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
.
4
,
A
u
g
u
s
t
2020
,
p
p
.
4
3
1
3
~
4
3
2
1
I
SS
N:
2
0
8
8
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v
1
0
i
4
.
pp
4
3
1
3
-
4
3
2
1
4313
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
Co
m
pa
riso
n b
etw
een handw
ritt
en
w
o
rd and sp
eech
record in
real
-
ti
m
e
using
C
NN
a
rchi
tectures
J
a
v
ier
O
.
P
inzó
n
-
Are
na
s
,
Ro
bin
s
o
n J
i
m
é
nez
-
M
o
re
no
F
a
c
u
lt
y
o
f
En
g
in
e
e
rin
g
,
Un
iv
e
rsid
a
d
M
i
li
tar
Nu
e
v
a
G
ra
n
a
d
a
,
Co
lo
m
b
ia
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Oct
1
0
,
2
0
1
9
R
ev
i
s
ed
Feb
5
,
2
0
2
0
A
cc
ep
ted
Feb
25
,
2
0
2
0
T
h
is
p
a
p
e
r
p
re
se
n
ts
th
e
d
e
v
e
lo
p
m
e
n
t
o
f
a
s
y
ste
m
o
f
c
o
m
p
a
riso
n
b
e
tw
e
e
n
w
o
rd
s
sp
o
k
e
n
a
n
d
w
rit
ten
b
y
m
e
a
n
s
o
f
d
e
e
p
lea
rn
i
n
g
tec
h
n
iq
u
e
s.
T
h
e
re
a
re
u
se
d
1
0
w
o
rd
s
a
c
q
u
ired
b
y
m
e
a
n
s
o
f
a
n
a
u
d
io
f
u
n
c
ti
o
n
a
n
d
,
th
e
se
sa
m
e
w
o
rd
s,
a
re
w
rit
ten
b
y
h
a
n
d
a
n
d
a
c
q
u
ired
b
y
a
w
e
b
c
a
m
,
in
su
c
h
a
w
a
y
a
s
to
v
e
rify
if
th
e
tw
o
d
a
ta
m
a
tch
a
n
d
sh
o
w
w
h
e
th
e
r
o
r
n
o
t
it
is
t
h
e
re
q
u
ired
w
o
rd
.
F
o
r
t
h
is,
2
d
if
f
e
r
e
n
t
CNN
a
rc
h
it
e
c
tu
re
s
w
e
re
u
se
d
f
o
r
e
a
c
h
f
u
n
c
ti
o
n
,
w
h
e
re
f
o
r
v
o
ice
re
c
o
g
n
it
io
n
,
a
s
u
it
a
b
le
CNN
w
a
s
u
se
d
t
o
i
d
e
n
ti
f
y
c
o
m
p
lete
w
o
rd
s
b
y
m
e
a
n
s
o
f
th
e
ir
f
e
a
tu
re
s
o
b
tain
e
d
w
it
h
m
e
l
f
re
q
u
e
n
c
y
c
e
p
stra
l
c
o
e
ff
icie
n
ts,
w
h
il
e
f
o
r
h
a
n
d
w
rit
in
g
,
a
f
a
ste
r
R
-
CNN
w
a
s
u
se
d
,
so
t
h
a
t
it
b
o
t
h
l
o
c
a
tes
a
n
d
i
d
e
n
ti
f
ies
th
e
c
a
p
tu
re
d
w
o
rd
.
T
o
im
p
le
m
e
n
t
th
e
sy
st
e
m
,
a
n
e
a
s
y
-
to
-
u
se
g
ra
p
h
ica
l
in
terf
a
c
e
w
a
s
d
e
v
e
lo
p
e
d
,
w
h
ich
u
n
it
e
s
th
e
tw
o
n
e
u
ra
l
n
e
tw
o
rk
s
f
o
r
it
s
o
p
e
ra
ti
o
n
.
W
it
h
th
is,
tes
ts
we
re
p
e
r
fo
rm
e
d
in
re
a
l
-
ti
m
e
,
o
b
tain
i
n
g
a
g
e
n
e
ra
l
a
c
c
u
ra
c
y
o
f
9
5
.
2
4
%
,
a
ll
o
w
in
g
sh
o
w
in
g
th
e
g
o
o
d
p
e
rf
o
rm
a
n
c
e
o
f
th
e
i
m
p
le
m
e
n
ted
s
y
ste
m
,
a
d
d
in
g
th
e
re
sp
o
n
se
sp
e
e
d
f
a
c
to
r,
b
e
in
g
les
s t
h
a
n
2
0
0
m
s in
m
a
k
in
g
th
e
c
o
m
p
a
riso
n
.
K
ey
w
o
r
d
s
:
C
NN
Fas
ter
R
-
C
NN
Han
d
w
r
iti
n
g
r
ec
o
g
n
itio
n
Sp
ee
ch
r
ec
o
g
n
i
tio
n
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
:
R
o
b
in
s
o
n
J
i
m
én
ez
-
Mo
r
en
o
,
Me
ch
atr
o
n
ic
s
E
n
g
in
ee
r
i
n
g
P
r
o
g
r
a
m
,
Fac
u
lt
y
o
f
E
n
g
i
n
ee
r
i
n
g
,
Un
i
v
er
s
id
ad
Militar
Nu
e
v
a
Gr
an
ad
a,
C
ar
r
er
a
1
1
#
1
0
1
-
8
0
,
B
o
g
o
tá
D.
C
.
,
C
o
lo
m
b
ia.
E
m
ail:
r
o
b
in
s
o
n
.
j
i
m
en
ez
@
u
n
i
m
ilit
ar
.
ed
u
.
co
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
i
m
p
le
m
e
n
tatio
n
o
f
p
atter
n
r
ec
o
g
n
itio
n
ap
p
licatio
n
s
i
n
d
if
f
er
e
n
t
t
y
p
e
s
o
f
d
ata,
s
u
c
h
as
s
i
g
n
al
s
f
r
o
m
s
e
n
s
o
r
s
[
1
]
,
au
d
io
[
2
,
3
]
,
o
r
ev
en
in
i
m
a
g
es
[
4
,
5
]
,
h
as
b
ee
n
g
r
o
w
i
n
g
ex
p
o
n
e
n
tia
ll
y
,
g
i
v
i
n
g
w
a
y
to
th
e
cr
ea
tio
n
o
f
a
g
r
ea
t
v
ar
iet
y
o
f
tech
n
iq
u
e
s
to
co
v
er
ea
c
h
t
y
p
e
o
f
p
atter
n
.
W
ith
i
n
t
h
e
s
e
tech
n
iq
u
es,
t
h
er
e
is
an
ar
ea
ca
lled
d
ee
p
lear
n
in
g
[
6
]
,
w
h
ich
co
n
tain
s
r
o
b
u
s
t
p
atter
n
r
ec
o
g
n
itio
n
m
et
h
o
d
s
s
u
c
h
as
r
ec
u
r
r
en
t
n
eu
r
al
n
e
t
w
o
r
k
s
[
7
,
8
]
.
T
h
at
a
r
e
m
ai
n
l
y
u
s
ed
f
o
r
s
p
ee
ch
r
ec
o
g
n
itio
n
an
d
w
r
i
tin
g
,
as
w
ell
as
d
e
ep
b
elief
n
eu
r
al
n
et
w
o
r
k
s
[
9
]
,
u
s
ed
f
o
r
i
m
ag
e
r
ec
o
g
n
itio
n
an
d
n
a
tu
r
al
la
n
g
u
a
g
e.
An
o
th
er
m
et
h
o
d
o
f
d
ee
p
lear
n
in
g
t
h
at
h
a
s
b
ee
n
e
v
o
lv
i
n
g
,
m
ai
n
l
y
s
i
n
ce
ea
r
l
y
2
0
1
2
[
6
]
,
is
th
e
co
n
v
o
lu
t
io
n
al
n
e
u
r
o
n
al
n
et
w
o
r
k
(
C
NN)
[
1
0
]
,
w
h
ic
h
w
a
s
o
r
ig
in
al
l
y
u
s
ed
to
r
ec
o
g
n
ize
o
n
l
y
p
atter
n
s
i
n
i
m
a
g
es.
Ho
w
e
v
er
,
d
u
e
to
t
h
e
d
em
o
n
s
tr
atio
n
o
f
its
ca
p
ab
ilit
y
to
r
ec
o
g
n
ize
u
p
to
1
0
0
0
d
if
f
e
r
en
t
ca
te
g
o
r
ies
[
1
1
]
an
d
to
s
u
p
p
o
r
t a
lar
g
e
n
u
m
b
er
o
f
h
id
d
en
la
y
er
s
,
w
h
ich
e
v
e
n
i
m
p
r
o
v
ed
its
p
er
f
o
r
m
an
ce
[
1
2
,
1
3
]
,
its
ap
p
licatio
n
h
as
b
ee
n
ex
te
n
d
ed
to
d
if
f
er
en
t
f
ield
s
,
b
ein
g
u
s
ed
f
o
r
s
p
ee
ch
r
ec
o
g
n
itio
n
[
1
4
,
1
5
]
,
elec
tr
o
m
y
o
g
r
ap
h
i
c
s
ig
n
al
s
[
1
6
,
1
7
]
,
an
d
ev
en
as
a
m
ea
n
s
o
f
h
u
m
a
n
-
m
ac
h
i
n
e
in
ter
ac
tio
n
[
1
8
,
1
9
]
.
W
ith
r
esp
ec
t
to
s
p
ee
ch
r
ec
o
g
n
itio
n
,
C
N
Ns
h
a
v
e
b
ee
n
ap
p
lied
s
u
cc
es
s
f
u
ll
y
i
n
d
i
f
f
er
en
t
d
ev
e
lo
p
m
en
ts
,
s
u
c
h
a
s
t
h
e
o
n
e
s
h
o
w
n
in
[
2
0
]
,
w
h
er
e
d
ee
p
C
NN
ar
ch
itectu
r
es
ar
e
co
m
p
ar
ed
ag
ain
s
t
a
d
ee
p
n
eu
r
al
n
et
w
o
r
k
f
o
r
lar
g
e
-
s
ca
le
s
p
ee
c
h
tas
k
s
,
h
av
i
n
g
a
r
elati
v
e
i
m
p
r
o
v
e
m
e
n
t
o
f
1
2
%
to
1
4
%.
I
n
an
o
th
er
d
ev
elo
p
m
e
n
t,
th
is
tec
h
n
iq
u
e
is
u
s
ed
to
r
ec
o
g
n
ize
6
d
if
f
er
e
n
t
la
n
g
u
a
g
es,
o
b
tai
n
in
g
a
w
o
r
d
er
r
o
r
r
ate
o
f
1
1
.
8
%
[
2
1
]
.
T
h
ese
w
o
r
k
s
u
s
u
a
ll
y
u
s
e
p
h
o
n
e
m
e
s
to
r
ec
o
g
n
ize
t
h
e
w
o
r
d
s
,
h
o
w
e
v
er
,
f
o
r
an
ap
p
licatio
n
w
h
er
e
n
o
t
th
e
w
h
o
le
o
f
a
lan
g
u
ag
e
is
u
s
ed
,
b
u
t
s
o
m
e
o
f
th
e
w
o
r
d
s
,
ex
tr
ac
ti
n
g
th
e
p
h
o
n
e
m
es
w
o
u
ld
r
esu
lt
in
a
v
e
r
y
co
m
p
le
x
tas
k
f
o
r
th
e
s
i
m
p
l
icit
y
o
f
t
h
e
w
o
r
k
,
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
.
4
,
A
u
g
u
s
t 2
0
2
0
:
4
3
1
3
-
4321
4314
f
o
r
th
is
r
ea
s
o
n
,
as
p
ar
t
o
f
t
h
e
co
n
tr
ib
u
tio
n
,
th
i
s
p
ap
er
p
r
o
p
o
s
es
u
s
in
g
a
C
NN
b
ased
o
n
co
m
p
lete
-
w
o
r
d
r
ec
o
g
n
itio
n
,
s
o
t
h
at
t
h
e
i
n
p
u
t
to
th
e
n
et
w
o
r
k
i
s
a
w
h
o
le
p
r
o
ce
s
s
ed
au
d
io
,
m
a
k
i
n
g
its
i
m
p
le
m
en
ta
tio
n
m
u
c
h
ea
s
ier
f
o
r
th
e
r
ec
o
g
n
itio
n
o
f
th
e
w
o
r
d
s
r
eq
u
ir
ed
.
On
th
e
o
th
er
h
an
d
,
r
eg
ar
d
in
g
th
e
lo
ca
tio
n
o
f
o
b
j
ec
ts
in
im
ag
es,
b
ec
au
s
e
C
NN
r
ec
o
g
n
iz
es
a
s
in
g
le
o
b
j
ec
t
in
a
to
tal
i
m
a
g
e,
a
v
ar
ia
n
t
w
a
s
d
ev
elo
p
ed
to
allo
w
it
t
o
id
en
tify
s
ev
er
al
o
b
j
ec
ts
w
it
h
in
t
h
e
s
a
m
e
i
m
a
g
e,
g
en
er
ati
n
g
a
n
e
w
v
ar
iatio
n
o
f
th
e
n
e
t
w
o
r
k
ca
lled
r
eg
io
n
-
b
a
s
ed
C
NN
(
R
-
C
N
N)
[
2
2
]
th
at,
in
co
m
b
i
n
atio
n
w
i
th
a
tech
n
iq
u
e
o
f
id
en
ti
f
y
i
n
g
r
eg
i
o
n
s
o
f
in
ter
est
(
R
o
I
)
,
th
e
n
et
wo
r
k
is
ab
le
to
lo
ca
te
an
d
id
en
tify
d
i
f
f
er
e
n
t
o
b
j
ec
ts
b
u
t
w
it
h
a
p
r
o
ce
s
s
in
g
ti
m
e
o
f
m
o
r
e
th
an
1
0
s
ec
o
n
d
s
,
m
a
k
i
n
g
i
t
v
er
y
s
lo
w
f
o
r
an
y
ap
p
li
ca
tio
n
in
r
ea
l
-
ti
m
e.
Fo
r
th
is
r
ea
s
o
n
,
o
th
er
v
ar
iati
o
n
s
w
er
e
i
m
p
le
m
e
n
ted
,
r
ea
ch
in
g
a
n
et
w
o
r
k
ca
lled
Fas
ter
R
-
C
NN
[
2
3
]
th
at,
in
s
tead
o
f
u
s
i
n
g
a
s
ep
ar
ate
al
g
o
r
ith
m
to
d
etec
t
R
o
I
s
,
h
a
s
a
r
eg
io
n
p
r
o
p
o
s
al
n
et
w
o
r
k
(
R
P
N)
w
it
h
w
h
ich
it
s
h
ar
es
t
h
e
w
ei
g
h
ts
a
n
d
lear
n
i
n
g
,
w
h
ic
h
n
o
t
o
n
l
y
in
cr
ea
s
e
s
it
s
r
o
b
u
s
tn
e
s
s
b
u
t
also
r
ed
u
ce
s
t
h
e
p
r
o
ce
s
s
in
g
ti
m
e
th
e
n
et
w
o
r
k
tak
es
to
lo
ca
te
all
th
e
o
b
j
ec
ts
o
f
in
ter
est
in
an
i
m
ag
e.
Ho
w
ev
er
,
alt
h
o
u
g
h
t
h
er
e
ar
e
w
o
r
k
s
t
h
at
C
NN
h
a
v
e
u
s
ed
to
r
ec
o
g
n
iz
e
h
an
d
w
r
itte
n
w
o
r
d
s
,
s
u
c
h
as
t
h
o
s
e
p
r
esen
ted
in
[
2
4
,
2
5
]
,
th
e
v
ar
iatio
n
to
lo
ca
te
th
e
m
w
i
th
o
u
t
th
e
n
ee
d
f
o
r
an
ad
d
itio
n
al
alg
o
r
ith
m
h
a
s
n
o
t
b
ee
n
u
s
ed
,
f
o
r
th
at
r
ea
s
o
n
,
t
h
is
p
ap
er
ex
p
lo
r
es
th
e
u
s
e
o
f
a
Fas
ter
R
-
C
NN
to
l
o
ca
te
1
0
d
if
f
er
e
n
t h
a
n
d
w
r
itte
n
w
o
r
d
s
,
to
v
er
if
y
t
h
eir
o
p
er
atio
n
.
T
h
is
w
o
r
k
,
in
ad
d
itio
n
to
th
e
af
o
r
em
e
n
tio
n
ed
,
s
ee
k
s
to
i
m
p
le
m
en
t
a
v
er
s
at
ile
ap
p
licatio
n
o
f
tr
u
t
h
ch
ec
k
b
et
w
ee
n
w
h
at
a
u
s
er
s
a
y
s
a
n
d
w
h
at
t
h
e
y
w
r
ite,
th
er
e
f
o
r
e,
it
is
p
r
o
p
o
s
ed
to
m
ak
e
th
e
co
m
p
ar
is
o
n
o
f
1
0
d
if
f
er
e
n
t
w
o
r
d
s
,
u
s
i
n
g
a
s
p
ee
ch
r
ec
o
g
n
it
io
n
s
y
s
te
m
a
n
d
o
n
e
o
f
lo
ca
tio
n
a
n
d
id
en
ti
f
icat
io
n
o
f
h
a
n
d
w
r
iti
n
g
,
u
n
i
ted
in
a
s
i
m
p
le
i
n
ter
f
ac
e,
w
h
ic
h
allo
w
s
d
o
in
g
t
h
is
ta
s
k
i
n
r
ea
l
-
ti
m
e.
T
h
is
p
ap
er
is
d
iv
i
d
ed
in
to
4
s
ec
tio
n
s
,
w
h
er
e
s
ec
tio
n
2
d
escr
ib
es
th
e
ar
ch
itect
u
r
es
i
m
p
le
m
en
te
d
f
o
r
s
p
ee
ch
r
ec
o
g
n
itio
n
an
d
h
an
d
w
r
itte
n
w
o
r
d
lo
ca
tio
n
an
d
id
en
ti
f
icatio
n
,
an
d
th
e
in
ter
f
ac
e
d
ev
elo
p
ed
f
o
r
th
e
tas
k
.
I
n
s
ec
tio
n
3
,
th
e
r
esu
lts
o
f
t
h
e
r
ea
l
-
ti
m
e
test
p
er
f
o
r
m
ed
ar
e
p
r
esen
ted
.
Fin
all
y
,
s
ec
t
io
n
4
s
h
o
w
s
th
e
c
o
n
clu
s
io
n
s
r
ea
ch
ed
.
2.
M
E
T
H
O
DS A
ND
M
AT
E
RI
AL
S
T
h
e
d
ev
elo
p
m
e
n
t
o
f
t
h
is
w
o
r
k
is
b
ased
o
n
th
e
r
ec
o
g
n
itio
n
o
f
1
0
d
if
f
er
e
n
t
w
o
r
d
s
i
n
th
e
Sp
an
is
h
lan
g
u
a
g
e
t
h
at
ca
n
b
e
w
r
it
ten
b
y
a
u
s
er
o
n
a
p
ap
er
,
w
h
ic
h
ar
e:
“A
b
aj
o
”
(
Do
w
n
)
,
“A
b
r
a”
(
Op
en
)
,
“A
r
r
ib
a”
(
Up
)
,
“
Av
a
n
ce
”
(
A
d
v
a
n
ce
)
,
“Ci
er
r
e”
(
C
lo
s
e)
,
“
Der
e”
(
s
h
o
r
t
e
n
in
g
m
o
d
if
ica
tio
n
o
f
R
i
g
h
t)
,
“
I
zq
u
i”
(
s
h
o
r
ten
in
g
m
o
d
i
f
icat
io
n
o
f
L
e
f
t)
,
“
No
”,
“
P
ar
e”
(
Sto
p
)
,
“
Si”
(
Yes)
,
ad
d
itio
n
all
y
,
th
e
r
ec
o
g
n
itio
n
o
f
th
ese
is
s
u
b
j
ec
t
to
w
h
at
t
h
e
u
s
er
s
a
y
s
b
y
m
ea
n
s
o
f
h
is
v
o
ice.
T
ak
in
g
i
n
t
o
ac
co
u
n
t
th
i
s
,
th
e
d
ev
e
lo
p
m
en
t
d
ep
en
d
s
o
n
th
e
i
m
p
le
m
e
n
tatio
n
o
f
t
w
o
m
ai
n
f
u
n
ct
io
n
s
,
w
h
ich
ar
e:
o
n
e
r
esp
o
n
s
ib
le
f
o
r
r
ec
o
g
n
izi
n
g
w
h
at
t
h
e
p
er
s
o
n
h
as
w
r
itte
n
a
n
d
an
o
th
er
r
esp
o
n
s
ib
le
f
o
r
p
r
o
ce
s
s
in
g
th
e
i
n
p
u
t
au
d
io
,
in
s
u
ch
a
w
a
y
t
h
at
it
id
en
ti
f
ies
t
h
e
w
o
r
d
th
at
t
h
e
u
s
er
s
a
y
s
.
Fo
r
th
i
s
,
t
w
o
d
if
f
er
e
n
t
C
N
N
ar
ch
itec
tu
r
es
ar
e
b
u
ilt
a
n
d
tr
ain
ed
i
n
o
r
d
er
to
ex
ec
u
te
s
aid
f
u
n
ctio
n
s
,
a
n
d
th
ese
ar
e
i
n
teg
r
ated
i
n
to
a
g
r
ap
h
ic
i
n
ter
f
ac
e
w
h
ich
i
s
w
h
at
th
e
u
s
er
u
s
es.
I
ts
d
ev
elo
p
m
e
n
t
is
ex
p
lai
n
ed
b
elo
w
.
2
.
1
.
CNN
f
o
r
ha
nd
w
rit
ing
re
co
g
nitio
n
T
h
e
f
u
n
c
tio
n
o
r
ien
ted
to
r
ec
o
g
n
ize
w
h
a
t
th
e
u
s
er
w
r
o
te
h
as
t
w
o
p
ar
ts
th
at
ar
e
th
e
lo
ca
tio
n
o
f
th
e
te
x
t
an
d
its
s
u
b
s
eq
u
en
t
cla
s
s
i
f
icati
o
n
.
T
h
e
o
b
j
ec
tiv
e
w
as
to
i
m
p
l
e
m
en
t
a
s
in
g
le
n
e
u
r
al
n
et
w
o
r
k
th
at
is
ca
p
ab
le
o
f
ca
r
r
y
i
n
g
o
u
t
b
o
th
p
ar
ts
,
f
o
r
t
h
is
r
ea
s
o
n
,
it
w
as
p
r
o
p
o
s
ed
to
i
m
p
le
m
e
n
t
a
n
e
u
r
al
n
et
w
o
r
k
b
ased
o
n
r
e
g
io
n
s
,
t
y
p
e
Fas
ter
R
-
C
NN
th
at,
th
a
n
k
s
to
its
h
ig
h
r
esp
o
n
s
e
ca
p
ac
ity
,
allo
w
s
t
h
e
d
ev
elo
p
m
en
t
o
f
alg
o
r
ith
m
s
ca
p
ab
le
o
f
lo
ca
tin
g
a
n
d
r
ec
o
g
n
izi
n
g
wo
r
d
s
in
r
ea
l
-
ti
m
e.
B
ased
o
n
th
i
s
,
th
e
ar
ch
itect
u
r
e
s
h
o
w
n
in
T
ab
le
1
is
p
r
o
p
o
s
ed
,
w
h
ic
h
co
n
s
i
s
ts
o
f
s
m
a
ll
s
q
u
ar
e
f
ilter
s
o
f
5
x
5
,
3
x
3
,
an
d
2
x
2
,
s
in
ce
in
ter
m
s
o
f
r
eso
l
u
tio
n
,
th
e
w
o
r
d
s
d
o
n
o
t
h
av
e
lar
g
e
s
ize,
g
i
v
i
n
g
t
h
e
p
o
s
s
ib
ilit
y
t
h
at
th
e
n
et
w
o
r
k
is
ab
le
to
lear
n
b
o
th
g
en
er
alitie
s
an
d
m
o
r
e
s
p
ec
if
ic
d
etails
o
f
ea
c
h
w
o
r
d
.
L
i
k
e
w
i
s
e,
ze
r
o
s
ar
e
ad
d
ed
to
th
e
ed
g
es
o
f
t
h
e
i
m
a
g
es
t
h
at
p
as
s
t
h
r
o
u
g
h
t
h
e
n
et
w
o
r
k
b
y
m
ea
n
s
o
f
p
ad
d
in
g
,
s
o
t
h
at
i
f
a
tr
ac
e
is
v
er
y
clo
s
e
to
th
e
ed
g
e,
th
e
n
et
w
o
r
k
i
s
ab
le
to
lear
n
o
r
tak
e
in
to
ac
co
u
n
t
its
ch
ar
ac
ter
is
tic
s
.
A
d
d
itio
n
al
l
y
,
co
u
p
led
to
th
e
C
NN,
th
er
e
is
a
s
ec
o
n
d
p
at
h
,
w
h
ic
h
i
s
a
r
eg
io
n
p
r
o
p
o
s
al
n
et
w
o
r
k
(
R
P
N)
,
in
c
h
ar
g
e
o
f
l
ea
r
n
in
g
t
h
e
r
eg
io
n
s
o
r
lo
ca
tio
n
o
f
th
e
w
o
r
d
s
.
T
h
ese
t
w
o
ar
e
j
o
in
ed
b
y
m
ea
n
s
o
f
a
r
eg
io
n
o
f
in
ter
est
p
o
o
lin
g
(
R
o
I
-
Ma
x
p
o
o
lin
g
)
,
w
h
ic
h
allo
w
s
to
h
a
v
e
d
y
n
a
m
ic
s
izes
o
f
t
h
e
i
n
co
m
in
g
f
ea
t
u
r
e
m
ap
s
to
th
is
la
y
er
,
i.e
.
d
e
p
en
d
in
g
o
n
th
e
b
o
x
lo
ca
ted
in
th
e
R
P
N,
it
p
er
f
o
r
m
s
th
e
co
r
r
esp
o
n
d
in
g
d
o
w
n
s
a
m
p
li
n
g
to
o
b
tain
t
h
e
r
eq
u
ir
ed
in
p
u
t
s
ize
in
th
e
n
e
x
t
la
y
er
.
A
d
atab
ase
co
n
tai
n
i
n
g
a
to
tal
o
f
1
2
0
0
g
r
a
y
s
ca
le
i
m
a
g
es
i
s
b
u
ilt,
o
f
w
h
ic
h
1
0
0
0
ar
e
u
s
ed
to
ca
r
r
y
o
u
t
t
h
e
tr
ai
n
in
g
o
f
t
h
e
n
et
w
o
r
k
a
n
d
2
0
0
to
v
alid
ate
it.
E
ac
h
i
m
a
g
e
w
as
ta
k
en
d
ir
ec
tl
y
b
y
a
w
eb
ca
m
a
n
d
co
n
t
ain
s
t
h
e
1
0
w
o
r
d
s
to
b
e
id
en
tif
ied
,
w
h
ic
h
w
er
e
m
an
u
all
y
lab
eled
,
as s
h
o
w
n
i
n
Fig
u
r
e
1
.
W
ith
th
is
,
t
h
e
tr
ain
in
g
o
f
t
h
e
n
eu
r
al
n
et
w
o
r
k
i
s
p
er
f
o
r
m
ed
an
d
later
its
v
alid
atio
n
.
I
n
th
i
s
last
s
tep
,
th
e
n
et
w
o
r
k
o
b
tain
s
9
8
.
9
%
o
f
av
er
ag
e
ac
c
u
r
ac
y
in
ter
m
s
o
f
th
e
id
e
n
ti
f
icatio
n
o
f
t
h
e
w
o
r
d
s
w
it
h
i
n
th
e
i
m
ag
e,
an
d
a
m
ea
n
a
v
er
ag
e
p
r
ec
is
io
n
(
m
A
P
)
in
t
h
e
lo
ca
tio
n
o
f
t
h
e
b
o
u
n
d
i
n
g
b
o
x
e
s
o
f
9
9
.
2
3
%.
T
h
e
av
er
ag
e
p
r
ec
is
io
n
o
f
ea
ch
w
o
r
d
is
s
h
o
w
n
i
n
Fi
g
u
r
e
2
,
w
i
th
t
h
e
w
o
r
d
"
Yes"
b
e
in
g
th
e
least
ac
c
u
r
ate,
w
it
h
9
8
%,
m
ai
n
l
y
b
ec
au
s
e
th
e
esti
m
ated
b
o
x
te
n
d
ed
to
b
e
a
litt
le
lar
g
er
th
a
n
t
h
e
g
r
o
u
n
d
tr
u
th
s
et.
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:
2
0
8
8
-
8708
C
o
mp
a
r
is
o
n
b
etw
ee
n
h
a
n
d
w
r
itten
w
o
r
d
a
n
d
s
p
ee
ch
r
ec
o
r
d
in
…
(
Ja
vier O
.
P
in
z
ó
n
-
A
r
en
a
s
)
4315
T
ab
le
1
.
Fas
ter
R
-
C
NN
a
r
ch
it
ec
tu
r
e
I
n
p
u
t
2
0
x
3
0
C
o
n
v
o
l
u
t
i
o
n
(
R
e
L
U
+
B
N
)
5
x
5
F
=
3
2
S
=
1
/
P
=
2
C
o
n
v
o
l
u
t
i
o
n
(
R
e
L
U
+
B
N
)
5
x
5
F
=
3
2
S
=
1
/
P
=
2
M
a
x
P
o
o
l
i
n
g
2
x
2
S
=
2
C
o
n
v
o
l
u
t
i
o
n
(
R
e
L
U
+
B
N
)
2
x
2
F
=
9
6
S
=
1
/
P
=
2
C
o
n
v
o
l
u
t
i
o
n
(
R
e
L
U
+
B
N
)
3
x
3
F
=
9
6
S
=
1
/
P
=
1
M
a
x
P
o
o
l
i
n
g
2
x
3
S
=
2
C
o
n
v
o
l
u
t
i
o
n
(
R
e
L
U
+
B
N
)
2
x
2
F
=
1
2
8
S
=
1
P
a
t
h
1
P
a
t
h
2
R
P
N
R
o
I
-
C
o
n
v
o
l
u
t
i
o
n
(
R
e
L
U
)
3
x
3
F
=
1
2
8
S
=
1
/
P
=
2
F
u
l
l
y
-
C
o
n
v
o
l
u
t
i
o
n
1
x
1
F
=
8
S
=
1
S
o
f
t
max
R
P
N
R
o
I
-
M
a
x
P
o
o
l
i
n
g
-
S
=
2
F
C
(
R
e
L
U
+
D
r
o
p
o
u
t
)
5
1
2
F
C
(
R
e
L
U
+
D
r
o
p
o
u
t
)
1
0
2
4
FC
11
S
o
f
t
max
C
l
a
ssi
f
i
c
a
t
i
o
n
N
o
t
e
:
S
=
S
t
r
i
d
e
,
P
=
P
a
d
d
i
n
g
,
F
=
F
i
l
t
e
r
s
,
B
N
=
B
a
t
c
h
N
o
r
mal
i
z
a
t
i
o
n
,
F
C
=
F
u
l
l
y
C
o
n
n
e
c
t
e
d
Fig
u
r
e
1
.
Sa
m
p
le
o
f
d
atab
ase
w
it
h
a
n
d
w
ith
o
u
t la
b
els
Fig
u
r
e
2
.
B
eh
av
io
r
o
f
th
e
R
o
I
d
etec
tio
n
at
d
if
f
er
en
t r
ec
all
le
v
els in
ea
c
h
ca
teg
o
r
y
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
.
4
,
A
u
g
u
s
t 2
0
2
0
:
4
3
1
3
-
4321
4316
2
.
2
.
CNN
f
o
r
ha
nd
w
rit
ing
re
co
g
nitio
n
T
o
im
p
le
m
en
t
t
h
e
s
p
ee
ch
r
ec
o
g
n
itio
n
f
u
n
ct
io
n
o
f
th
e
u
s
er
,
f
ir
s
t,
th
e
d
atab
ase
f
o
r
th
e
tr
ain
in
g
is
p
r
ep
ar
ed
,
in
s
u
c
h
a
w
a
y
t
h
at
th
e
d
i
m
e
n
s
io
n
s
o
f
t
h
e
n
et
w
o
r
k
i
n
p
u
t
ar
e
o
b
tain
ed
.
Fo
r
th
is
,
it
i
s
o
b
tain
ed
1
1
5
au
d
io
s
,
o
f
len
g
t
h
f
r
o
m
2
s
ec
o
n
d
s
to
1
6
k
Hz,
o
f
e
ac
h
w
o
r
d
o
f
d
i
f
f
er
e
n
t
p
eo
p
le
in
v
ar
ied
en
v
ir
o
n
m
e
n
ts
,
i.e
.
co
n
tr
o
lled
an
d
n
o
t
co
n
tr
o
lled
o
r
w
it
h
n
o
is
e,
m
ak
i
n
g
a
to
tal
o
f
1
1
5
0
d
ata
in
t
h
e
d
atab
ase.
Ho
w
e
v
er
,
as
it
is
w
an
ted
to
m
a
k
e
au
d
io
ac
q
u
is
itio
n
co
n
ti
n
u
o
u
s
l
y
,
th
er
e
is
a
n
ee
d
to
ad
d
an
ad
d
itio
n
al
ca
te
g
o
r
y
,
w
h
ic
h
i
n
th
is
ca
s
e
is
ca
lled
“
O
tr
o
s
”,
w
h
ich
co
n
ta
in
s
d
if
f
er
e
n
t
s
o
u
n
d
s
o
f
en
v
ir
o
n
m
e
n
t
(
p
eo
p
le
s
p
e
ak
in
g
,
b
lo
w
s
,
etc.
)
an
d
o
f
th
e
s
a
m
e
u
s
er
s
(
s
tr
o
n
g
b
r
ea
th
in
g
,
co
u
g
h
i
n
g
,
etc.
)
to
av
o
id
th
at
th
e
n
et
w
o
r
k
co
n
f
u
s
e
s
th
ese
s
o
u
n
d
s
w
it
h
th
e
w
o
r
d
s
.
Fro
m
th
i
s
c
ate
g
o
r
y
,
2
4
0
au
d
io
s
ar
e
r
ec
o
r
d
ed
.
I
n
o
r
d
er
f
o
r
th
e
au
d
io
s
to
b
e
u
s
ed
in
a
C
NN
f
o
cu
s
ed
o
n
s
p
ee
ch
r
ec
o
g
n
i
tio
n
,
p
r
o
ce
s
s
in
g
o
f
ea
ch
au
d
io
is
p
er
f
o
r
m
ed
b
y
m
ea
n
s
o
f
th
e
Me
l Fr
eq
u
e
n
c
y
C
ep
s
tr
al
C
o
e
f
f
icien
t
s
(
MF
C
C
)
,
i
n
s
u
c
h
a
w
a
y
t
h
at
a
f
ea
t
u
r
e
m
ap
is
o
b
tain
ed
to
b
e
en
ter
ed
in
to
th
e
C
NN.
Fo
r
th
e
p
r
o
ce
s
s
in
g
,
a
w
in
d
o
w
o
f
2
0
m
s
is
u
s
ed
w
i
th
a
s
h
i
f
t
o
f
1
0
m
s
,
o
b
tain
in
g
1
2
co
ef
f
ic
ien
t
s
p
er
f
r
a
m
e
a
n
d
a
to
tal
o
f
1
9
9
f
r
a
m
es.
T
o
h
av
e
a
clea
r
er
f
ea
t
u
r
e
m
ap
,
f
lo
o
r
f
ilter
is
ap
p
lied
to
th
e
m
ap
,
an
d
th
en
o
b
tain
th
e
f
ir
s
t
(
∆)
an
d
s
ec
o
n
d
(
∆∆
)
d
er
iv
ativ
e
f
r
o
m
t
h
e
MFC
C
o
b
tain
ed
,
to
ac
q
u
ir
e
b
etter
ch
ar
ac
ter
is
t
ics
o
f
t
h
e
s
o
u
n
d
a
n
d
th
a
t
t
h
e
n
e
t
w
o
r
k
h
as
a
g
r
ea
ter
p
o
s
s
ib
ilit
y
to
lear
n
th
e
te
m
p
o
r
al
v
ar
iatio
n
s
o
f
ea
ch
w
o
r
d
.
An
ex
a
m
p
le
o
f
t
h
ese
m
ap
s
i
s
p
r
esen
ted
in
Fig
u
r
e
3
.
Fin
all
y
,
th
e
d
atab
ase
o
f
th
e
f
ea
tu
r
e
m
ap
s
o
b
tain
ed
is
d
iv
id
ed
,
tak
in
g
1
0
0
m
ap
s
p
er
w
o
r
d
an
d
2
0
0
f
r
o
m
t
h
e
ca
teg
o
r
y
"
Oth
er
"
f
o
r
tr
ain
in
g
,
an
d
1
5
p
er
w
o
r
d
an
d
4
0
f
r
o
m
"
Otr
o
s
"
f
o
r
v
alid
atio
n
.
On
ce
th
e
d
atab
ase
is
o
b
tai
n
ed
,
C
NN
is
p
r
o
p
o
s
ed
f
o
r
th
e
ap
p
licatio
n
as s
h
o
w
n
i
n
T
a
b
le
2
.
Fig
u
r
e
3
.
Featu
r
e
m
ap
s
o
f
th
e
w
o
r
d
“A
b
aj
o
T
ab
le
2
.
C
NN
A
r
ch
i
tectu
r
e
p
r
o
p
o
s
ed
I
n
p
u
t
1
2
x
1
9
9
x
3
C
o
n
v
o
l
u
t
i
o
n
(
+
B
a
t
c
h
N
o
r
mal
i
z
a
t
i
o
n
)
5
x
5
F
=
3
2
S
=
1
/
P
=
2
C
o
n
v
o
l
u
t
i
o
n
5
x
5
F
=
3
2
S
=
1
/
P
=
2
M
a
x
P
o
o
l
i
n
g
2
x
1
S
=
2
x
1
C
o
n
v
o
l
u
t
i
o
n
3
x
3
F
=
6
4
S=
1
/
P
=
1
C
o
n
v
o
l
u
t
i
o
n
3
x
3
F
=
6
4
S
=
1
/
P
=
1
M
a
x
P
o
o
l
i
n
g
2
x
3
S
=
2
x
2
C
o
n
v
o
l
u
t
i
o
n
2
x
2
F
=
1
2
8
S
=
1
/
P
=
1
C
o
n
v
o
l
u
t
i
o
n
3
x
3
F
=
1
2
8
S
=
1
/
P
=
1
M
a
x
P
o
o
l
i
n
g
2
x
2
S
=
2
x
2
F
u
l
l
y
C
o
n
n
e
c
t
e
d
5
1
2
D
r
o
p
o
u
t
2
0
%
D
i
sco
n
n
e
c
t
e
d
F
u
l
l
y
C
o
n
n
e
c
t
e
d
2
0
4
8
D
r
o
p
o
u
t
2
0
%
D
i
sco
n
n
e
c
t
e
d
F
u
l
l
y
C
o
n
n
e
c
t
e
d
11
S
o
f
t
M
a
x
C
l
a
ssi
f
i
c
a
t
i
o
n
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:
2
0
8
8
-
8708
C
o
mp
a
r
is
o
n
b
etw
ee
n
h
a
n
d
w
r
itten
w
o
r
d
a
n
d
s
p
ee
ch
r
ec
o
r
d
in
…
(
Ja
vier O
.
P
in
z
ó
n
-
A
r
en
a
s
)
4317
Sin
ce
t
h
e
r
ec
o
g
n
it
io
n
to
b
e
m
a
d
e
is
o
f
t
h
e
co
m
p
lete
w
o
r
d
s
,
s
q
u
ar
e
f
ilter
s
ar
e
p
r
o
p
o
s
ed
,
in
s
u
ch
a
w
a
y
th
at
t
h
e
n
et
w
o
r
k
lear
n
s
t
h
e
ch
ar
ac
ter
is
tic
s
b
o
th
i
n
ter
m
s
o
f
t
h
e
co
ef
f
icie
n
t
s
a
n
d
th
e
te
m
p
o
r
al
v
ar
iatio
n
s
.
T
h
e
n
et
w
o
r
k
co
n
s
i
s
ts
o
f
3
p
a
ck
ets
2
co
n
v
o
lu
tio
n
s
p
lu
s
a
m
ax
p
o
o
lin
g
,
w
ith
t
h
e
d
if
f
er
e
n
ce
th
at
in
th
e
f
ir
s
t,
a
n
o
r
m
al
izatio
n
la
y
er
i
s
ad
d
ed
in
th
e
f
ir
s
t
co
n
v
o
l
u
tio
n
an
d
d
o
w
n
s
a
m
p
lin
g
is
m
ad
e
o
n
l
y
in
th
e
co
e
f
f
ic
ien
t
s
,
s
o
th
at
i
n
t
h
e
f
o
llo
w
i
n
g
co
n
v
o
l
u
tio
n
s
,
th
e
te
m
p
o
r
al
f
e
a
t
u
r
e
s
a
r
e
m
a
i
n
t
a
i
n
e
d
.
L
i
k
e
w
i
s
e
,
p
a
d
d
i
n
g
i
s
a
d
d
e
d
t
o
t
h
e
v
o
l
u
m
e
s
,
i
n
s
u
c
h
a
w
ay
t
h
a
t
t
h
e
c
h
a
r
a
c
t
e
r
i
s
t
i
c
s
f
o
u
n
d
i
n
t
h
e
f
i
r
s
t
a
n
d
l
a
s
t
c
o
e
f
f
i
c
i
e
n
t
s
a
r
e
t
a
k
e
n
i
n
t
o
a
c
c
o
u
n
t
a
t
t
h
e
m
o
m
e
n
t
t
h
a
t
t
h
e
l
e
a
r
n
i
n
g
f
i
l
t
e
r
s
p
a
s
s
o
v
e
r
t
h
e
m
.
Fin
all
y
,
t
h
e
tr
ain
i
n
g
o
f
t
h
e
n
et
w
o
r
k
i
s
ca
r
r
ie
d
o
u
t,
w
it
h
w
h
ic
h
an
ac
c
u
r
ac
y
o
f
1
0
0
%
an
d
9
0
%
v
alid
atio
n
i
s
o
b
tain
ed
,
as
s
h
o
w
n
in
T
ab
le
3
.
I
n
th
is
o
n
e,
it
ca
n
b
e
o
b
s
er
v
ed
th
at
th
e
w
o
r
d
s
th
at
g
o
t
t
h
e
m
o
s
t
er
r
o
n
eo
u
s
class
if
ica
tio
n
s
w
er
e
“A
r
r
ib
a”
,
“Av
an
ce
”,
“Ci
er
r
e”
,
an
d
“
I
zq
u
i”,
m
ain
l
y
w
h
e
n
th
er
e
w
a
s
a
lo
t
o
f
a
m
b
ien
t
n
o
is
e
w
it
h
p
eo
p
le
talk
in
g
,
w
h
ic
h
ca
u
s
es
th
e
c
h
ar
ac
ter
is
tics
m
ap
to
ch
a
n
g
e
to
s
o
m
e
e
x
ten
t,
ca
u
s
in
g
it
to
b
e
co
n
f
u
s
ed
w
it
h
o
th
er
w
o
r
d
s
.
T
ab
le
3
.
C
lass
if
icatio
n
o
f
ea
c
h
w
o
r
d
au
d
io
f
o
r
v
alid
atio
n
W
o
r
d
A
b
a
j
o
A
b
r
a
A
r
r
i
b
a
A
v
a
n
c
e
C
i
e
r
r
e
D
e
r
e
I
z
q
u
i
No
P
a
r
e
Si
O
t
r
o
s
O
v
e
r
a
l
l
a
cc
.
T
r
u
e
p
o
si
t
i
v
e
s
15
14
12
12
12
15
12
15
14
14
36
9
0
%
2
.
3
.
G
ra
ph
ic
us
er
inte
rf
a
ce
T
o
c
o
n
ca
ten
ate
t
h
e
o
p
er
atio
n
o
f
th
e
t
w
o
tr
ain
ed
C
NNs,
a
b
asic
g
r
ap
h
ic
u
s
er
in
ter
f
ac
e
(
GU
I
)
is
m
ad
e,
s
o
th
at
it
i
s
ea
s
y
to
u
s
e
f
o
r
an
y
u
s
er
.
T
h
is
i
n
ter
f
ac
e
is
c
o
m
p
o
s
ed
o
f
a
w
i
n
d
o
w
th
a
t
allo
w
s
v
i
s
u
a
lizi
n
g
i
n
r
ea
l
-
ti
m
e
th
e
w
o
r
d
w
r
it
ten
b
y
m
ea
n
s
o
f
a
w
eb
ca
m
.
A
d
d
itio
n
all
y
,
t
h
er
e
ar
e
4
b
u
tto
n
s
,
a
m
o
n
g
w
h
ic
h
2
ar
e
to
ac
tiv
ate
a
n
d
d
ea
ctiv
ate
t
h
e
web
ca
m
,
a
n
d
th
e
o
t
h
er
2
,
to
ac
tiv
ate
o
r
d
ea
ctiv
ate
t
h
e
d
ata
c
o
llectio
n
b
y
m
ea
n
s
o
f
a
m
icr
o
p
h
o
n
e.
Fi
n
all
y
,
th
er
e
is
a
b
o
x
t
h
at
s
h
o
w
s
th
e
w
o
r
d
s
aid
b
y
t
h
e
u
s
er
,
w
h
ic
h
c
h
a
n
g
e
s
co
lo
r
d
ep
en
d
in
g
o
n
2
ca
s
es: g
r
ee
n
,
if
t
h
e
w
o
r
d
m
en
tio
n
ed
b
y
t
h
e
u
s
er
is
t
h
e
s
a
m
e
as t
h
e
o
n
e
h
e
r
ec
o
g
n
ized
i
n
th
e
i
m
a
g
e,
o
r
r
ed
,
if
th
e
w
o
r
d
d
o
es
n
o
t
m
a
tch
th
e
o
n
e
r
ec
o
g
n
ized
in
t
h
e
i
m
ag
e.
Fo
r
th
e
i
n
ter
f
ac
e
to
s
tar
t
w
o
r
k
i
n
g
,
b
o
th
th
e
ca
m
er
a
a
n
d
th
e
au
d
io
i
n
p
u
t
m
u
s
t b
e
ac
tiv
a
ted
.
An
ex
a
m
p
le
o
f
th
e
i
n
ter
f
ac
e
i
s
s
h
o
w
n
i
n
Fig
u
r
e
4
.
Fig
u
r
e
4
.
Gr
ap
h
ic
u
s
er
i
n
ter
f
a
ce
3.
RE
SU
L
T
S AN
D
AN
AL
Y
SI
S
Var
io
u
s
r
ea
l
-
ti
m
e
test
s
o
f
th
e
GUI
ar
e
p
er
f
o
r
m
ed
to
s
ee
th
e
p
er
f
o
r
m
a
n
ce
o
f
b
o
th
th
e
p
r
o
p
o
s
ed
n
et
w
o
r
k
f
o
r
s
p
ee
ch
r
ec
o
g
n
iti
o
n
an
d
t
h
e
n
et
w
o
r
k
s
p
ec
ializ
ed
in
f
in
d
i
n
g
t
h
e
w
o
r
d
s
.
Fo
r
th
i
s
,
a
co
n
t
in
u
o
u
s
r
ec
o
r
d
in
g
alg
o
r
ith
m
is
ad
d
ed
,
in
s
u
c
h
a
w
a
y
t
h
at
o
n
ce
t
h
e
a
u
d
io
is
s
tar
ted
,
th
e
d
ata
is
tak
en
co
n
tin
u
o
u
s
l
y
,
h
o
w
ev
er
,
i
n
o
r
d
er
to
p
r
ev
en
t
en
v
ir
o
n
m
e
n
tal
n
o
is
e
f
r
o
m
b
ei
n
g
r
ec
o
g
n
ized
,
an
a
m
p
li
tu
d
e
t
h
r
esh
o
ld
o
f
t
h
e
au
d
io
s
ig
n
al
is
u
s
ed
w
it
h
r
esp
ec
t
to
its
av
er
ag
e
f
o
r
a
tim
e
o
f
2
0
0
m
s
.
I
f
th
e
av
er
ag
e
o
f
t
h
e
am
p
lit
u
d
e
ex
ce
ed
s
th
e
v
al
u
e
o
f
0
.
1
,
it
is
d
eter
m
in
ed
th
at
th
er
e
is
p
r
esen
ce
o
f
v
o
ice,
w
h
er
eb
y
th
e
d
ata
is
tak
en
f
r
o
m
a
ti
m
e
ti=t
-
7
0
0
m
s
to
t
f
=ti+2
0
0
0
m
s
,
w
h
e
r
e
t
i
s
t
h
e
c
u
r
r
e
n
t
t
im
e
,
t
i
i
s
t
h
e
t
i
m
e
o
f
b
e
g
i
n
n
i
n
g
o
f
d
a
t
a
c
o
l
l
e
c
t
i
o
n
a
n
d
t
f
is
t
h
e
f
i
n
a
l
t
i
m
e
o
f
d
a
t
a
c
o
l
l
e
c
t
i
o
n
,
o
b
t
a
i
n
i
n
g
a
n
a
u
d
i
o
o
f
2
s
e
c
o
n
d
s
.
I
f
t
h
e
t
h
r
e
s
h
o
l
d
i
s
n
o
t
s
u
r
p
a
s
s
e
d
,
t
h
e
a
u
d
i
o
d
a
t
a
i
s
n
o
t
s
a
v
e
d
.
W
it
h
t
h
i
s
,
it
i
s
p
r
o
ce
ed
ed
to
p
e
r
f
o
r
m
th
e
test
s
,
w
h
er
e
u
s
er
s
w
r
ite
th
e
w
o
r
d
s
i
n
d
iv
id
u
all
y
,
p
lacin
g
th
e
m
i
n
a
n
y
p
o
s
it
io
n
o
f
t
h
e
i
m
ag
e
an
d
e
v
e
n
w
it
h
a
ce
r
tai
n
d
eg
r
ee
o
f
i
n
cli
n
atio
n
to
in
c
r
ea
s
e
th
e
d
i
f
f
ic
u
lt
y
o
f
th
eir
r
ec
o
g
n
itio
n
.
A
f
ter
it
is
lo
ca
ted
,
th
e
u
s
er
s
s
a
y
s
o
m
e
o
f
t
h
e
w
o
r
d
s
,
eit
h
er
th
e
o
n
e
t
h
e
y
w
r
o
te
o
r
an
o
th
er
,
i
n
s
u
c
h
a
w
ay
t
h
a
t
t
h
e
i
n
t
e
r
f
a
c
e
s
h
o
w
s
t
h
e
m
i
f
w
h
a
t
t
h
e
y
s
a
y
m
a
t
c
h
w
i
t
h
w
h
a
t
t
h
e
y
w
r
o
t
e
.
T
h
e
s
e
t
e
s
t
s
a
r
e
c
a
r
r
i
e
d
o
u
t
w
i
t
h
6
u
s
e
r
s
,
e
a
c
h
o
n
e
w
i
t
h
7
t
r
i
e
s
,
i
.
e
.
t
h
e
r
e
w
e
r
e
d
o
n
e
4
2
t
e
s
t
s
,
o
b
t
a
i
n
i
n
g
t
h
e
r
e
s
u
l
t
s
s
h
o
w
n
i
n
T
a
b
l
e
4
.
I
n
ea
ch
o
f
th
e
test
s
,
th
e
s
p
e
ec
h
r
ec
o
g
n
itio
n
n
et
w
o
r
k
w
as
ab
le
t
o
c
o
r
r
ec
tly
id
en
ti
f
y
e
ac
h
w
o
r
d
m
en
tio
n
ed
b
y
t
h
e
u
s
e
r
s
.
R
e
g
ar
d
in
g
th
e
id
en
ti
f
ica
tio
n
o
f
th
e
w
o
r
d
s
in
th
e
i
m
a
g
es,
9
5
.
2
4
%
o
f
g
en
er
al
ac
cu
r
ac
y
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
.
4
,
A
u
g
u
s
t 2
0
2
0
:
4
3
1
3
-
4321
4318
w
a
s
o
b
tain
ed
,
b
ein
g
ab
le
to
r
ec
o
g
n
ize
2
0
w
o
r
d
s
co
r
r
ec
tly
an
d
in
d
icate
an
er
r
o
r
o
f
m
a
tc
h
in
g
w
h
en
t
h
e
u
s
er
s
n
a
m
ed
a
n
o
th
er
w
o
r
d
2
0
ti
m
e
s
.
E
ig
h
t
o
f
th
e
s
e
test
s
ca
n
b
e
o
b
s
er
v
ed
in
Fi
g
u
r
e
5
,
w
h
e
r
e
w
h
en
th
e
w
o
r
d
w
a
s
co
r
r
ec
tl
y
clas
s
i
f
ied
,
th
e
i
d
en
tific
atio
n
co
n
f
id
e
n
ce
w
as
g
r
ea
ter
th
a
n
9
0
%,
in
o
t
h
er
w
o
r
d
s
,
th
e
n
et
w
o
r
k
w
a
s
"
s
u
r
e"
th
at
i
t
w
as
t
h
e
w
o
r
d
.
On
t
h
e
o
th
er
h
a
n
d
,
th
e
id
en
t
if
icatio
n
n
et
w
o
r
k
o
f
t
h
e
w
o
r
d
s
in
t
h
e
i
m
a
g
e
h
ad
2
er
r
o
n
eo
u
s
class
i
f
icat
io
n
s
,
t
h
at
is
,
w
h
en
t
h
e
u
s
er
m
en
tio
n
ed
a
w
o
r
d
an
d
an
o
th
er
w
as
o
n
th
e
w
eb
ca
m
,
th
i
s
w
a
s
r
ec
o
g
n
ized
as
tr
u
e,
h
o
w
e
v
er
,
as
s
h
o
w
n
i
n
Fi
g
u
r
e
5
.
g
a
n
d
Fi
g
u
r
e
5
.
h
,
th
e
co
n
f
id
en
ce
is
e
v
en
lo
w
er
th
a
n
8
0
%,
w
h
ic
h
is
s
o
l
v
ed
w
it
h
a
co
n
f
id
en
ce
t
h
r
es
h
o
ld
,
s
o
t
h
a
t
w
h
e
n
a
w
o
r
d
i
s
m
is
c
lass
if
ie
d
,
if
t
h
is
t
h
r
esh
o
ld
is
n
o
t e
x
ce
ed
ed
,
it is
ta
k
en
a
s
a
n
eg
ati
v
e.
T
ab
le
4
.
R
esu
lts
o
b
tain
ed
f
r
o
m
th
e
te
s
ts
R
i
g
h
t
s
p
e
e
c
h
r
e
c
o
g
n
i
t
i
o
n
T
r
u
e
p
o
si
t
i
v
e
s
T
r
u
e
n
e
g
a
t
i
v
e
s
F
a
l
s
e
p
o
si
t
i
v
e
s
F
a
l
s
e
n
e
g
a
t
i
v
e
s
O
v
e
r
a
l
l
a
c
c
u
r
a
c
y
42
20
20
2
0
9
5
.
2
4
%
(
a)
(
b
)
(
c)
(
d
)
(
e)
(
f
)
(
g
)
(
h
)
Fig
u
r
e
5
.
Sa
m
p
les o
f
t
h
e
tes
t p
er
f
o
r
m
ed
,
(
a,
b
,
c)
th
e
w
r
itte
n
w
o
r
d
is
th
e
o
n
e
to
ld
b
y
t
h
e
u
s
e
r
,
(
d
,
e,
f
)
th
e
w
r
itte
n
w
o
r
d
d
if
f
er
s
f
r
o
m
w
h
at
t
h
e
u
s
er
s
aid
,
(
g
,
h
)
th
e
w
r
itte
n
w
o
r
d
is
w
r
o
n
g
l
y
id
en
tifie
d
,
p
ass
in
g
it a
s
co
r
r
ec
t c
o
m
p
ar
ed
w
it
h
w
h
at
w
as
s
aid
b
y
th
e
u
s
e
r
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:
2
0
8
8
-
8708
C
o
mp
a
r
is
o
n
b
etw
ee
n
h
a
n
d
w
r
itten
w
o
r
d
a
n
d
s
p
ee
ch
r
ec
o
r
d
in
…
(
Ja
vier O
.
P
in
z
ó
n
-
A
r
en
a
s
)
4319
I
n
o
r
d
er
to
b
etter
u
n
d
er
s
tan
d
h
o
w
t
h
e
u
s
ed
Fas
ter
R
-
C
N
N
is
b
eh
av
i
n
g
,
th
e
ac
t
iv
at
io
n
s
o
b
tai
n
ed
b
y
al
l
th
e
w
o
r
d
s
i
n
th
e
la
s
t
la
y
er
o
f
t
h
e
n
et
w
o
r
k
ar
e
s
h
o
w
n
F
i
g
u
r
e
6
.
A
s
ca
n
b
e
s
ee
n
,
t
h
e
n
et
w
o
r
k
ac
tiv
a
tes
th
e
m
o
s
t
r
elev
an
t
r
eg
io
n
s
o
f
ea
ch
w
o
r
d
,
ev
en
in
th
e
w
o
r
d
“
C
ier
r
e”
,
it
f
o
cu
s
e
s
o
n
th
e
w
h
o
le
o
f
it,
w
h
er
e
th
e
r
ed
co
lo
r
m
ea
n
s
g
r
ea
ter
ac
tiv
at
io
n
.
On
t
h
e
o
th
er
h
an
d
,
in
ce
r
tain
ca
te
g
o
r
ies
it
ac
tiv
ate
s
th
e
ed
g
es
o
f
th
e
s
h
ee
t
w
h
er
e
it
h
as
b
ee
n
w
r
itte
n
,
m
ai
n
l
y
i
n
th
e
w
o
r
d
“
Si”,
w
h
er
e
p
o
s
s
ib
l
y
co
n
f
u
s
e
t
h
e
s
h
ad
o
w
w
i
th
t
h
e
s
ec
tio
n
o
f
th
e
letter
"
i
"
,
h
o
w
e
v
er
,
n
o
t
f
i
n
d
in
g
m
o
r
e
f
ea
t
u
r
es,
t
h
e
n
et
w
o
r
k
d
o
es n
o
t id
en
ti
f
y
it
as a
w
o
r
d
.
As
f
o
r
t
h
e
p
r
o
ce
s
s
i
n
g
ti
m
e
o
f
th
e
alg
o
r
it
h
m
,
o
n
ce
y
o
u
h
av
e
f
i
n
is
h
ed
tak
i
n
g
t
h
e
au
d
io
,
th
e
p
r
o
ce
s
s
o
f
o
b
tain
in
g
th
e
MFC
C
s
,
u
s
i
n
g
th
e
v
o
ice
r
ec
o
g
n
itio
n
n
et
w
o
r
k
,
u
s
in
g
t
h
e
Fas
ter
R
-
C
NN
a
n
d
d
is
p
la
y
i
n
g
t
h
e
r
es
u
lt
i
n
th
e
g
r
ap
h
ic
in
ter
f
ac
e,
it
tak
es
a
n
av
er
a
g
e
o
f
1
5
1
m
s
,
i.e
.
o
n
ce
th
e
u
s
er
f
in
is
h
es
s
a
y
i
n
g
t
h
e
w
o
r
d
,
th
e
alg
o
r
it
h
m
id
en
ti
f
ies
w
h
et
h
er
o
r
n
o
t it
m
atch
e
s
w
it
h
w
h
a
t is
w
r
itte
n
al
m
o
s
t i
n
s
ta
n
tl
y
.
Fig
u
r
e
6
.
Fas
ter
R
-
C
N
N
ac
tiv
atio
n
s
f
o
r
s
o
m
e
w
o
r
d
s
4.
CO
NCLU
SI
O
N
T
h
is
p
ap
er
d
escr
ib
ed
th
e
i
m
p
le
m
e
n
tatio
n
o
f
a
co
m
p
ar
is
o
n
s
y
s
te
m
(
ex
ec
u
t
ed
i
n
r
ea
l
-
ti
m
e
)
b
et
w
ee
n
h
an
d
w
r
itte
n
an
d
s
p
o
k
e
n
w
o
r
d
s
b
y
a
u
s
er
s
o
th
at
ea
c
h
o
f
th
e
f
u
n
c
tio
n
s
d
e
v
elo
p
ed
w
as
i
m
p
le
m
e
n
ted
w
it
h
d
if
f
er
e
n
t
C
NN
ar
ch
i
tectu
r
e
s
,
w
h
ic
h
p
r
esen
ted
a
v
alid
atio
n
ac
cu
r
ac
y
o
f
9
8
.
9
%
an
d
9
0
%
f
o
r
th
e
r
ec
o
g
n
itio
n
o
f
h
an
d
w
r
iti
n
g
an
d
s
p
ee
ch
,
r
esp
ec
tiv
el
y
.
On
th
e
o
th
er
h
an
d
,
w
h
e
n
p
er
f
o
r
m
i
n
g
th
e
test
s
in
r
ea
l
-
ti
m
e,
th
e
s
p
ee
ch
r
ec
o
g
n
it
io
n
f
u
n
cti
o
n
d
id
n
o
t
h
av
e
er
r
o
r
s
w
h
e
n
r
ec
o
g
n
izi
n
g
ea
c
h
w
o
r
d
th
at
t
h
e
u
s
er
s
aid
,
th
a
t
is
,
it
id
en
ti
f
ied
1
0
0
%
o
f
th
e
4
2
t
ests
p
er
f
o
r
m
ed
.
A
d
d
itio
n
all
y
,
f
o
r
th
e
s
y
s
t
e
m
o
f
r
ec
o
g
n
itio
n
an
d
lo
ca
lizatio
n
o
f
th
e
h
a
n
d
w
r
itte
n
w
o
r
d
s
,
t
h
e
d
if
f
ic
u
lt
y
o
f
th
e
te
s
ts
w
a
s
in
cr
ea
s
ed
,
n
o
t
o
n
l
y
b
y
n
o
t
co
n
tr
o
lli
n
g
th
e
a
m
b
ien
t
l
ig
h
t,
b
u
t
also
b
y
p
laci
n
g
t
h
e
m
w
it
h
a
ce
r
tai
n
d
e
g
r
ee
o
f
in
cl
in
at
i
o
n
,
in
o
r
d
er
to
test
it
s
p
er
f
o
r
m
an
ce
,
o
b
tain
i
n
g
an
ac
cu
r
ac
y
o
f
9
5
.
2
4
%
o
f
th
e
g
en
er
al
s
y
s
te
m
d
e
v
elo
p
ed
,
w
h
er
e
th
er
e
w
er
e
o
n
l
y
2
ca
s
es
o
f
f
alse
p
o
s
iti
v
es,
i.e
.
w
h
e
n
th
e
u
s
er
s
aid
an
y
o
f
th
e
w
o
r
d
s
,
th
e
s
y
s
te
m
er
r
o
n
eo
u
s
l
y
s
aid
th
a
t
th
e
w
r
itte
n
o
n
e
m
atc
h
ed
,
also
,
n
o
f
alse
n
eg
a
tiv
e
s
w
er
e
o
b
tain
ed
d
u
r
in
g
th
e
tes
ts
p
er
f
o
r
m
ed
.
W
ith
th
is
,
it
ca
n
b
e
o
b
s
er
v
ed
th
e
v
er
y
g
o
o
d
p
er
f
o
r
m
a
n
ce
o
f
t
h
e
g
e
n
er
al
s
y
s
te
m
i
m
p
le
m
en
ted
an
d
t
h
e
g
r
e
at
ca
p
ac
ity
t
h
at
t
h
e
C
NN
s
h
a
v
e
to
b
e
u
s
ed
b
o
th
in
s
p
ee
ch
r
ec
o
g
n
itio
n
u
s
i
n
g
co
m
p
lete
w
o
r
d
s
an
d
in
lo
ca
tio
n
an
d
r
ec
o
g
n
itio
n
o
f
h
a
n
d
w
r
itin
g
,
in
ad
d
itio
n
to
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
.
4
,
A
u
g
u
s
t 2
0
2
0
:
4
3
1
3
-
4321
4320
th
e
f
ac
t
th
at
it
i
s
ca
p
ab
le
o
f
m
ak
in
g
t
h
e
co
m
p
ar
is
o
n
al
m
o
s
t
i
m
m
ed
iatel
y
r
e
g
ar
d
less
o
f
w
h
e
th
er
2
r
o
b
u
s
t
n
eu
r
al
n
et
w
o
r
k
s
ar
e
b
ein
g
u
s
ed
,
w
i
th
a
ti
m
e
o
f
1
5
1
m
s
.
T
ak
in
g
i
n
to
ac
co
u
n
t
t
h
e
ab
o
v
e,
t
h
is
s
y
s
te
m
ca
n
b
e
i
m
p
le
m
en
ted
an
d
co
m
p
le
m
e
n
ted
f
o
r
its
ap
p
licatio
n
in
d
if
f
er
en
t
f
u
tu
r
e
w
o
r
k
s
,
f
o
r
ex
a
m
p
le,
b
e
u
s
ed
in
co
u
n
tr
ies
w
h
er
e
th
e
t
r
a
d
i
t
i
o
n
a
l
s
y
s
t
e
m
o
f
v
o
t
e
c
o
u
n
t
i
n
g
i
n
e
l
e
c
t
o
r
a
l
s
y
s
t
e
m
s
i
s
s
t
i
l
l
u
s
e
d
,
s
o
t
h
a
t
w
h
e
n
t
h
e
p
e
r
s
o
n
s
a
y
s
t
h
e
n
u
m
b
e
r
o
f
v
o
t
e
s
f
o
r
a
c
a
n
d
i
d
a
t
e
,
t
h
e
n
u
m
b
e
r
w
r
i
t
t
e
n
o
n
t
h
e
f
o
r
m
m
u
s
t
b
e
t
h
e
s
a
m
e
,
m
a
i
n
l
y
t
o
h
e
l
p
a
v
o
i
d
f
r
a
u
d
i
n
e
l
e
c
t
i
o
n
s
.
ACK
NO
WL
E
D
G
E
M
E
NT
S
T
h
e
au
th
o
r
s
ar
e
g
r
atef
u
l
to
th
e
Un
iv
er
s
id
ad
Militar
Nu
ev
a
Gr
an
ad
a,
w
h
ic
h
,
th
r
o
u
g
h
its
Vice
-
r
ec
to
r
y
f
o
r
r
esear
ch
,
f
in
a
n
ce
s
th
e
p
r
esen
t
p
r
o
j
ec
t
w
it
h
co
d
e
I
MP
-
I
NG
-
2
2
9
0
(
2
0
1
7
-
2
0
1
8
)
an
d
ti
tled
"
Pro
to
tip
o
d
e
r
o
b
o
t a
s
is
ten
te
p
ar
a
cir
u
g
ía,
"
f
r
o
m
w
h
ic
h
th
e
p
r
ese
n
t
w
o
r
k
is
d
er
iv
ed
.
RE
F
E
R
E
NC
E
S
[1
]
A
.
Ha
a
g
,
S
.
G
o
ro
n
z
y
,
P
.
S
c
h
a
ich
,
a
n
d
J.
W
il
li
a
m
s,
“
E
m
o
ti
o
n
re
c
o
g
n
it
io
n
u
sin
g
b
io
-
se
n
s
o
rs:
F
i
rst
ste
p
s
t
o
w
a
rd
s
a
n
a
u
to
m
a
ti
c
s
y
ste
m
,
”
T
u
to
ria
l
a
n
d
Res
e
a
rc
h
W
o
rk
sh
o
p
o
n
A
ff
e
c
ti
v
e
Dia
lo
g
u
e
S
y
ste
ms
,
p
p
.
3
6
-
4
8
,
2
0
0
4
.
[
2
]
G
.
D
.
S
r
e
e
,
P
.
C
h
a
n
d
r
a
s
e
k
h
a
r
,
a
n
d
B
.
V
e
n
k
a
t
e
s
h
u
l
u
,
“
S
V
M
b
a
s
e
d
s
p
e
e
c
h
e
m
o
t
i
o
n
r
e
c
o
g
n
i
t
i
o
n
c
o
m
p
a
r
e
d
w
i
t
h
G
M
M
-
U
B
M
a
n
d
N
N
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
E
n
g
i
n
e
e
r
i
n
g
S
c
i
e
n
c
e
a
n
d
C
o
m
p
u
t
i
n
g
,
v
o
l
.
6
,
n
o
.
1
1
,
p
p
.
3
2
9
3
-
3
2
9
8
,
2
0
1
6
.
[
3
]
H
.
F
.
P
a
r
d
e
d
e
,
A
.
R
.
Y
u
l
i
a
n
i
,
a
n
d
R
.
S
u
s
t
i
k
a
,
“
C
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
a
n
d
f
e
a
t
u
r
e
t
r
a
n
s
f
o
r
m
a
t
i
o
n
f
o
r
d
i
s
t
a
n
t
s
p
e
e
c
h
r
e
c
o
g
n
i
t
i
o
n
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
E
l
e
c
t
r
i
c
a
l
&
C
o
m
p
u
t
e
r
E
n
g
i
n
e
e
r
i
n
g
,
v
o
l
.
8
,
n
o
.
6
,
p
p
.
5
3
8
1
-
5
3
8
8
,
2
0
1
8
[4
]
D.
Cireş
a
n
,
U.
M
e
ier
,
a
n
d
J.
S
c
h
m
id
h
u
b
e
r,
“
M
u
lt
i
-
c
o
lu
m
n
d
e
e
p
n
e
u
ra
l
n
e
tw
o
rk
s
f
o
r
i
m
a
g
e
c
las
si
f
ica
ti
o
n
,
”
I
E
E
E
C
o
n
f
e
r
e
n
c
e
o
n
C
o
m
p
u
t
e
r
V
i
s
i
o
n
a
n
d
P
a
t
t
e
r
n
R
e
c
o
g
n
i
t
i
o
n
,
p
p
.
1
-
8
,
2
0
1
2
[5
]
M
.
S
.
H.
A
l
-
T
a
m
i
m
i,
“
Co
m
b
in
in
g
c
o
n
v
o
l
u
ti
o
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
s
a
n
d
sla
n
tl
e
t
tran
sf
o
rm
f
o
r
a
n
e
ffe
c
ti
v
e
i
m
a
g
e
re
tri
e
v
a
l
sc
h
e
m
e
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
&
Co
mp
u
ter
En
g
i
n
e
e
rin
g
,
v
o
l.
9
,
n
o
.
5
,
p
p
.
4
3
8
2
-
4
3
9
5
,
2
0
1
9
[6
]
Y.
L
e
Cu
n
,
Y.
Be
n
g
io
,
a
n
d
G
.
Hin
to
n
,
“
De
e
p
lea
rn
in
g
,
”
Na
t
u
re
,
v
o
l.
5
2
1
,
n
o
.
7
5
5
3
,
p
p
.
4
3
6
,
2
0
1
5
[7
]
R.
P
a
sc
a
n
u
,
C.
G
u
lce
h
re
,
K.
C
h
o
,
a
n
d
Y
.
Be
n
g
io
,
“
Ho
w
to
c
o
n
stru
c
t
d
e
e
p
re
c
u
rre
n
t
n
e
u
ra
l
n
e
tw
o
rk
s,”
2nd
In
ter
n
a
t
io
n
a
l
C
o
n
fer
e
n
c
e
o
n
L
e
a
rn
in
g
Rep
re
se
n
ta
ti
o
n
s (
ICL
R
)
,
p
p
.
1
-
1
3
,
2
0
1
4.
[8
]
M
.
A
.
Jish
a
n
,
K.
R.
M
a
h
m
u
d
,
a
n
d
A
.
K.
A
l
A
z
a
d
,
“
Na
tu
ra
l
lan
g
u
a
g
e
d
e
sc
rip
ti
o
n
o
f
ima
g
e
s
u
sin
g
h
y
b
rid
re
c
u
rre
n
t
n
e
u
ra
l
n
e
tw
o
rk
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
Co
m
p
u
ter
En
g
i
n
e
e
rin
g
,
v
o
l.
9
,
n
o
4
,
p
p
.
2
9
3
2
-
2
9
4
0
,
2
0
1
9
[9
]
G
.
E.
Hin
to
n
,
S
.
Os
i
n
d
e
ro
,
a
n
d
Y.
W
.
T
e
h
,
“
A
f
a
st
l
e
a
rn
in
g
a
lg
o
rit
h
m
f
o
r
d
e
e
p
b
e
li
e
f
n
e
ts,”
Ne
u
r
a
l
C
o
mp
u
ta
ti
o
n
,
v
o
l.
1
8
,
n
o
.
7
,
p
p
.
1
5
2
7
-
1
5
5
4
,
2
0
0
6
[1
0
]
M
.
D.
Zeiler
a
n
d
R.
F
e
rg
u
s,
“
Vi
su
a
li
z
in
g
a
n
d
u
n
d
e
rsta
n
d
in
g
c
o
n
v
o
lu
ti
o
n
a
l
n
e
tw
o
rk
s,”
Eu
ro
p
e
a
n
C
o
n
fer
e
n
c
e
o
n
Co
mp
u
ter
V
isio
n
,
p
p
.
8
1
8
-
8
3
3
,
2
0
1
4
.
[1
1
]
A
.
Kriz
h
e
v
sk
y
,
I.
S
u
tsk
e
v
e
r
,
a
n
d
G
.
E.
Hin
to
n
,
“
Im
a
g
e
n
e
t
c
las
si
f
ic
a
ti
o
n
w
it
h
d
e
e
p
c
o
n
v
o
l
u
ti
o
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
s,”
Ad
v
a
n
c
e
s i
n
Ne
u
ra
l
In
f
o
rm
a
ti
o
n
Pro
c
e
ss
in
g
S
y
ste
ms
,
p
p
.
1
0
9
7
-
1
1
0
5
,
2
0
1
2
.
[1
2
]
K.
S
im
o
n
y
a
n
a
n
d
A
.
Zi
ss
e
rm
a
n
,
“
V
e
ry
d
e
e
p
c
o
n
v
o
l
u
ti
o
n
a
l
n
e
tw
o
rk
s
f
o
r
larg
e
-
sc
a
l
e
ima
g
e
re
c
o
g
n
it
io
n
,
”
In
ter
n
a
t
io
n
a
l
C
o
n
fer
e
n
c
e
o
n
L
e
a
rn
in
g
Rep
re
se
n
ta
ti
o
n
s (
ICL
R
)
,
p
p
.
1
-
1
4
,
2
0
1
5.
[1
3
]
C.
S
z
e
g
e
d
y
,
V
.
V
a
n
h
o
u
c
h
e
,
S
.
L
o
ffe
,
J.
S
h
len
s,
a
n
d
Z
.
B.
W
o
jn
a
,
“
Re
th
in
k
i
n
g
th
e
in
c
e
p
t
io
n
a
rc
h
it
e
c
tu
re
f
o
r
c
o
m
p
u
ter v
isio
n
,
”
IEE
E
Co
n
fer
e
n
c
e
o
n
Co
m
p
u
ter
Vi
si
o
n
a
nd
Pa
tt
e
rn
Rec
o
g
n
it
i
o
n
,
p
p
.
2
8
1
8
-
2
8
2
6
,
2
0
1
6
.
[1
4
]
O.
A
b
d
e
l
-
Ha
m
id
,
e
t
a
l
,
“
Co
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
s
f
o
r
sp
e
e
c
h
re
c
o
g
n
it
io
n
,
”
IEE
E/
AC
M
T
ra
n
sa
c
ti
o
n
s
on
A
u
d
io
,
S
p
e
e
c
h
,
a
n
d
L
a
n
g
u
a
g
e
Pro
c
e
ss
in
g
,
v
o
l
.
2
2
,
n
o
.
1
0
,
p
p
.
1
5
3
3
-
1
5
4
5
,
2
0
1
4
[1
5
]
S
.
G
a
n
a
p
a
th
y
a
n
d
V.
P
e
d
d
in
ti
,
“3
-
D
CNN
m
o
d
e
ls
f
o
r
f
a
r
-
f
ield
m
u
lt
i
-
c
h
a
n
n
e
l
s
p
e
e
c
h
re
c
o
g
n
it
io
n
,
”
IEE
E
In
ter
n
a
t
io
n
a
l
C
o
n
fer
e
n
c
e
o
n
Ac
o
u
stics
,
S
p
e
e
c
h
a
n
d
S
ig
n
a
l
Pro
c
e
ss
in
g
,
p
p
.
5
4
9
9
-
5
5
0
3
,
2
0
1
8
.
[1
6
]
J.
O.
P
.
A
re
n
a
s,
R.
J
.
M
o
re
n
o
,
a
n
d
R.
D.
H
.
Be
leñ
o
,
“
EM
G
sig
n
a
l
a
c
q
u
isit
io
n
a
n
d
p
ro
c
e
ss
in
g
a
p
p
li
c
a
ti
o
n
w
it
h
CNN
t
e
stin
g
f
o
r
M
AT
LAB,
”
In
ter
n
a
ti
o
n
a
l
Rev
iew o
f
A
u
to
m
a
ti
c
Co
n
tro
l,
v
o
l.
1
1
,
n
o
.
1
,
p
p
.
4
4
-
5
1
,
2
0
1
8
.
[1
7
]
L
.
Ch
e
n
,
J.
F
u
,
Y.
W
u
,
H
.
L
i,
a
n
d
B.
Z
h
e
n
g
,
“
Ha
n
d
g
e
stu
re
re
c
o
g
n
it
io
n
u
sin
g
c
o
m
p
a
c
t
CNN
v
ia
su
rf
a
c
e
e
lec
tro
m
y
o
g
ra
p
h
y
sig
n
a
ls
,
”
S
e
n
so
rs
,
v
o
l.
2
0
,
n
o
.
3
,
p
p
.
1
-
14
,
2
0
2
0
.
[1
8
]
C.
Nu
z
z
i,
e
t
a
l.
,
“
De
e
p
lea
rn
i
n
g
b
a
se
d
m
a
c
h
in
e
v
isio
n
:
F
irst
ste
p
s
to
w
a
rd
s
a
h
a
n
d
g
e
stu
re
re
c
o
g
n
it
io
n
se
t
u
p
f
o
r
c
o
ll
a
b
o
ra
ti
v
e
ro
b
o
ts,”
W
o
rk
sh
o
p
o
n
M
e
tro
l
o
g
y
f
o
r In
d
u
stry
4
.
0
a
n
d
Io
T
,
p
p
.
2
8
-
33
,
2
0
1
8
.
[1
9
]
S.
C.
Hs
u
,
Y.
W
.
W
a
n
g
,
a
n
d
C.
L
.
Hu
a
n
g
,
“
Hu
m
a
n
o
b
jec
t
id
e
n
ti
f
ica
ti
o
n
f
o
r
h
u
m
a
n
-
ro
b
o
t
in
tera
c
ti
o
n
b
y
u
sin
g
f
a
st
R
-
CNN
,
”
2nd
IEE
E
I
n
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
R
o
b
o
ti
c
Co
m
p
u
t
i
n
g
,
p
p
.
2
0
1
-
2
0
4
,
2
0
1
8
.
[2
0
]
T.
N.
S
a
in
a
th
,
e
t
a
l
,
“
De
e
p
c
o
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
s
f
o
r
lar
g
e
-
sc
a
le
sp
e
e
c
h
tas
k
s,”
Ne
u
ra
l
Ne
two
rk
s,
v
o
l.
6
4
,
p
p
.
3
9
-
4
8
,
2
0
1
5
.
[2
1
]
T
.
S
e
rc
u
,
C.
P
u
h
rsc
h
,
B.
Kin
g
sb
u
ry
,
a
n
d
Y.
L
e
c
u
n
,
“
V
e
ry
d
e
e
p
m
u
lt
il
in
g
u
a
l
c
o
n
v
o
lu
t
io
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
s
f
o
r
L
V
CS
R,
”
IEE
E
I
n
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Ac
o
u
stics
,
S
p
e
e
c
h
a
n
d
S
i
g
n
a
l
Pr
o
c
e
ss
in
g
,
p
p
.
4
9
5
5
-
4
9
5
9
,
2
0
1
6
.
[2
2
]
R.
G
irsh
ick
,
J.
Do
n
a
h
u
e
,
T
.
Da
rre
ll
,
a
n
d
J.
M
a
li
k
,
“
Rich
f
e
a
tu
re
h
iera
rc
h
ies
f
o
r
a
c
c
u
ra
te
o
b
jec
t
d
e
tec
ti
o
n
a
n
d
se
m
a
n
ti
c
se
g
m
e
n
tatio
n
,
”
IEE
E
c
o
n
fer
e
n
c
e
o
n
c
o
mp
u
ter
v
isio
n
a
n
d
p
a
tt
e
rn
re
c
o
g
n
it
io
n
,
p
p
.
5
8
0
-
587
,
2
0
1
4
.
[2
3
]
S
.
Re
n
,
K.
He
,
R.
G
irsh
ick
,
a
n
d
J.
S
u
n
,
“
F
a
ste
r
R
-
CNN
:
T
o
w
a
rd
s
re
a
l
-
ti
m
e
o
b
jec
t
d
e
tec
ti
o
n
w
it
h
r
e
g
io
n
p
r
o
p
o
sa
l
n
e
tw
o
rk
s,”
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
Pa
tt
e
rn
An
a
lys
is a
n
d
M
a
c
h
in
e
I
n
telli
g
e
n
c
e
,
v
o
l.
3
9
,
n
o
.
6
,
p
p
.
1
1
3
7
-
1
1
4
9
,
2
0
1
5
.
[2
4
]
S
.
S
u
d
h
o
lt
a
n
d
G
.
A
.
F
in
k
,
“
P
H
OCN
e
t:
A
d
e
e
p
c
o
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
f
o
r
w
o
rd
sp
o
tt
in
g
in
h
a
n
d
w
rit
ten
d
o
c
u
m
e
n
ts,”
1
5
t
h
I
n
ter
n
a
t
io
n
a
l
C
o
n
fer
e
n
c
e
o
n
,
Fr
o
n
ti
e
rs
i
n
Ha
n
d
w
riti
n
g
Rec
o
g
n
it
i
o
n
,
p
p
.
2
7
7
-
2
8
2
,
2
0
1
6
.
[
2
5
]
K.
Zag
o
ris,
I
.
P
ra
ti
k
a
k
is
,
a
n
d
B.
G
a
to
s,
“
Un
su
p
e
rv
ise
d
w
o
rd
s
p
o
tt
in
g
i
n
h
isto
rica
l
h
a
n
d
w
rit
ten
d
o
c
u
m
e
n
t
ima
g
e
s
u
s
i
n
g
d
o
c
u
m
e
n
t
-
o
r
i
e
n
t
e
d
l
o
c
a
l
f
e
a
t
u
r
e
s
,
”
I
E
E
E
T
r
a
n
s
a
c
t
i
o
n
s
o
n
I
m
a
g
e
P
r
o
c
e
s
s
i
n
g
,
v
o
l
.
2
6
,
n
o
.
8
,
p
p
.
4
0
3
2
-
4
0
4
1
,
2
0
1
7
.
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:
2
0
8
8
-
8708
C
o
mp
a
r
is
o
n
b
etw
ee
n
h
a
n
d
w
r
itten
w
o
r
d
a
n
d
s
p
ee
ch
r
ec
o
r
d
in
…
(
Ja
vier O
.
P
in
z
ó
n
-
A
r
en
a
s
)
4321
B
I
O
G
RAP
H
I
E
S
O
F
AUTH
O
RS
J
a
v
ier
O
r
la
n
d
o
Pi
n
z
ó
n
Ar
e
n
a
s
w
a
s
b
o
rn
i
n
S
o
c
o
rr
o
-
S
a
n
tan
d
e
r,
C
o
lo
m
b
ia,
i
n
1
9
9
0
.
He
re
c
e
iv
e
d
h
is
d
e
g
re
e
in
M
e
c
h
a
tro
n
ics
E
n
g
in
e
e
rin
g
(C
u
m
L
a
u
d
e
)
in
2
0
1
3
,
S
p
e
c
ializa
ti
o
n
i
n
En
g
in
e
e
rin
g
P
ro
jec
t
M
a
n
a
g
e
m
e
n
t
in
2
0
1
6
,
a
n
d
M
.
S
c
.
in
M
e
c
h
a
tro
n
ics
En
g
i
n
e
e
rin
g
in
2
0
1
9
,
a
t
th
e
Un
iv
e
rsid
a
d
M
i
li
tar
Nu
e
v
a
G
ra
n
a
d
a
-
UMN
G
.
He
h
a
s
e
x
p
e
rien
c
e
in
th
e
a
re
a
s
o
f
a
u
to
m
a
ti
o
n
,
e
lec
tro
n
ic
c
o
n
tr
o
l,
a
n
d
m
a
c
h
in
e
lea
rn
in
g
.
C
u
rre
n
tl
y
,
h
e
is
stu
d
y
in
g
a
P
h
.
D.
in
A
p
p
li
e
d
S
c
ien
c
e
s
a
n
d
w
o
rk
in
g
a
s
a
G
r
a
d
u
a
te
As
sista
n
t
a
t
t
h
e
UMNG
w
it
h
e
m
p
h
a
sis
o
n
Ro
b
o
ti
c
s a
n
d
M
a
c
h
in
e
L
e
a
rn
i
n
g
,
E
-
m
a
il
:
u
3
9
0
0
2
3
1
@
u
n
im
il
it
a
r.
e
d
u
.
c
o
Ro
b
i
n
s
o
n
J
i
m
é
n
e
z
M
o
r
e
n
o
w
a
s
b
o
r
n
i
n
B
o
g
o
tá,
C
o
lo
m
b
ia,
i
n
1
9
7
8
.
He
re
c
e
iv
e
d
th
e
En
g
i
n
e
e
r
d
e
g
re
e
in
El
e
c
tro
n
ics
a
t
th
e
F
ra
n
c
isc
o
Jo
sé
d
e
Ca
ld
a
s
District
Un
iv
e
rsit
y
-
UD
-
in
2
0
0
2
.
M
.
S
c
.
in
In
d
u
strial
A
u
to
m
a
ti
o
n
f
ro
m
th
e
Un
iv
e
rsid
a
d
Na
c
io
n
a
l
d
e
Co
lo
m
b
ia
-
2
0
1
2
a
n
d
P
h
.
D
.
in
En
g
in
e
e
ri
n
g
a
t
th
e
F
ra
n
c
isc
o
Jo
sé
d
e
Ca
ld
a
s
District
Un
iv
e
rsit
y
-
2
0
1
8
.
He
is
c
u
rre
n
tl
y
w
o
rk
in
g
a
s
a
P
ro
f
e
ss
o
r
in
t
h
e
M
e
c
h
a
tro
n
ics
En
g
in
e
e
rin
g
P
r
o
g
ra
m
a
t
th
e
Un
iv
e
rsid
a
d
M
il
it
a
r
N
u
e
v
a
G
ra
n
a
d
a
-
UMN
G
.
H
e
h
a
s
e
x
p
e
rien
c
e
in
th
e
a
re
a
s
o
f
In
stru
m
e
n
t
a
ti
o
n
a
n
d
El
e
c
tro
n
ic
Co
n
tr
o
l,
a
c
ti
n
g
m
a
in
ly
in
Ro
b
o
t
ics
,
c
o
n
t
ro
l
,
p
a
tt
e
rn
re
c
o
g
n
it
io
n
,
a
n
d
im
a
g
e
p
ro
c
e
ss
in
g
.
E
-
m
a
il
:
ro
b
in
so
n
.
ji
m
e
n
e
z
@u
n
im
il
it
a
r.
e
d
u
.
c
o
Evaluation Warning : The document was created with Spire.PDF for Python.