a
ired
p
e
o
p
le
w
o
rld
w
id
e
a
re
u
n
a
b
le
to
trav
e
l
sa
fe
l
y
a
n
d
a
u
to
n
o
m
o
u
sly
b
e
c
a
u
se
th
e
y
a
re
p
h
y
sic
a
ll
y
u
n
a
b
le
to
p
e
rc
e
iv
e
e
ffe
c
ti
v
e
v
isu
a
l
in
f
o
rm
a
ti
o
n
d
u
rin
g
t
h
e
ir
d
a
il
y
li
v
e
s.
In
th
is
re
se
a
rc
h
,
w
e
stu
d
y
h
o
w
to
e
x
tr
a
c
t
th
e
c
h
a
ra
c
ter
in
f
o
rm
a
ti
o
n
o
f
th
e
ro
a
d
sig
n
a
n
d
tran
sm
it
it
t
o
t
h
e
v
isu
a
ll
y
im
p
a
ired
e
ff
e
c
ti
v
e
l
y
,
so
th
e
y
c
a
n
u
n
d
e
rsta
n
d
e
a
sie
r
.
Ex
p
e
rim
e
n
tal
m
e
th
o
d
is
to
a
p
p
l
y
th
e
M
a
x
i
m
a
ll
y
S
tab
le
Ex
tern
a
l
Re
g
io
n
a
n
d
S
tr
o
k
e
W
id
th
T
ra
n
s
f
o
r
m
m
e
th
o
d
in
P
h
a
se
I
so
th
a
t
th
e
v
isu
a
ll
y
i
m
p
a
ired
p
e
rso
n
c
a
n
re
c
o
g
n
ize
th
e
letters
o
n
th
e
ro
a
d
sig
n
s.
It
is
to
c
o
n
v
e
y
tex
t
in
f
o
rm
a
ti
o
n
t
o
th
e
d
isa
b
led
.
T
h
e
re
su
lt
o
f
P
h
a
se
I
u
sin
g
sa
m
p
le
s
o
f
s
i
m
p
le
ro
a
d
sig
n
s
w
a
s
to
e
x
trac
t
th
e
sig
n
in
f
o
rm
a
ti
o
n
a
f
ter
d
iv
id
in
g
th
e
e
x
a
c
t
c
h
a
ra
c
ter
a
re
a
,
b
u
t
th
e
a
c
c
u
ra
c
y
w
a
s
n
o
t
g
o
o
d
f
o
r
t
h
e
Ha
n
g
u
l
(Ko
re
a
n
c
h
a
ra
c
ters
)
in
f
o
r
m
a
ti
o
n
.
T
h
e
in
it
ial
e
x
p
e
ri
m
e
n
tal
re
su
lt
s
in
th
e
P
h
a
se
II
su
c
c
e
e
d
e
d
in
tra
n
sm
it
ti
n
g
th
e
tex
t
in
f
o
rm
a
ti
o
n
on
P
h
a
se
I
to
t
h
e
v
isu
a
ll
y
im
p
a
ired
.
In
t
h
e
f
u
tu
re
,
it
w
il
l
b
e
re
q
u
ired
t
o
d
e
v
e
lo
p
a
w
e
a
ra
b
le
c
h
a
ra
c
ter
re
c
o
g
n
it
io
n
sy
ste
m
t
h
a
t
c
a
n
b
e
a
tt
a
c
h
e
d
to
th
e
v
isu
a
ll
y
i
m
p
a
ired
.
In
o
rd
e
r
t
o
p
e
rf
o
rm
th
is
tas
k
,
w
e
n
e
e
d
to
d
e
v
e
lo
p
a
n
d
v
e
rify
a
m
in
iatu
rize
d
a
n
d
w
e
a
r
a
b
le ch
a
ra
c
ter r
e
c
o
g
n
it
io
n
sy
ste
m
.
In
th
is
p
a
p
e
r,
w
e
e
x
a
m
in
e
d
th
e
m
e
th
o
d
o
f
re
c
o
g
n
izin
g
ro
a
d
sig
n
c
h
a
ra
c
ters
o
n
th
e
ro
a
d
a
n
d
p
re
se
n
t
ed
a
p
o
ss
i
b
il
it
y
th
a
t
m
a
y
b
e
a
p
p
l
ica
b
le
to
o
u
r
f
in
a
l
d
e
v
e
lo
p
m
e
n
t.
K
ey
w
o
r
d
s
:
C
h
ar
ac
ter
r
ec
o
g
n
itio
n
s
y
s
te
m
MSE
R
OC
R
V
is
u
a
ll
y
i
m
p
air
ed
W
ea
r
ab
le
d
ev
ice
Co
p
y
rig
h
t
©
2
0
2
0
In
stit
u
te o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
rin
g
a
n
d
S
c
ien
c
e
.
Al
l
rig
h
ts
re
se
rv
e
d
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
J
ae
jo
o
n
Ki
m
Sc
h
o
o
l o
f
C
o
m
p
u
ter
a
n
d
C
o
m
m
u
n
i
c
at
io
n
,
Dae
g
u
U
n
i
v
er
s
i
t
y
,
2
0
1
Dae
g
u
d
ae
-
r
o
,
G
y
eo
n
g
s
a
n
,
G
y
eo
n
g
b
u
k
,
3
8
4
5
3
,
So
u
th
Ko
r
ea
.
E
m
ail:
j
j
k
i
m
i
s
u
@
d
ae
g
u
.
ac
.
k
r
1.
I
NT
RO
D
UCT
I
O
N
T
h
er
e
ar
e
m
a
n
y
p
eo
p
le
w
it
h
v
is
u
a
l
d
is
ab
ilit
ies
i
n
th
e
w
o
r
ld
as
w
el
l
as
p
eo
p
le
r
esid
e
in
Ko
r
ea
.
Au
x
iliar
y
d
e
v
ices
h
a
v
e
b
ee
n
ac
ti
v
el
y
d
e
v
elo
p
ed
to
p
r
o
v
id
e
v
is
u
all
y
i
m
p
air
ed
p
er
s
o
n
s
w
i
th
ac
c
u
r
ate
in
f
o
r
m
atio
n
o
n
r
o
ad
s
ig
n
s
d
u
r
in
g
w
a
lk
i
n
g
,
w
h
ic
h
ca
n
b
e
p
r
o
v
en
as
cr
u
c
ial
p
ar
ts
o
f
ev
er
y
d
ay
s
o
cial
ac
tiv
i
ties
.
E
v
en
m
o
r
e,
v
ar
io
u
s
d
e
v
ices
s
u
ch
a
s
p
r
o
d
u
cts
th
at
r
ea
d
o
r
p
r
o
v
id
e
d
ir
ec
tio
n
s
f
o
r
b
lin
d
h
av
e
b
ee
n
ac
tiv
e
l
y
d
ev
elo
p
ed
in
th
e
p
ast.
Var
io
u
s
m
e
th
o
d
s
ar
e
cu
r
r
en
t
l
y
in
d
e
v
elo
p
m
e
n
t
to
au
to
m
atica
ll
y
e
x
tr
ac
t
th
e
r
e
g
io
n
s
u
r
r
o
u
n
d
i
n
g
a
ch
ar
ac
ter
w
h
e
n
t
h
at
c
h
ar
ac
ter
is
d
etec
ted
to
b
e
in
t
h
e
p
r
esen
c
e
o
f
a
v
is
u
all
y
i
m
p
air
ed
p
er
s
o
n
[1
-
5]
.
T
h
e
m
et
h
o
d
co
n
s
is
t
s
o
f
m
a
in
l
y
f
ir
s
t
r
ec
o
g
n
izi
n
g
t
h
e
ch
ar
ac
ter
s
tr
in
g
in
its
r
eg
io
n
an
d
th
en
s
e
n
d
in
g
th
e
s
ca
n
n
ed
in
f
o
r
m
a
tio
n
b
ac
k
to
th
e
u
s
er
.
D
u
e
to
ar
ti
f
icial
in
tell
ig
e
n
ce
tec
h
n
o
lo
g
y
,
a
d
d
i
tio
n
al
f
ea
t
u
r
es
ca
n
in
cl
u
d
e
a
d
ictio
n
ar
y
f
u
n
ct
io
n
in
ca
s
e
th
e
u
s
er
d
o
es
n
o
t
u
n
d
er
s
tan
d
th
e
m
ea
n
in
g
o
f
t
h
e
s
ig
n
in
q
u
es
tio
n
.
I
n
r
ec
en
t
y
ea
r
s
,
v
ar
io
u
s
tec
h
n
iq
u
es
u
s
i
n
g
a
r
tif
ic
ial
i
n
tel
lig
e
n
ce
in
i
m
a
g
e
r
ec
o
g
n
i
tio
n
,
n
atu
r
al
la
n
g
u
a
g
e
p
r
o
ce
s
s
in
g
,
a
n
d
n
at
u
r
al
la
n
g
u
ag
e
g
e
n
er
atio
n
h
a
v
e
b
ee
n
d
ev
elo
p
ed
to
h
elp
p
eo
p
le
w
it
h
v
i
s
u
al
i
m
p
air
m
en
t
s
to
m
ai
n
tai
n
a
co
m
p
r
eh
en
s
i
v
e
an
d
p
r
o
d
u
c
tiv
e
lif
e
[
6
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
A
p
p
lica
tio
n
o
n
ch
a
r
a
cter reco
g
n
itio
n
s
ystem
o
n
r
o
a
d
s
ig
n
fo
r
visu
a
lly
imp
a
ir
ed
:
ca
s
e
...
(
Ja
ejo
o
n
K
im)
779
I
n
th
is
p
ap
er
,
w
e
tr
ied
to
d
ev
elo
p
a
v
ar
iet
y
o
f
w
ea
r
ab
le
as
s
is
ti
v
e
aid
s
f
o
r
t
h
e
v
is
u
all
y
i
m
p
a
ir
ed
to
eli
m
i
n
ate
th
e
i
n
co
n
v
en
ien
ce
o
f
w
a
lk
i
n
g
o
r
r
ea
l
lif
e.
I
n
o
r
d
er
to
d
o
th
is
,
in
P
h
ase
I
,
Ma
x
i
m
all
y
Stab
le
E
x
ter
n
al
R
eg
io
n
(
MSE
R
)
an
d
Stro
k
e
W
i
d
th
T
r
an
s
f
o
r
m
(
SW
T
)
f
ea
tu
r
es
w
er
e
ap
p
lied
t
o
ch
ar
ac
ter
r
ec
o
g
n
itio
n
o
f
r
o
ad
s
ig
n
s
a
n
d
p
ed
estrian
s
i
g
n
s
[
7
,
8
]
.
I
n
P
h
ase
I
I
,
w
e
tr
ied
to
d
esig
n
a
te
x
t
r
ec
o
g
n
itio
n
s
y
s
t
e
m
b
ased
o
n
T
T
S
(
T
ex
t
-
to
-
Sp
ee
ch
)
co
n
v
er
ter
[
9
]
th
at
tr
an
s
m
its
t
h
e
te
x
t r
ec
o
g
n
i
tio
n
in
f
o
r
m
at
io
n
to
th
e
v
is
u
all
y
i
m
p
air
ed
.
Th
is
p
ap
er
p
r
o
p
o
s
es
a
p
lan
f
o
r
th
e
v
i
s
u
al
l
y
i
m
p
air
ed
to
m
ee
t
t
h
ese
d
e
v
elo
p
m
e
n
ts
a
n
d
to
p
r
esen
t
th
e
ap
p
licab
le
p
o
s
s
ib
ilit
ie
s
.
S
ec
tio
n
2
b
r
ief
l
y
d
is
c
u
s
s
es
th
e
r
elate
d
w
o
r
k
s
f
o
r
ch
ar
ac
ter
s
tr
in
g
r
ec
o
g
n
itio
n
tech
n
o
lo
g
y
c
u
r
r
en
t
l
y
in
d
e
v
el
o
p
m
e
n
t
f
o
r
th
e
v
i
s
u
a
ll
y
i
m
p
a
ir
ed
an
d
ch
ar
ac
ter
is
tic
s
o
f
in
f
o
r
m
atio
n
g
u
id
i
n
g
s
y
s
te
m
.
Sectio
n
3
d
escr
ib
e
s
e
x
p
er
i
m
e
n
tal
i
m
p
le
m
e
n
tatio
n
o
f
ch
ar
ac
ter
r
ec
o
g
n
itio
n
f
o
r
o
u
r
p
r
o
p
o
s
ed
s
y
s
te
m
.
Sectio
n
s
4
an
d
5
co
n
clu
d
e
t
h
e
p
ap
er
b
y
d
is
cu
s
s
i
n
g
ex
p
er
i
m
e
n
tal
r
esu
l
ts
,
f
u
tu
r
e
w
o
r
k
s
an
d
o
th
er
p
o
s
s
ib
le
s
o
lu
tio
n
s
.
2.
RE
L
AT
E
D
WO
RK
2
.
1
.
Cha
ra
ct
er
is
t
ics a
nd
t
re
nd
o
f
info
r
m
a
t
io
n g
uid
i
ng
s
y
s
t
e
m
B
lin
d
p
eo
p
le
ar
e
d
iv
id
ed
in
to
lo
w
v
is
io
n
an
d
t
h
e
b
li
n
d
.
L
o
w
v
is
io
n
r
ef
er
s
to
v
i
s
io
n
p
r
o
b
le
m
s
t
h
at
ca
n
n
o
t
b
e
i
m
p
r
o
v
ed
b
y
m
ed
ic
al
o
r
o
p
tical
m
ea
n
s
d
u
e
to
b
ir
th
d
ef
ec
t
s
o
r
ac
q
u
ir
ed
e
y
e
d
is
e
ases
.
A
cc
o
r
d
in
g
to
a
r
ep
o
r
t
o
f
Yo
n
s
ei
Un
iv
er
s
it
y
Me
d
ical
C
en
ter
,
th
e
s
u
icid
e
r
ate
o
f
th
e
lo
w
v
is
io
n
is
t
w
ice
th
at
o
f
th
e
n
o
r
m
al
p
er
s
o
n
,
an
d
t
h
e
p
s
y
ch
o
lo
g
ical
s
tr
ess
d
u
e
to
th
e
lo
w
v
is
io
n
is
lar
g
e,
an
d
t
h
e
n
ee
d
f
o
r
an
au
x
i
liar
y
d
e
v
ice
to
he
lp
th
e
lo
w
v
is
io
n
is
in
cr
ea
s
i
n
g
[
1
0
]
.
I
n
ad
d
itio
n
,
a
r
ep
o
r
t
f
r
o
m
th
e
UN
es
t
i
m
a
tes
t
h
at
t
h
e
w
o
r
ld
‟
s
p
o
p
u
latio
n
ag
e
6
0
o
r
o
ld
er
w
ill
in
cr
ea
s
e
f
r
o
m
1
1
%
i
n
2
0
1
2
to
2
2
%
b
y
2
0
2
0
as
s
h
o
w
n
i
n
Fi
g
u
r
e
1
.
E
s
p
ec
iall
y
i
n
No
r
t
h
Am
er
ica
a
n
d
E
u
r
o
p
e,
th
e
d
is
tr
ib
u
tio
n
o
f
p
o
p
u
latio
n
a
g
ed
6
0
y
ea
r
s
o
r
o
ld
er
is
r
ep
o
r
ted
to
b
e
s
ig
n
i
f
ica
n
tl
y
h
ig
h
er
t
h
a
n
th
e
av
er
a
g
e
[
1
1
]
.
I
n
th
e
p
o
in
t
o
f
w
o
r
ld
m
ar
k
e
t,
Fig
u
r
e
2
in
d
icate
d
t
h
at
t
h
e
g
lo
b
al
eld
er
l
y
a
n
d
d
is
ab
led
ass
is
ti
v
e
d
ev
ices
m
ar
k
et
w
as
v
al
u
ed
at
US$
1
4
,
1
0
9
.
1
m
ill
io
n
i
n
2
0
1
5
an
d
is
ex
p
ec
ted
to
ex
p
an
d
at
a
C
A
GR
o
f
7
.
4
% d
u
r
in
g
t
h
e
f
o
r
ec
ast p
er
io
d
(
2
0
1
6
–
2
0
2
4)
[
1
2
]
.
Fig
u
r
e
1
.
T
h
e
s
tatis
tcs o
f
U
N,
g
en
d
er
ed
in
n
o
v
atio
n
s
(
2
0
1
2
v
s
.
2
0
5
0
)
[
1
1
]
Fig
u
r
e
2
.
Glo
b
al
eld
er
ly
an
d
d
is
ab
led
ass
is
t
iv
e
d
ev
ices
m
ar
k
et
s
ize
an
d
f
o
r
ec
ast
(
u
n
i
t: US$
Millio
n
)
[
1
2
]
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
10
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
0
:
778
-
785
780
2
.
2
.
Cha
ra
ct
er
info
r
m
a
t
io
n r
ec
o
g
nitio
n sy
s
t
e
m
T
o
allo
w
a
v
is
u
all
y
i
m
p
ar
ed
p
er
s
o
n
to
r
ec
o
g
n
ize
a
c
h
ar
ac
t
er
,
o
u
r
a
p
p
licatio
n
m
u
s
t
b
e
a
b
le
to
f
in
d
th
e
c
h
ar
ac
ter
r
eg
io
n
i
n
w
h
ic
h
th
e
s
i
g
n
o
r
th
e
m
e
s
s
a
g
e
i
s
w
r
itte
n
.
T
h
e
m
o
s
t
w
ell
-
k
n
o
wn
m
et
h
o
d
u
s
ed
f
o
r
b
in
ar
izatio
n
is
a
g
lo
b
al
b
i
n
ar
i
za
tio
n
w
h
ic
h
i
s
b
ased
o
n
a
s
i
n
g
le
t
h
r
esh
o
ld
[
1
3
-
1
5
]
.
Ho
w
e
v
er
,
th
e
s
e
m
et
h
o
d
s
h
av
e
ch
r
ac
ter
i
s
tics
s
u
c
h
as
p
e
r
f
o
r
m
an
ce
is
s
u
es
w
it
h
th
e
o
v
er
all
b
r
ig
h
tn
e
s
s
o
f
th
e
in
p
u
t
d
o
cu
m
e
n
t
i
m
ag
e
ar
e
n
o
t c
o
n
s
ta
n
t.
An
o
th
er
m
et
h
o
d
u
s
ed
f
o
r
b
in
a
r
izatio
n
is
th
e
Ma
x
i
m
al
l
y
Sta
b
le
E
x
ter
n
al
R
e
g
io
n
(
M
SER)
alg
o
r
ith
m
,
w
h
ic
h
is
u
s
ed
in
a
m
o
r
e
r
o
b
u
s
t
m
a
n
n
er
o
f
f
i
n
d
in
g
a
r
eg
io
n
w
it
h
a
s
m
all
r
ate
o
f
ch
an
g
e
w
h
ile
ch
a
n
g
in
g
th
e
th
r
es
h
o
ld
v
alu
e
i
n
o
r
d
er
to
d
is
ti
n
g
u
is
h
ch
ar
ac
ter
s
in
b
o
th
b
r
ig
h
t
an
d
d
ar
k
r
eg
io
n
s
.
T
h
e
ch
ar
ac
ter
ar
ea
is
g
en
er
all
y
ea
s
y
to
d
is
ti
n
g
u
is
h
f
r
o
m
th
e
b
ac
k
g
r
o
u
n
d
a
n
d
h
as
a
co
n
s
ta
n
t
b
r
i
g
h
t
n
es
s
v
al
u
e.
T
h
is
c
h
ar
ac
ter
is
tic
o
f
th
e
ch
ar
ac
ter
r
eg
io
n
is
s
u
itab
l
e
to
b
e
d
etec
ted
b
y
th
e
MSE
R
m
et
h
o
d
[
7
,
8
]
.
Va
r
io
u
s
ap
p
licatio
n
s
h
a
v
e
b
ee
n
d
ev
elo
p
ed
b
y
v
ar
io
u
s
co
m
p
a
n
ies,
all
o
f
w
h
ic
h
atte
m
p
t
to
ex
tr
ac
t
a
ch
ar
ac
ter
s
tr
in
g
b
y
ex
tr
ac
ti
n
g
h
o
r
izo
n
tal
an
d
v
er
tical
b
o
u
n
d
ar
y
co
m
p
o
n
en
ts
f
r
o
m
t
h
e
in
p
u
t
i
m
a
g
e.
T
h
ese
m
et
h
o
d
s
b
asicall
y
eith
er
f
i
n
d
th
e
p
o
r
tio
n
w
it
h
th
e
g
r
ea
test
b
o
u
n
d
ar
y
co
m
p
o
n
en
t
o
r
ex
tr
ac
t
a
c
h
ar
ac
ter
s
tr
in
g
t
h
at
is
ir
r
esp
ec
tiv
e
o
f
m
i
s
j
u
d
g
m
e
n
t
o
r
d
is
to
r
tio
n
.
T
h
e
A
B
B
YY
Fin
eRead
er
E
n
g
in
e,
O
C
R
SDK
(
ch
ar
ac
ter
r
ec
o
g
n
itio
n
d
ev
elo
p
m
e
n
t
to
o
lk
it
)
s
u
p
p
lied
b
y
R
etia,
w
as
ap
p
lied
to
th
e
d
ev
elo
p
m
en
t
o
f
a
'
p
r
i
n
t
-
to
-
s
p
ee
ch
co
n
v
er
ter
'
,
w
h
ich
i
s
a
d
ev
ice
f
o
r
r
ec
o
g
n
izi
n
g
p
r
in
ted
ch
ar
ac
ter
s
f
o
r
v
is
u
all
y
i
m
p
air
ed
an
d
co
n
v
er
tin
g
th
e
m
in
to
v
o
ice
[
16
-
18
]
.
A
B
B
YY
[
19
]
d
ev
elo
p
e
d
A
B
B
YY
F
in
e
R
ea
d
er
E
n
g
i
n
e
w
it
h
a
g
o
al
o
f
a
s
tab
le
„
p
r
in
t
-
to
-
s
p
ee
ch
‟
co
n
v
e
r
ter
f
o
r
th
e
v
is
u
all
y
i
m
p
air
ed
.
T
h
e
OC
R
k
e
y
tec
h
n
o
lo
g
ies o
f
A
B
B
YY
Fi
n
eRead
er
E
n
g
i
n
e
h
av
e
s
e
v
er
al
m
ai
n
f
ea
tu
r
e
s
s
u
c
h
as:
a.
Ko
r
ea
n
,
C
h
i
n
ese,
E
n
g
l
is
h
,
n
u
m
b
er
s
an
d
s
p
ec
ial
c
h
ar
ac
ter
s
s
u
p
p
o
r
t
b.
Ver
tical
r
ea
d
in
g
,
M
u
lti r
ea
d
in
g
c.
B
atch
s
ca
n
n
i
n
g
,
b
atc
h
co
n
v
er
s
io
n
p
o
s
s
ib
le
d.
I
t c
an
s
av
e
w
o
r
d
,
tex
t
f
ile
etc.
B
y
u
tili
zi
n
g
th
i
s
en
g
i
n
e,
th
e
v
is
u
all
y
i
m
p
air
ed
u
s
er
ca
n
co
n
v
er
t
th
e
d
esire
d
d
o
cu
m
e
n
t
to
s
p
ee
ch
th
r
o
u
g
h
t
h
i
s
d
ev
ice.
W
h
i
le
in
t
h
e
p
ast b
o
o
k
s
,
r
ec
eip
ts
a
n
d
o
th
er
r
ea
d
ab
le
d
o
cu
m
e
n
t
s
co
u
ld
n
o
t b
e
r
ea
d
w
it
h
o
u
t
B
r
aille
d
o
cu
m
en
t
s
,
t
h
e
en
g
i
n
e
n
o
w
m
a
k
e
s
it
p
o
s
s
ib
le
to
d
ir
ec
tl
y
s
ca
n
a
d
esire
d
d
o
cu
m
en
t
o
r
im
a
g
e
b
y
u
s
in
g
th
e
„
s
o
u
n
d
‟
w
it
h
O
C
R
tec
h
n
o
lo
g
y
a
n
d
co
n
v
er
t
it
s
o
th
e
u
s
e
r
ca
n
h
ea
r
th
e
b
y
v
o
ice.
T
h
e
„
f
i
n
g
er
r
ea
d
er
‟
[
20]
p
r
o
d
u
ct
d
ev
elo
p
ed
b
y
MI
T
w
o
r
k
s
b
y
p
laci
n
g
a
f
in
g
er
n
ea
r
t
h
e
b
o
o
k
.
A
h
ig
h
-
r
eso
lu
tio
n
ca
m
er
a
is
u
s
ed
to
s
ca
n
th
e
c
h
ar
ac
ter
s
co
r
r
esp
o
n
d
in
g
t
o
th
e
f
i
n
g
er
‟
s
to
u
c
h
w
h
ile
also
r
ea
d
in
g
th
e
m
o
u
t lo
u
d
a
f
ter
b
ein
g
s
ca
n
n
ed
.
A
f
te
r
s
ev
er
al
y
ea
r
s
o
f
r
esear
ch
,
O
r
C
a
m
,
a
n
I
s
r
ae
li
co
m
p
a
n
y
,
i
s
n
o
w
d
ev
e
lo
p
in
g
n
e
w
p
r
o
d
u
cts
th
at
u
s
e
f
ac
ial
r
ec
o
g
n
itio
n
tech
n
o
lo
g
y
to
s
ca
n
n
o
t
o
n
l
y
c
h
ar
ac
ter
s
b
u
t
al
s
o
v
ar
io
u
s
co
lo
r
s
[
21
]
.
T
h
er
e
ar
e
also
„
s
m
ar
t
g
las
s
es
‟
f
o
r
p
eo
p
le
w
ith
p
ar
tial
v
is
i
o
n
lo
s
s
.
T
h
ese
„
s
m
ar
t
g
la
s
s
es
‟
w
o
r
k
b
y
v
ie
w
in
g
i
m
a
g
es
i
n
3
D
g
las
s
es.
Ho
w
e
v
er
ev
en
th
o
u
g
h
th
e
s
e
g
las
s
es
h
a
v
e
b
ee
n
p
r
o
v
en
u
s
ef
u
l
r
ed
u
cin
g
in
co
n
v
e
n
ie
n
ce
o
f
th
e
v
i
s
u
all
y
d
is
ab
led
,
th
e
y
ar
e
s
till
v
er
y
e
x
p
en
s
iv
e.
T
h
er
ef
o
r
e
th
er
e
w
ill
b
e
d
if
f
ic
u
lties
in
m
ak
in
g
t
h
e
m
co
m
m
er
ciall
y
av
ai
lab
le
f
o
r
ev
er
y
o
n
e
.
Fig
u
r
e
3
s
h
o
w
s
th
e
n
u
m
b
er
o
f
p
eo
p
le
in
Ko
r
ea
w
it
h
s
ig
h
t
d
is
ab
ilit
ies
i
s
ar
o
u
n
d
ab
o
u
t
2
5
0
,
0
0
0
p
e
o
p
le
as
o
f
2
0
1
6
.
Un
f
o
r
tu
n
atel
y
,
t
h
e
n
u
m
b
er
o
f
p
e
o
p
le
w
it
h
v
is
u
al
d
is
a
b
ilit
ies
is
g
r
o
w
i
n
g
lar
g
er
ea
ch
y
ea
r
b
y
m
o
r
e
th
a
n
10
%.
I
n
ad
d
itio
n
,
th
e
t
y
p
es o
f
d
is
ab
ilit
ies
ar
e
d
is
cr
ib
ed
in
Fi
g
u
r
e
4
.
Fig
u
r
e
3
.
T
h
e
s
tatis
tics
o
f
d
i
s
a
b
ilit
ies i
n
Ko
r
ea
,
2
0
0
7
an
d
2
0
1
6
[
2
2
]
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
A
p
p
lica
tio
n
o
n
ch
a
r
a
cter reco
g
n
itio
n
s
ystem
o
n
r
o
a
d
s
ig
n
fo
r
visu
a
lly
imp
a
ir
ed
:
ca
s
e
...
(
Ja
ejo
o
n
K
im)
781
Fig
u
r
e
4
.
T
h
e
ty
p
es o
f
v
is
u
all
y
i
m
p
air
ed
.
(
a)
Scen
er
y
s
ee
n
w
i
th
n
o
r
m
al
e
y
e,
(
b
)
Scen
er
y
s
ee
n
w
it
h
e
y
e
s
w
it
h
ca
tar
ac
ts
,
(
c)
Scen
er
y
s
ee
n
w
it
h
e
y
e
s
w
it
h
g
la
u
co
m
a,
(
d
)
Scen
er
y
s
ee
n
w
it
h
e
y
e
s
w
it
h
d
iab
etic
r
etin
o
p
ath
y
,
(
e)
Scen
er
y
s
ee
n
w
it
h
e
y
e
s
w
it
h
m
ac
u
lar
d
eg
e
n
er
ati
on
[
2
3
]
3.
E
XP
E
R
I
M
E
NT
A
L
I
M
P
L
E
M
E
NT
AT
I
O
N
On
e
o
f
th
e
m
o
s
t
i
m
p
o
r
ta
n
t
p
ar
ts
o
f
t
h
e
ap
p
licatio
n
is
to
p
r
o
v
id
e
in
f
o
r
m
atio
n
th
a
t
is
o
n
l
y
n
e
ce
s
s
ar
y
i
n
o
r
d
er
to
n
ev
er
co
n
f
u
s
e
t
h
e
u
s
er
.
T
h
er
ef
o
r
e
,
th
e
ap
p
licatio
n
m
u
s
t
b
e
ab
le
to
s
elec
t
o
n
ly
t
h
e
r
elev
a
n
t
ar
ea
a
m
o
n
g
s
t
t
h
e
o
th
er
„
b
ac
k
g
r
o
u
n
d
n
o
is
e‟
.
Fo
r
ex
a
m
p
le,
th
e
ap
p
licatio
n
m
u
s
t
b
e
ab
le
t
o
d
is
tin
g
u
is
h
b
et
w
ee
n
a
r
o
ad
s
ig
n
an
d
th
e
o
t
h
er
n
at
u
r
al
b
ac
k
g
r
o
u
n
d
s
u
r
r
o
u
n
d
i
n
g
th
e
s
i
g
n
.
I
n
th
i
s
p
ap
er
,
w
e
ai
m
to
d
etec
t
th
e
te
x
t
ar
ea
in
th
e
i
m
a
g
e
p
r
o
v
id
ed
w
i
th
t
h
e
v
i
s
u
a
ll
y
i
m
p
air
ed
an
d
t
o
r
ec
o
g
n
ize
th
e
i
n
f
o
r
m
atio
n
p
r
o
v
id
ed
in
th
e
ar
ea
as
a
ch
ar
ac
ter
an
d
d
eliv
er
it
to
th
e
v
is
u
all
y
i
m
p
air
ed
.
I
n
o
r
d
er
to
im
p
le
m
e
n
t
o
u
r
g
o
al,
we
u
s
ed
MSE
R
a
n
d
OC
R
m
et
h
o
d
s
f
o
r
d
etec
tin
g
i
n
f
o
r
m
atio
n
ar
ea
an
d
tex
t
d
etec
tio
n
as
a
s
tar
t
i
n
g
p
o
in
t
o
f
r
esear
ch
.
B
ased
o
n
th
e
P
h
ase
I
o
f
th
e
r
esear
ch
,
w
e
ai
m
to
d
ev
elo
p
a
w
ea
r
ab
le
d
ev
ice
t
h
at
p
r
o
v
id
es
r
o
ad
in
f
o
r
m
atio
n
ab
o
u
t
th
e
v
i
s
u
all
y
i
m
p
air
ed
b
y
u
t
ilizi
n
g
t
h
e
T
T
S (
T
ex
t
-
to
-
Sp
ee
ch
)
f
u
n
ct
io
n
o
f
a
t
y
p
ical
s
m
ar
tp
h
o
n
e.
3
.
1
.
F
ea
t
ure
ex
t
ra
ct
io
n
w
it
h
M
SE
R
I
n
f
o
r
m
a
tio
n
o
n
th
e
r
o
ad
s
ig
n
s
d
o
n
o
t
co
n
tain
as
m
u
c
h
tex
t
u
al
in
f
o
r
m
a
tio
n
w
h
e
n
co
m
p
ar
ed
to
b
o
o
k
s
o
r
n
e
w
s
p
ap
er
s
.
A
s
a
m
et
h
o
d
of
th
e
v
is
u
all
y
i
m
p
air
ed
to
ar
r
iv
e
at
th
eir
d
esire
d
d
esti
n
atio
n
,
t
h
e
y
h
ad
to
p
r
ev
io
u
s
r
el
y
t
h
e
u
s
a
g
e
o
f
a
g
u
id
an
ce
s
tic
k
.
W
h
ile
it
m
a
y
to
p
r
ac
tical
a
n
d
h
a
s
b
ee
n
u
s
ed
as
f
r
o
m
th
e
b
eg
i
n
n
i
n
g
of
b
lin
d
p
eo
p
l
e,
a
d
if
f
er
en
t
m
et
h
o
d
ca
n
b
e
d
ev
elo
p
ed
.
T
h
ese
d
ay
s
it
is
p
o
s
s
ib
le
to
d
etec
t
th
e
te
x
t
i
n
a
n
i
m
ag
e
u
s
in
g
n
o
m
o
r
e
th
a
n
a
s
m
ar
tp
h
o
n
e
w
i
t
h
a
ca
m
er
a.
B
y
u
s
i
n
g
p
i
x
el
s
p
an
s
a
s
n
o
d
es,
b
lo
b
s
ca
n
b
e
g
e
n
er
ated
w
h
ic
h
ca
n
d
etec
t
tex
t
i
n
a
n
i
m
a
g
e
m
o
r
e
q
u
ick
l
y
a
n
d
ac
cu
r
atel
y
.
I
n
ad
d
i
tio
n
,
as
t
h
e
u
s
ag
e
o
f
s
m
ar
tp
h
o
n
e
i
n
cr
ea
s
e
d
v
er
y
r
a
p
id
ly
,
a
n
o
p
tical
ch
ar
ac
ter
r
ec
o
g
n
itio
n
ap
p
licatio
n
t
h
at
r
ec
o
g
n
ize
s
a
c
h
ar
ac
ter
of
an
i
m
a
g
e
ca
p
tu
r
ed
b
y
a
ca
m
er
a
m
o
u
n
ted
o
n
a
s
m
ar
t
p
h
o
n
e
an
d
d
is
p
la
y
s
it
o
n
a
d
is
p
la
y
c
an
b
e
d
is
tr
ib
u
ted
in
th
e
n
ea
r
f
u
t
u
r
e
an
d
p
r
o
v
e
m
o
r
e
p
r
ac
tical
co
m
p
ar
ed
to
a
g
u
id
an
ce
s
tic
k
.
T
h
e
m
o
s
t
i
m
p
o
r
tan
t
s
tep
i
s
to
d
etec
t
th
e
b
lo
b
w
h
ich
d
o
es
a
ca
n
d
id
ate
r
eg
io
n
ex
is
t
s
i
n
t
h
e
i
m
a
g
e
t
h
at
is
m
o
s
t
li
k
el
y
to
co
n
ta
in
t
h
e
t
ex
t
o
f
t
h
e
r
o
ad
s
ig
n
.
T
h
e
S
WT
alg
o
r
ith
m
[
2
4
,
2
5
]
is
a
m
et
h
o
d
o
f
d
etec
tin
g
tex
t
b
y
d
eter
m
in
in
g
w
h
ich
r
e
g
io
n
h
as
lit
tle
ch
a
n
g
e
in
t
h
e
s
h
ap
e
an
d
th
ic
k
n
e
s
s
.
W
h
e
n
t
h
e
p
ar
a
m
eter
s
ar
e
g
i
v
e
n
,
th
e
al
g
o
r
ith
m
d
etec
t
s
t
h
e
tex
t
w
it
h
i
n
t
h
e
i
m
a
g
e
u
s
i
n
g
th
o
s
e
p
ar
am
eter
s
.
T
h
e
m
ai
n
ad
v
a
n
ta
g
e
o
f
SW
T
ar
e
th
at
it
ca
n
d
etec
t
tex
t
in
a
n
i
m
a
g
e
w
it
h
o
u
t
a
s
ep
ar
ate
lear
n
in
g
p
r
o
ce
s
s
,
b
u
t
it
also
h
as
a
d
r
a
w
b
a
ck
in
w
h
ich
it
ta
k
es
a
lo
t
o
f
t
i
m
e
to
d
etec
t
te
x
t
in
a
n
i
m
a
g
e
b
ec
au
s
e
o
f
its
co
m
p
li
ca
ted
o
p
er
atio
n
.
T
h
e
MSE
R
al
g
o
r
ith
m
f
o
r
r
o
b
u
s
t
tex
t
d
etec
tio
n
is
w
id
el
y
u
s
ed
as
an
alg
o
r
ith
m
f
o
r
d
etec
tin
g
b
lo
b
as
an
ag
g
r
eg
ate
r
eg
io
n
o
f
p
ix
els
d
if
f
er
en
t
f
r
o
m
s
u
r
r
o
u
n
d
in
g
p
ix
el
s
i
n
in
t
en
s
it
y
.
T
h
e
MSE
R
al
g
o
r
ith
m
h
as
t
h
e
ad
v
a
n
ta
g
e
th
a
t
it
ca
n
d
etec
t
th
e
b
lo
b
f
aste
r
th
an
t
h
e
SW
T
alg
o
r
ith
m
,
b
u
t
it
h
as
a
d
is
ad
v
an
ta
g
e
th
at
t
h
e
ac
cu
r
ac
y
is
s
o
m
e
w
h
at
lac
k
i
n
g
,
s
u
ch
as
d
etec
tin
g
th
e
tic
k
le
o
r
n
o
is
e
b
y
b
lo
b
.
3
.
2
.
Appl
ica
t
io
n pr
oc
edure
Fig
u
r
e
5
is
an
o
v
er
all
b
lo
ck
d
iag
r
a
m
th
at
s
h
o
w
s
t
h
e
p
r
o
ce
s
s
o
f
o
u
r
ap
p
licatio
n
af
ter
ca
p
tu
r
in
g
th
e
i
m
a
g
e
an
d
ex
tr
ac
tin
g
th
e
t
ex
t.
I
n
th
is
p
ap
er
,
w
e
ai
m
ed
to
test
th
e
d
esig
n
p
o
s
s
ib
ilit
y
o
f
an
ass
is
ti
v
e
d
ev
ice
w
h
ic
h
ca
n
b
e
h
elp
f
u
l
f
o
r
th
e
v
is
u
all
y
i
m
p
a
ir
ed
th
r
o
u
g
h
P
h
a
s
e
I
p
r
o
ce
s
s
in
t
w
o
s
tag
e
s
.
T
h
e
f
ir
s
t
s
tep
ca
p
tu
r
es
i
m
a
g
es
of
a
s
m
ar
tp
h
o
n
e
o
r
f
u
t
u
r
e
ter
m
i
n
al
d
ev
ice
th
a
t
th
e
b
l
in
d
h
a
s
an
d
t
h
en
r
e
m
o
v
e
s
n
o
n
-
te
x
t
r
eg
io
n
s
u
s
i
n
g
MSE
R
.
T
h
e
MSE
R
ex
tr
ac
tio
n
i
m
p
le
m
en
t
s
t
h
e
f
o
llo
w
i
n
g
s
tep
s
:
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
10
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
0
:
778
-
785
782
F
ig
u
r
e
5
.
T
h
e
o
v
er
all
b
lo
ck
d
i
ag
r
a
m
a.
S
w
ee
p
th
r
es
h
o
ld
o
f
in
te
n
s
it
y
b.
E
x
tr
ac
t c
o
n
n
ec
ted
co
m
p
o
n
en
t
s
c.
Fin
d
a
th
r
e
s
h
o
ld
o
n
m
a
x
i
m
all
y
s
tab
le
d.
A
p
p
r
o
x
i
m
ate
a
r
eg
io
n
w
i
th
a
n
ellip
s
e
e.
Kee
p
r
eg
io
n
d
escr
ip
to
r
s
On
e
o
f
t
h
e
m
o
s
t
i
m
p
o
r
tan
t
f
e
atu
r
es
o
f
t
h
e
M
SER
a
lg
o
r
it
h
m
i
s
t
h
at
it
d
etec
ts
m
o
s
t
te
x
t
in
a
n
i
m
a
g
e
b
u
t
also
ex
tr
ac
t
s
a
s
tab
le
r
eg
i
o
n
,
n
o
t
a
tex
t
r
eg
io
n
.
T
o
m
ak
e
th
e
al
g
o
r
ith
m
s
u
it
o
u
r
n
ee
d
s
,
w
e
r
e
m
o
v
ed
n
o
n
-
tex
t
ar
ea
s
u
s
i
n
g
p
ar
a
m
eter
s
th
at
d
is
ti
n
g
u
is
h
w
o
r
d
s
b
y
th
eir
g
eo
m
etr
ic
al
p
r
o
p
er
ties
w
it
h
i
n
t
h
e
i
m
ag
e.
I
n
ad
d
itio
n
,
SW
T
ca
n
b
e
u
s
ed
to
d
is
tin
g
u
is
h
m
o
r
e
s
o
p
h
is
tic
ated
n
o
n
-
te
x
t
ar
ea
s
.
I
n
th
i
s
p
a
p
er
,
w
e
ap
p
l
ied
th
i
s
m
et
h
o
d
to
f
in
d
an
d
r
em
o
v
e
n
o
n
-
te
x
t
ar
ea
s
.
I
n
o
r
d
er
to
m
er
g
e
in
d
i
v
id
u
al
te
x
t
ar
ea
s
to
ex
tr
ac
t
in
f
o
r
m
at
io
n
o
n
a
s
in
g
le
w
o
r
d
o
r
a
tex
t
lin
e
af
ter
f
in
all
y
co
n
f
i
r
m
in
g
t
h
e
tex
t
ar
ea
,
a
n
eig
h
b
o
r
in
g
te
x
t
ar
ea
is
s
ea
r
ch
ed
to
f
o
r
m
a
b
o
u
n
d
in
g
b
o
x
o
f
t
h
ese
ar
ea
s
,
an
d
f
i
n
all
y
,
th
e
i
n
f
o
r
m
atio
n
is
ex
tr
ac
ted
.
I
n
P
h
ase
I
I
,
th
e
f
in
a
l
d
etec
ted
tex
t
in
f
o
r
m
atio
n
is
co
n
v
er
ted
i
n
to
v
o
ice
in
f
o
r
m
a
tio
n
on
t
h
e
v
is
u
a
ll
y
i
m
p
air
ed
to
b
e
t
r
an
s
m
it
ted
o
r
ap
p
lied
to
f
u
tu
r
e
ap
p
licab
le
an
d
w
ea
r
ab
le
d
ev
ices.
4.
RE
SU
L
T
AND
DI
SCUS
SI
O
N
4
.
1
.
P
ha
s
e
I
I
n
o
u
r
ex
p
er
i
m
e
n
t,
w
e
u
s
ed
a
t
y
p
ical
i
m
ag
e
o
f
th
e
r
o
ad
s
ig
n
w
h
ic
h
th
e
v
is
u
all
y
i
m
p
a
ir
ed
ca
n
ea
s
il
y
en
co
u
n
ter
o
n
th
e
r
o
ad
.
T
h
e
lef
t
co
lu
m
n
o
f
Fig
u
r
e
6
s
h
o
w
s
te
s
t
d
ata
s
et.
T
h
ese
i
m
a
g
es
w
er
e
s
o
r
ted
in
o
r
d
er
to
m
ak
e
te
s
t
co
n
d
itio
n
s
m
o
r
e
d
if
f
icu
l
t.
W
h
en
co
m
p
ar
in
g
th
e
a
m
o
u
n
t
o
f
tex
t
to
th
e
w
h
o
le
i
m
a
g
e
th
er
e
is
n
o
t
m
u
c
h
te
x
t
co
n
tain
ed
i
n
t
h
e
i
m
ag
e
in
th
e
f
ir
s
t
p
lace
.
T
h
er
ef
o
r
e,
th
ese
i
m
a
g
es
w
er
e
id
ea
l
in
te
s
ti
n
g
i
f
o
u
r
p
r
o
p
o
s
ed
m
eth
o
d
o
f
co
m
b
i
n
in
g
MSE
R
a
n
d
SW
T
to
ex
tr
ac
t
th
e
n
ec
es
s
ar
y
i
n
f
o
r
m
atio
n
.
T
h
ese
te
s
t
co
n
d
itio
n
s
w
er
e
also
i
m
p
o
r
tan
t
b
ec
au
s
e
t
h
e
en
d
g
o
al
o
f
o
u
r
ap
p
licatio
n
is
to
d
is
tr
ib
u
te
it
s
o
th
at
it
ca
n
b
e
u
s
ed
d
u
r
in
g
ev
er
y
d
a
y
ac
ti
v
itie
s
.
T
h
u
s
,
w
e
m
ad
e
s
u
r
e
th
at
t
h
e
i
m
a
g
es
w
er
e
s
elec
ted
v
er
y
ca
r
e
f
u
l
l
y
in
o
r
d
er
to
s
i
m
u
late
w
h
at
t
h
e
u
s
er
m
i
g
h
t e
n
co
u
n
ter
w
h
ile
u
s
i
n
g
t
h
e
ap
p
licatio
n
.
I
n
Fig
u
r
e
6
,
w
e
s
h
o
w
t
h
e
r
es
u
lts
o
f
ap
p
ly
in
g
MSE
R
,
g
eo
m
etr
ic
ch
ar
ac
ter
is
tics
,
an
d
th
en
ap
p
ly
in
g
SW
T
to
f
in
d
n
o
n
-
te
x
t
ar
ea
s
i
n
ca
p
tu
r
ed
i
m
ag
e
s
u
s
i
n
g
m
o
b
ile
p
h
o
n
es
o
f
t
h
e
v
is
u
all
y
i
m
p
air
ed
.
T
h
ese
r
esu
lts
ar
e
s
h
o
w
n
o
n
r
ig
h
t
co
lu
m
n
o
f
Fig
u
r
e
6
an
d
T
ab
le
1
s
h
o
w
s
th
e
f
i
n
al
d
etec
ted
ch
ar
ac
ter
in
f
o
r
m
atio
n
o
f
ea
ch
test
i
m
ag
e.
I
n
o
u
r
test
ca
s
e
o
f
i
m
a
g
es
co
n
tai
n
i
n
g
o
n
l
y
E
n
g
li
s
h
c
h
ar
ac
ter
s
,
ac
c
u
r
ac
y
w
as
h
i
g
h
in
th
e
i
m
p
le
m
e
n
ted
m
et
h
o
d
.
Ho
w
e
v
er
,
w
h
en
w
e
ap
p
lied
o
u
r
m
et
h
o
d
to
an
im
a
g
e
w
h
er
e
K
o
r
ea
n
an
d
E
n
g
lis
h
w
er
e
m
ix
ed
,
w
e
co
u
ld
s
ee
a
d
r
o
p
o
f
f
in
p
er
f
o
r
m
a
n
ce
.
T
h
r
o
u
g
h
th
i
s
r
esu
l
t
w
e
ca
n
co
n
f
i
r
m
th
at
a
d
if
f
er
en
t
d
etec
tio
n
en
g
in
e
i
s
r
eq
u
ir
ed
f
o
r
th
e
d
etec
tio
n
o
f
m
u
l
tip
le
lan
g
u
a
g
es.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
A
p
p
lica
tio
n
o
n
ch
a
r
a
cter reco
g
n
itio
n
s
ystem
o
n
r
o
a
d
s
ig
n
fo
r
visu
a
lly
imp
a
ir
ed
:
ca
s
e
...
(
Ja
ejo
o
n
K
im)
783
(
a)
(
b
)
(
c)
(
d
)
(
e)
(
f
)
(
g
)
(
h
)
Fig
u
r
e
6
.
T
h
e
test
d
ata
s
et
an
d
ex
p
er
i
m
e
n
tal
r
es
u
lts
T
ab
le
1
.
I
m
p
le
m
e
n
tal
a
n
d
d
etec
ted
r
esu
lts
T
e
st
d
a
t
a
D
e
t
e
c
t
e
d
T
e
x
t
I
n
f
o
r
mat
i
o
n
F
i
g
u
r
e
6
-
(
a
)
C
y
c
l
i
st
s
A
H
E
A
D
S
L
O
W
F
i
g
u
r
e
6
-
(
c
)
A
H
E
A
D
F
i
g
u
r
e
6
-
(
e
)
P
ED
X
I
N
G
F
i
g
u
r
e
6
-
(
g
)
‟
.
„
i
"
.
S
i
6
S
L
O
W
4
.
2
.
P
ha
s
e
I
I
T
T
S
is
a
t
y
p
e
o
f
s
p
ee
c
h
s
y
n
th
esis
p
r
o
g
r
a
m
t
h
at
m
a
k
es
t
h
e
c
o
n
ten
t
s
o
f
co
m
p
u
ter
d
o
cu
m
e
n
ts
,
s
u
ch
a
s
h
elp
f
iles
a
n
d
w
eb
p
ag
es,
s
o
u
n
d
h
u
m
a
n
-
r
ea
d
ab
le.
T
T
S
ca
n
al
s
o
r
ea
d
im
a
g
e
in
f
o
r
m
at
io
n
f
o
r
p
eo
p
le
w
it
h
v
is
u
a
l
i
m
p
air
m
e
n
t
s
.
T
h
er
e
ar
e
m
a
n
y
T
T
S
p
r
o
d
u
cts
o
n
th
e
m
ar
k
et
,
in
clu
d
i
n
g
R
ea
d
P
lease
2
0
0
0
,
P
r
o
v
er
b
e
Sp
ee
ch
Un
it,
an
d
T
ex
tA
lo
u
d
.
L
u
ce
n
t
an
d
A
T
&
T
h
av
e
th
eir
o
w
n
p
r
o
d
u
cts
ca
lled
“
T
ex
t
-
to
-
Sp
ee
ch
.
”
I
n
th
is
r
esear
ch
,
w
e
ap
p
lie
d
Mic
r
o
s
o
f
t
Sp
ee
ch
A
P
I
'
s
s
p
ee
ch
to
te
x
t
f
u
n
c
tio
n
alit
y
i
n
co
n
s
id
er
atio
n
o
f
e
f
f
i
cien
t
a
s
p
ec
ts
[
2
6
]
.
Fig
u
r
e
7
s
h
o
w
s
s
i
m
p
l
if
ied
b
lo
ck
d
iag
r
a
m
f
o
r
T
T
S f
u
n
c
tio
n
.
T
h
is
s
y
s
te
m
w
o
r
k
s
w
ell
f
o
r
o
u
r
test
d
ata.
(
a)
(
b
)
(
c)
Fig
u
r
e
7
.
T
h
e
b
lo
ck
d
iag
r
am
f
o
r
tex
t
-
to
-
s
p
ee
ch
p
r
o
ce
s
s
.
(
a)
in
p
u
t
i
m
a
g
e,
(
b
)
ca
p
tu
r
ed
letter
s
o
n
n
o
tep
ad
,
an
d
(
c)
s
p
ee
ch
o
u
t
w
i
th
te
x
t.t
x
t f
ile
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
10
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
0
:
778
-
785
784
5.
CO
NCLU
SI
O
N
I
n
th
is
p
ap
er
,
w
e
p
r
o
p
o
s
e
d
a
ch
ar
ac
ter
r
ec
o
g
n
itio
n
s
y
s
te
m
f
o
r
v
is
u
a
l
i
m
p
air
ed
p
er
s
o
n
s
an
d
ad
d
e
d
in
a
v
o
ice
g
u
id
a
n
ce
ap
p
licatio
n
im
p
le
m
e
n
tat
io
n
p
r
o
ce
s
s
.
Ou
r
g
o
al
is
to
ad
d
a
ch
ar
ac
ter
r
ec
o
g
n
itio
n
s
y
s
te
m
f
o
r
th
e
v
is
u
all
y
i
m
p
air
ed
to
th
e
s
m
ar
tp
h
o
n
e
ter
m
i
n
al
t
h
r
o
u
g
h
H/W
p
r
o
d
u
ctio
n
.
W
e
p
r
o
p
o
s
ed
a
m
et
h
o
d
to
d
etec
t
th
e
ch
ar
ac
ter
s
an
d
ap
p
l
y
t
h
e
v
o
ice
g
u
id
a
n
ce
f
r
o
m
th
e
i
m
a
g
e
s
ca
p
tu
r
ed
in
P
h
a
s
e
I
.
Usi
n
g
th
e
f
ea
t
u
r
es o
f
M
SER
an
d
SW
T
,
it
s
h
o
w
ed
th
e
p
o
s
s
i
b
ilit
y
o
f
ex
tr
ac
tin
g
th
e
letter
s
o
f
r
o
ad
s
ig
n
s
,
an
d
in
P
h
a
s
e
I
I
,
it
s
h
o
w
ed
th
e
r
es
u
l
t
o
f
d
eliv
er
in
g
t
h
e
ex
tr
ac
ted
te
x
t
in
f
o
r
m
atio
n
to
t
h
e
u
s
er
s
.
Ho
w
e
v
er
,
in
th
is
p
ap
er
,
w
e
ca
n
lo
o
k
at
v
ar
io
u
s
f
u
t
u
r
e
p
o
in
t
s
to
co
n
s
id
er
.
So
f
ar
,
th
e
r
o
ad
g
u
id
a
n
ce
s
er
v
ice
ap
p
licatio
n
f
o
r
th
e
v
i
s
u
all
y
i
m
p
air
ed
h
as
b
ee
n
d
ev
el
o
p
ed
u
s
in
g
a
b
lu
e
p
itch
er
co
m
m
u
n
icatio
n
b
ased
H/W
au
x
iliar
y
d
ev
ice
s
u
c
h
as
a
b
ea
co
n
i
n
a
s
p
ec
if
ic
a
r
ea
.
W
ith
th
e
d
ev
elo
p
m
e
n
t
o
f
I
o
T
tech
n
o
lo
g
y
,
th
e
o
b
j
ec
ts
an
d
d
ata
r
elate
d
t
o
I
o
T
h
av
e
b
ee
n
in
cr
ea
s
ed
g
r
ea
tl
y
.
I
t
is
n
ec
ess
ar
y
to
d
ev
el
o
p
a
ca
m
er
a
-
b
ased
v
id
eo
in
f
o
r
m
at
io
n
g
u
id
an
ce
s
er
v
ice
ap
p
t
h
at
u
tili
ze
s
a
v
o
ice
g
u
id
an
ce
s
er
v
ice
tech
n
o
lo
g
y
th
at
p
r
o
v
id
es
ap
p
r
o
p
r
iate
in
f
o
r
m
a
tio
n
.
I
n
a
d
d
itio
n
,
a
r
tif
icial
in
telli
g
en
ce
tech
n
iq
u
e
s
s
u
ch
as
i
m
a
g
e
r
ec
o
g
n
itio
n
,
n
a
tu
r
al
lan
g
u
a
g
e
p
r
o
ce
s
s
in
g
,
a
n
d
n
at
u
r
al
lan
g
u
ag
e
g
e
n
er
atio
n
ca
n
b
e
u
s
ed
to
en
ab
le
b
lin
d
p
eo
p
le
to
liv
e
a
m
o
r
e
co
m
p
r
e
h
en
s
iv
e
a
n
d
p
r
o
d
u
ctiv
e
lif
e.
ACK
NO
WL
E
D
G
E
M
E
NT
S
T
h
is
r
esear
ch
w
as s
u
p
p
o
r
ted
b
y
th
e
Dae
g
u
Un
iv
er
s
it
y
R
esear
ch
Gr
an
t,
2017.
RE
F
E
R
E
NC
E
S
[1
]
A
.
A
.
P
a
n
c
h
a
l,
e
t
a
l
.
,
“
Ch
a
ra
c
ter
d
e
tec
ti
o
n
a
n
d
re
c
o
g
n
i
ti
o
n
sy
ste
m
f
o
r
v
isu
a
ll
y
i
m
p
a
ired
p
e
o
p
le,”
Pro
c
e
e
d
in
g
s
o
f
IEE
E
In
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
n
Rec
e
n
t
T
re
n
d
s
in
El
e
c
tro
n
ics
,
I
n
fo
rm
a
t
io
n
&
Co
mm
u
n
ica
t
io
n
T
e
c
h
n
o
l
o
g
y
,
In
d
ia
,
p
p
.
1
4
9
2
-
1
4
9
6
,
2
0
1
6
.
[2
]
A
b
d
u
rra
s
y
id
,
e
t
a
l.
,
“
De
tec
ti
o
n
o
f
i
m
m
o
v
a
b
le
o
b
jec
ts
o
n
v
isu
a
ll
y
im
p
a
ired
p
e
o
p
le
w
a
lk
in
g
a
id
s,”
T
EL
KOM
NIKA
T
e
lec
o
mm
u
n
ica
ti
o
n
Co
mp
u
ti
n
g
E
lec
tro
n
ics
a
n
d
C
o
n
tr
o
l
,
v
o
l
.
1
7
,
p
p
.
5
80
-
5
8
5
,
2
0
1
9
.
[3
]
D
.
A
.
M
o
h
a
m
m
e
d
,
e
t
a
l.
,
“
O
f
f
-
li
n
e
h
a
n
d
w
rit
ten
c
h
a
ra
c
ter
re
c
o
g
n
it
io
n
u
si
n
g
a
n
in
teg
ra
ted
DBSCA
N
-
A
N
N
sc
h
e
m
e
,
”
In
d
o
n
e
sia
n
J
o
u
rn
a
l
o
f
El
e
c
trica
l
En
g
i
n
e
e
rin
g
a
n
d
C
o
mp
u
ter
S
c
ien
c
e
,
v
o
l.
1
4
,
p
p
.
1
4
4
3
-
1
4
5
1
,
2
0
1
9
.
[4
]
O
.
O.
Ola
d
a
y
o
,
“
Yo
ru
b
a
L
a
n
g
u
a
g
e
a
n
d
Nu
m
e
ra
ls‟
O
ff
li
n
e
In
terp
re
ter
Us
in
g
M
o
rp
h
o
l
o
g
ica
l
a
n
d
T
e
m
p
lat
e
M
a
tch
in
g
,
”
IAE
S
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
Arti
fi
c
i
a
l
I
n
telli
g
e
n
c
e
,
v
o
l.
3
,
p
p
.
6
4
-
7
2
,
2
0
1
4
.
[5
]
I
.
F
.
Bt
.
Ha
iru
m
a
n
a
n
d
O
.
M
.
F
o
o
n
g
,
“
OCR
sig
n
a
g
e
re
c
o
g
n
it
io
n
w
it
h
sk
e
w
&
sla
n
t
c
o
rre
c
ti
o
n
f
o
r
v
isu
a
ll
y
i
m
p
a
ired
p
e
o
p
le
,
”
Pr
o
c
e
e
d
in
g
s
o
f
IEE
E
In
t
e
rn
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Hy
b
rid
In
telli
g
e
n
t
S
y
ste
ms
,
M
e
lac
c
a
,
M
a
la
y
sia
,
p
p
.
3
0
6
-
3
1
0
,
2
0
1
1
.
[6
]
M
.
E.
P
o
ll
a
c
k
,
“
In
telli
g
e
n
t
T
e
c
h
n
o
l
o
g
y
f
o
r
a
n
Ag
in
g
P
o
p
u
lati
o
n
:
T
h
e
Us
e
o
f
AI
to
A
ss
ist
El
d
e
rs
w
it
h
Co
g
n
it
iv
e
Im
p
a
ir
m
e
n
t,
”
AI
M
a
g
a
zin
e
,
Ame
ric
a
n
Asso
c
i
a
ti
o
n
fo
r A
rti
fi
c
ia
l
In
t
e
ll
ig
e
n
c
e
,
v
o
l.
2
6
,
p
p
.
9
-
2
1
,
2
0
0
5
.
[7
]
J.
M
a
tas
,
e
t
a
l.
,
“
Ro
b
u
st
W
id
e
Ba
se
li
n
e
S
tere
o
f
ro
m
M
a
x
ima
ll
y
S
tab
le
Ex
trem
a
l
Re
g
io
n
s,”
Pro
c
e
e
d
in
g
s
o
f
th
e
Brit
ish
M
a
c
h
i
n
e
Vi
si
o
n
C
o
n
fer
e
n
c
e
,
BM
VC
,
p
p
.
3
8
4
-
3
9
3
,
2
0
0
2
.
[8
]
F
o
rss
e
n
a
n
d
P
er
-
E
rik
,
“
M
a
x
im
a
ll
y
S
tab
le
Co
lo
u
r
Re
g
io
n
s
f
o
r
Re
c
o
g
n
it
io
n
a
n
d
M
a
tch
in
g
,
”
Pro
c
e
e
d
in
g
s
o
f
IEE
E
Co
n
fer
e
n
c
e
o
n
Co
m
p
u
ter
Vi
si
o
n
a
n
d
Pa
tt
e
rn
Rec
o
g
n
it
io
n
,
CVP
R,
M
in
n
e
a
p
o
li
s,
US
A
,
2
0
0
7
.
[9
]
Q
.
Ye
a
n
d
D
.
Do
e
rm
a
n
n
,
“
T
e
x
t
De
tec
ti
o
n
a
n
d
Re
c
o
g
n
it
i
o
n
in
Im
a
g
e
r
y
:
A
S
u
rv
e
y
,
”
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
P
a
tt
e
rn
An
a
lys
is
a
n
d
M
a
c
h
i
n
e
In
tell
ig
e
n
c
e
,
v
o
l.
3
7
,
p
p
.
1
4
8
0
-
1
5
0
0
,
2
0
1
5
.
[1
0
]
T
.
H
.
Ri
m
,
e
t
a
l.
,
“
In
f
lu
e
n
c
e
o
f
v
i
su
a
l
a
c
u
it
y
o
n
su
icid
a
l
id
e
a
ti
o
n
,
s
u
icid
e
a
tt
e
m
p
ts
a
n
d
d
e
p
re
ss
io
n
i
n
S
o
u
t
h
Ko
re
a
,
”
Brit
ish
J
o
u
rn
a
l
o
f
Op
h
th
a
lmo
l
o
g
y
,
v
o
l.
9
9
,
p
p
.
1
1
1
2
-
1
1
1
9
,
2
0
1
5
.
[1
1
]
Ex
p
lo
ri
n
g
M
a
rk
e
ts
f
o
r
A
ss
isti
v
e
Tec
h
n
o
lo
g
ies
f
o
r
th
e
El
d
e
rl
y
,
UN
Ge
n
d
e
re
d
In
n
o
v
a
ti
o
n
s
.
A
v
a
il
a
b
le:
h
tt
p
s:/
/g
e
n
d
e
re
d
in
n
o
v
a
ti
o
n
s.sta
n
f
o
rd
.
e
d
u
/ca
se
-
stu
d
ies
/ro
b
o
ts.
h
tm
l
.
[1
2
]
G
lo
b
a
l
El
d
e
rly
a
n
d
Disa
b
led
A
ss
isti
v
e
De
v
i
c
e
s M
a
rk
e
t,
“
Co
h
e
re
n
t
M
a
rk
e
t
In
sig
h
ts,
”
2
0
1
7
.
[1
3
]
N.
Otsu
,
“
A
T
h
re
sh
o
ld
S
e
lec
ti
o
n
M
e
th
o
d
f
ro
m
G
ra
y
Lev
e
l
Histo
g
ra
m
,
”
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
S
y
ste
m
,
v
o
l.
1
9
,
p
p
.
62
-
66
,
1
9
7
9
.
[1
4
]
G
.
P
a
rk
,
e
t
a
l.
,
“
A
S
tu
d
y
o
n
En
h
a
n
c
e
d
Bi
n
a
riza
ti
o
n
M
e
t
h
o
d
b
y
Us
in
g
In
ten
sity
In
f
o
r
m
a
ti
o
n
,
”
Pro
c
e
e
d
in
g
s
o
f
th
e
S
p
rin
g
Co
n
fer
e
n
c
e
o
f
t
h
e
Ko
r
e
a
M
u
lt
ime
d
i
a
S
o
c
iety
,
p
p
.
4
4
1
-
4
4
5
,
2
0
0
3
.
[1
5
]
K.
M
u
n
a
d
i,
e
t
a
l
.
,
“
Im
p
ro
v
e
d
T
h
re
sh
o
ld
i
n
g
M
e
th
o
d
f
o
r
E
n
h
a
n
c
i
n
g
Ja
w
i
Bin
a
riza
ti
o
n
P
e
rf
o
rm
a
n
c
e
,
”
Pro
c
e
e
d
in
g
s
o
f
1
4
t
h
IAP
R
I
n
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
n
Do
c
u
me
n
t
An
a
lys
is
a
n
d
R
e
c
o
g
n
it
i
o
n
,
Ky
o
t
o
,
J
a
p
a
n
,
v
o
l.
1
,
p
p
.
1
1
0
8
-
1
1
1
3
,
2
0
1
7
.
[1
6
]
R.
S
m
it
h
,
“
A
n
O
v
e
r
v
ie
w
o
f
th
e
Tes
se
ra
c
t
OCR En
g
in
e
,
”
Pro
c
e
e
d
in
g
s o
f
9
t
h
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Do
c
u
me
n
t
An
a
lys
is
a
n
d
Rec
o
g
n
it
i
o
n
,
Bra
zi
l
,
2
0
0
7
.
[1
7
]
A
.
A
b
d
u
lk
a
d
e
r
a
n
d
M
.
R.
Ca
ss
e
y
,
“
L
o
w
c
o
st
c
o
rre
c
ti
o
n
o
f
OCR
e
rro
rs
u
sin
g
lea
rn
in
g
i
n
a
m
u
lt
i
-
e
n
g
in
e
e
n
v
iro
n
m
e
n
t,
”
Pro
c
e
e
d
in
g
s
o
f
IE
EE
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Do
c
u
me
n
t
A
n
a
lys
is
a
n
d
Rec
o
g
n
it
i
o
n
,
ICDAR2
0
0
9
,
p
p
.
5
7
6
-
5
8
0
,
2
0
0
9
.
[1
8
]
T
.
M
.
Bre
u
e
l,
e
t
a
l.
,
“
Hig
h
-
p
e
rf
o
rm
a
n
c
e
OC
R
f
o
r
p
rin
te
d
E
n
g
li
sh
a
n
d
f
ra
k
tu
r
u
si
n
g
L
S
T
M
n
e
tw
o
rk
s,
”
Pro
c
e
e
d
in
g
s
o
f
I
EE
E
In
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
n
D
o
c
u
me
n
t
An
a
lys
is
a
n
d
Rec
o
g
n
it
i
o
n
,
ICDAR
2
0
1
3
.
W
a
sh
i
n
g
t
o
n
,
US
A
,
2
0
1
3
.
[1
9
]
M
.
He
li
ń
sk
i,
M
.
Km
iec
ia
k
,
T
P
a
rk
o
ła
,
“
Re
p
o
rt
o
n
t
h
e
c
o
m
p
a
ris
o
n
o
f
T
e
ss
e
r
a
c
t
a
n
d
A
BBYY
F
i
n
e
Re
a
d
e
r
OCR
e
n
g
in
e
s
,
”
Im
p
a
c
t,
2
0
1
2
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
A
p
p
lica
tio
n
o
n
ch
a
r
a
cter reco
g
n
itio
n
s
ystem
o
n
r
o
a
d
s
ig
n
fo
r
visu
a
lly
imp
a
ir
ed
:
ca
s
e
...
(
Ja
ejo
o
n
K
im)
785
[2
0
]
R
.
S
h
il
k
ro
t,
J
.
Hu
b
e
r,
W
.
M
.
E
e
,
P
.
M
a
e
s,
S
.
C
.
Na
n
a
y
a
k
k
a
ra
,
“
F
in
g
e
rRe
a
d
e
r:
A
W
e
a
ra
b
le
De
v
ice
to
Ex
p
lo
re
P
ri
n
ted
T
e
x
t
o
n
t
h
e
G
o
,
”
Pro
c
e
e
d
in
g
s
o
f
th
e
3
3
rd
An
n
u
a
l
ACM
Co
n
fer
e
n
c
e
o
n
H
u
ma
n
F
a
c
to
rs
in
Co
mp
u
ti
n
g
S
y
ste
ms
,
CHI 2
0
1
5
,
S
e
o
u
l,
K
o
re
a
,
2
0
1
5
.
[2
1
]
M
.
W
a
isb
o
u
rd
,
O
.
M
.
A
h
m
e
d
,
J
.
Ne
wm
a
n
,
M
.
S
a
h
u
,
D
.
R
o
b
i
n
so
n
,
L
.
S
ia
m
,
C
.
B.
Re
a
m
e
r,
T
.
Zh
a
n
,
M
.
G
o
ld
ste
in
,
S
.
Ku
rtz,
M
.
R.
M
o
ste
r,
L
.
A
.
Ha
rk
,
L
.
J
.
Ka
tz
,
“
T
h
e
Eff
e
c
t
o
f
a
n
In
n
o
v
a
ti
v
e
V
isio
n
S
im
u
lato
r
(OrCa
m
)
o
n
Qu
a
li
ty
o
f
L
i
f
e
in
P
a
ti
e
n
ts
w
it
h
G
lau
c
o
m
a
,
”
J
o
u
rn
a
l
o
f
Vi
su
a
l
Im
p
a
rime
n
t
&
Bl
in
d
n
e
ss
,
V
o
.
1
1
3
,
No
.
4
,
p
p
.
3
3
2
-
3
4
0
,
2
0
1
9
.
[2
2
]
Ko
re
a
Em
p
lo
y
m
e
n
t
De
v
e
lo
p
m
e
n
t
In
stit
u
te
,
“
2
0
1
7
Disa
b
led
S
taticti
c
s
(Ko
re
a
n
)
,
”
2
0
1
7
.
[2
3
]
W
HO
,
“
Gl
o
b
a
l
d
a
ta o
n
v
isu
a
l
im
p
a
irm
e
n
ts
,”
2
0
1
0
.
[2
4
]
L
.
Li
a
n
d
C
.
L
.
Tan
,
“
Ch
a
ra
c
te
r
Re
c
o
g
n
it
io
n
u
n
d
e
r
S
e
v
e
re
P
e
rs
p
e
c
ti
v
e
Disto
rti
o
n
,
”
Pro
c
e
e
d
in
g
s
o
f
IEE
E
1
9
t
h
In
ter
n
a
t
io
n
a
l
C
o
n
fer
e
n
c
e
o
n
P
a
tt
e
rn
Rec
o
g
n
it
io
n
,
IC
PR
2
0
0
8
.
T
a
m
p
a
,
U
S
A
,
2
0
0
8
.
[2
5
]
K.
W
a
n
g
,
e
t
a
l.
,
“
En
d
-
to
-
e
n
d
sc
e
n
e
tex
t
re
c
o
g
n
it
io
n
,”
Pro
c
e
e
d
i
n
g
s
o
f
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
C
o
mp
u
ter
Vi
si
o
n
,
ICCV2
0
1
1
,
Ba
rc
e
lo
n
a
,
S
p
a
i
n
,
2
0
1
1
.
[2
6
]
Z.
Zh
a
n
g
,
“
M
icro
so
f
t
Kin
e
c
t
se
n
s
o
r
a
n
d
it
s ef
f
e
c
t
,
”
IEE
E
M
u
lt
iM
e
d
ia
,
V
o
l.
1
9
,
N
o
.
2
,
p
p
.
4
-
1
0
,
2
0
1
2
.
B
I
O
G
RAP
H
I
E
S
O
F
AUTH
O
RS
J
a
e
jo
o
n
K
i
m
re
c
e
iv
e
d
th
e
M
S
a
n
d
P
h
.
D.
d
e
g
re
e
f
ro
m
th
e
De
p
a
rtme
n
t
o
f
El
e
c
tri
c
a
l
En
g
i
n
e
e
rin
g
,
Io
w
a
S
tate
Un
iv
e
rsit
y
,
US
A
.
He
re
c
e
iv
e
d
Ba
c
h
e
lo
rs
f
ro
m
De
p
a
rtme
n
t
El
e
c
tro
n
ics
En
g
in
e
e
rin
g
a
n
d
M
a
th
e
m
a
ti
c
s,
Ha
n
y
a
n
g
Un
iv
e
rsit
y
,
Ko
re
a
.
F
ro
m
2
0
0
1
to
2
0
0
2
,
h
e
w
a
s
a
s
e
n
io
r
re
se
a
rc
h
e
r
a
t
ET
RI
(El
e
c
tro
n
ics
a
n
d
T
e
lec
o
m
m
u
n
ica
ti
o
n
Re
se
a
rc
h
I
n
stit
u
te).
He
is
c
u
rre
n
t
ly
a
p
ro
f
e
ss
o
r
a
t
Da
e
g
u
Un
iv
e
rsit
y
,
R
e
p
u
b
li
c
o
f
Ko
re
a
.
His
re
se
a
r
c
h
in
tere
sts
in
c
lu
d
e
im
a
g
e
p
ro
c
e
ss
in
g
,
n
e
u
ra
l
n
e
tw
o
rk
s an
d
n
o
n
-
d
e
stru
c
ti
v
e
e
v
a
lu
a
ti
o
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.