I
n
t
e
r
n
at
i
on
al
Jou
r
n
al
of
E
l
e
c
t
r
i
c
al
an
d
C
o
m
p
u
t
er E
n
g
i
n
eeri
n
g
(
I
J
E
C
E
)
V
o
l.
11
,
N
o.
5
,
O
ct
o
b
er
202
1
, p
p
.
42
81
~
428
8
I
SSN
:
2088
-
8708
,
D
O
I
:
10.
11
591/
i
j
ece
.
v
11
i
5
.
p
p42
81
-
428
8
4281
Jou
r
n
al
h
om
e
p
age
:
h
ttp
:
//ije
c
e
.
ia
e
s
c
o
r
e
.
c
o
m
D
o
w
n s
y
n
d
ro
m
e
d
et
ect
i
o
n
u
s
i
ng
m
o
dif
ie
d
a
d
ab
o
os
t
a
lg
o
ri
t
h
m
V
in
c
y
D
e
v
i V
.
K
1
,
R
a
j
e
s
h
R
2
1
D
ep
ar
t
m
en
t
o
f
C
o
m
p
u
t
er
S
ci
en
ce,
B
h
ar
at
h
i
ar
U
n
i
v
er
s
i
t
y
,
C
o
i
m
b
at
o
r
e,
I
n
d
i
a
2
D
ep
ar
t
m
en
t
o
f
C
o
m
p
u
t
er
S
ci
en
ce,
CH
RIS
T
(
D
ee
m
ed
t
o
b
e
U
n
iv
e
r
s
ity
)
,
B
a
ng
a
l
or
e
,
I
ndi
a
A
rt
i
cl
e I
n
f
o
AB
S
T
RAC
T
A
r
tic
le
h
is
to
r
y
:
R
ecei
v
ed
J
un
14
, 2
0
20
Re
v
i
se
d
M
ar
1
, 2
0
2
1
A
ccep
t
ed
M
ar
1
4
, 2
0
2
1
I
n h
um
a
n body
g
e
ne
t
i
c
c
ode
s
a
r
e
s
t
or
e
d i
n
t
he
g
e
ne
s
.
A
l
l
of
our
i
nhe
r
i
t
e
d
t
r
ai
t
s
ar
e a
s
s
o
ci
at
ed
w
i
t
h
t
h
es
e g
e
n
es
an
d
ar
e g
r
o
u
p
ed
as
s
t
r
u
ct
u
r
es
g
en
er
al
l
y
cal
l
ed
ch
r
o
m
o
s
o
m
es
.
I
n
t
y
p
i
c
al
cas
es
,
ea
ch
cel
l
co
n
s
i
s
t
s
o
f
2
3
p
ai
r
s
o
f
c
hr
om
os
om
e
s
,
out
of
w
hi
c
h e
a
c
h
pa
r
e
nt
c
o
nt
r
i
b
ut
e
s
ha
l
f
.
B
ut
i
f
a
pe
r
s
on
ha
s
a
pa
r
t
i
a
l
or
f
ul
l
c
opy
of
c
hr
o
m
os
o
m
e
21,
t
he
s
i
t
ua
t
i
o
n i
s
c
a
l
l
e
d D
ow
n
s
y
n
d
r
o
m
e
.
I
t r
e
s
u
lts
in
in
te
lle
c
tu
a
l
d
is
a
b
ili
ty
,
r
e
a
d
in
g
im
p
a
ir
m
e
n
t,
d
ev
el
o
p
m
en
t
al
d
el
a
y
,
an
d
o
t
h
er
m
ed
i
cal
ab
n
o
r
m
al
i
t
i
es
.
T
h
er
e i
s
n
o
s
p
eci
f
ic
t
r
e
a
t
m
e
nt
f
or
D
ow
n s
y
ndr
om
e
.
T
hus
,
e
a
r
l
y
de
t
e
c
t
i
on a
nd s
c
r
e
e
ni
ng
of
t
hi
s
di
s
a
bi
l
i
t
y
a
r
e
t
he
be
s
t
s
t
y
l
e
s
f
or
dow
n s
y
ndr
om
e
pr
e
v
e
nt
i
on.
I
n
t
hi
s
w
or
k
,
r
eco
g
n
i
t
i
o
n
o
f
D
o
w
n
s
y
n
d
r
o
m
e
u
t
i
l
i
zes
a s
et
o
f
f
aci
al
ex
p
r
es
s
i
o
n
i
m
a
g
es
.
S
o
li
d
g
e
o
m
e
tr
ic
d
e
s
c
r
ip
to
r
is
e
m
p
l
o
y
ed
f
o
r
ex
t
r
act
i
n
g
t
h
e f
aci
al
f
ea
t
u
r
es
f
r
o
m
t
h
e i
m
a
g
e s
et
.
A
n
A
d
aB
o
o
s
t
m
et
h
o
d
i
s
p
r
act
i
ced
t
o
g
at
h
er
t
h
e r
eq
u
i
r
ed
d
at
a
s
e
t
s
a
nd f
or
t
he
c
a
t
e
g
or
i
z
a
t
i
on.
T
he
e
x
t
r
a
c
t
e
d i
nf
or
m
a
t
i
on i
s
t
he
n a
s
s
i
g
ne
d
a
nd us
e
d
t
o
i
ns
t
r
uc
t
t
he
N
e
ur
a
l
N
e
t
w
or
k
us
i
ng
B
a
c
k
pr
opa
g
a
t
io
n
a
lg
o
r
ith
m
.
T
hi
s
w
or
k
r
e
c
or
de
d t
ha
t
t
he
pr
e
s
e
nt
e
d m
ode
l
m
e
e
t
s
t
he
r
e
qui
r
e
m
e
nt
w
i
t
h
9
8
.
6
7
%
accu
r
acy
.
Ke
y
wo
rd
s
:
A
da
B
oos
t
a
lg
o
r
ith
m
B
ack
p
r
o
p
ag
at
i
o
n
B
a
ye
s
c
la
s
s
if
ie
r
D
o
w
n s
ynd
r
o
m
e
F
a
c
ia
l p
o
in
ts
T
hi
s
i
s
an
ope
n ac
c
e
s
s
ar
t
i
c
l
e
u
nd
e
r
t
he
CC B
Y
-
SA
l
i
cen
s
e.
Co
rre
sp
o
n
d
i
n
g
Au
t
h
o
r
:
V
i
nc
y D
e
v
i
V
.
K
.
D
ep
ar
t
m
en
t
o
f
C
o
m
p
u
t
er
S
ci
e
n
ce
B
h
a
r
a
th
ia
r
U
n
iv
e
r
s
it
y
C
oi
m
ba
t
or
e
,
I
n
di
a
E
m
a
il: v
i
n
c
y
d
e
v
i1
5
@
g
m
a
i
l.
c
o
m
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
be
a
r
i
n
g
of
a
pa
r
t
i
a
l
or
f
ul
l
c
opy
of
c
h
r
o
m
os
o
m
e
21,
c
o
m
m
o
n
l
y
c
a
l
l
e
d T
r
i
s
om
y
21 r
e
s
ul
t
s
i
n
a
s
e
t
o
f
m
ed
i
cal
ab
n
o
r
m
al
i
t
i
e
s
cal
l
ed
d
o
w
n
s
ynd
r
o
m
e
(
D
S
)
.
A
n
i
nd
i
vi
d
ua
l
w
i
t
h
D
S
m
a
y
ha
v
e
i
nd
i
vi
d
ua
l
a
nd
/
o
r
i
nt
e
l
l
e
c
t
ua
l
d
e
ve
l
o
p
m
e
nt
i
s
s
ue
s
[
1
]
.
S
t
ud
i
e
s
ha
ve
s
ho
w
n
t
h
a
t
a s
o
u
l
s
u
f
f
er
i
n
g
f
r
o
m
D
S
e
x
h
i
b
i
t
d
i
v
er
s
e
f
aci
al
f
eat
u
r
es
[
2
]
.
T
h
e D
S
af
f
ect
ed
p
er
s
o
n
s
m
a
y
al
s
o
p
o
s
s
es
s
s
o
m
e ch
ar
act
er
i
s
t
i
c
s
t
h
at
ar
e a s
m
al
l
ch
i
n
,
a f
l
at
n
a
s
al
br
i
dg
e
,
f
eeb
l
e
m
u
s
c
le
to
n
e
,
a
n
d
tilte
d
e
y
e
s
[
3
]
.
D
i
a
g
n
o
s
i
s
o
f
D
S
m
a
y
b
e e
x
ec
u
t
ed
ear
l
i
er
o
r
a
f
te
r
p
a
r
tu
r
itio
n
.
I
n
th
e
f
ir
s
t c
a
s
e
,
b
io
-
c
h
e
m
i
cal
ev
al
u
at
i
o
n
a
nd
c
yt
o
ge
ne
t
i
c
t
e
s
t
s
c
a
n b
e
d
o
ne
f
o
r
t
he
d
i
a
g
no
s
i
s
o
f
D
S
.
I
n t
he
s
eco
n
d
cas
e,
t
h
e
e
xi
s
t
e
nc
e
o
f
a
num
be
r
of
m
i
n
or
phy
s
i
c
a
l
a
n
d
m
e
di
c
a
l
m
a
l
f
or
m
a
t
i
ons
i
n
a
d
d
itio
n
to
m
i
ni
a
t
ur
e
e
a
r
s
pr
ot
r
u
de
d
t
on
gue
,
up
s
l
a
nt
i
ng p
a
l
p
e
b
r
a
l
f
i
s
s
ur
e
s
,
ep
i
can
t
h
al
f
o
l
d
,
f
la
t f
a
c
ia
l
lo
o
k
s
,
a
n
d
e
x
tr
e
m
it
y
v
a
r
ia
b
le
p
o
i
n
t
s
t
o
t
h
e ch
a
n
ce f
o
r
D
S
[
4
]
.
F
o
r
au
t
o
m
at
ed
,
co
m
p
u
t
er
-
ai
d
ed
a
na
l
ys
i
s
o
f
D
S
,
t
h
e i
m
ag
e p
r
o
ces
s
,
an
d
an
al
y
s
i
s
ca
n
s
er
v
e as
an
e
f
f
ect
i
v
e an
d
p
o
w
er
f
u
l
cr
eat
u
r
e.
L
oos
et
al
.
[
5
]
s
t
u
d
i
ed
t
h
e ch
ar
act
er
i
s
t
i
c
f
aci
al
p
at
t
er
n
s
u
s
i
n
g
G
ab
o
r
w
a
v
el
et
f
eat
u
r
es
f
o
r
t
en
s
ynd
r
o
m
e
s
,
e
xc
l
ud
i
ng D
o
w
n
s
ynd
r
o
m
e
.
F
ur
t
he
r
s
t
ud
i
e
s
[
6
]
gi
ve
a
c
l
a
s
s
i
f
i
c
a
t
i
o
n f
o
r
1
4
s
ynd
r
o
m
e
s
,
b
ut
o
nl
y
21%
a
c
c
u
r
a
c
y
.
E
v
e
n
s
o,
t
h
a
t
m
e
t
h
od f
a
i
l
e
d
t
o di
s
t
i
ng
ui
s
h
t
he
s
ynd
r
o
m
e
s
f
r
o
m
a
t
i
d
y o
n
e
.
I
n i
m
p
r
o
ve
r
,
t
he
m
et
h
o
d
cal
l
s
f
o
r
m
an
u
al
l
ab
el
i
n
g
a
n
d
s
o
i
t
w
as
n
o
t
a f
u
l
l
y
a
u
t
o
m
at
ed
t
ec
h
n
i
q
u
e.
L
at
er
,
S
a
r
ay
d
e
m
i
r
et
al
.
[
7
]
pr
op
os
e
d a
G
a
bor
w
a
v
e
l
e
t
t
r
a
n
s
f
or
m
t
e
c
hn
i
qu
e
f
or
t
h
e
de
t
e
c
t
i
on
of
D
S
a
n
d t
h
e
t
e
c
h
n
i
qu
e
pos
s
e
s
s
e
s
h
i
gh
accu
r
ac
y
.
B
u
r
çi
n
a
n
d
V
as
i
f
d
e
v
el
o
p
ed
a
lo
c
a
l b
in
a
r
y
p
a
tte
r
n
(
L
B
P
)
b
as
ed
co
n
cep
t
f
or
t
h
e
D
S
di
a
gn
os
i
s
.
H
e
r
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN
:
20
88
-
8708
In
t
J
E
l
e
c
&
C
o
m
p
E
n
g
,
V
o
l.
11
, N
o
.
5
,
O
c
t
obe
r
20
2
1
:
4
2
8
1
-
4
2
8
8
4282
th
e
y
u
s
e
d
a
te
m
p
la
te
m
a
tc
h
i
n
g
s
tr
a
te
g
y
.
S
till,
t
h
e
s
e
m
e
th
o
d
s
ta
k
e
s
o
m
e
s
o
r
t o
f
m
a
n
u
a
l p
r
e
-
pr
oc
e
s
s
i
n
g f
or
t
h
e
su
c
c
e
ss
o
f
t
h
e
d
i
a
g
n
o
si
s p
r
o
c
e
s
s.
P
r
i
o
r
t
o
b
i
r
t
h
,
i
f
t
h
e s
cr
een
i
n
g
o
u
t
co
m
e r
ev
eal
s
t
h
e c
h
a
n
ce f
o
r
D
S
,
a f
o
r
m
al
d
i
a
g
n
o
s
i
s
can
b
e
es
t
ab
l
i
s
h
ed
.
F
o
r
ex
a
m
i
n
i
n
g
t
h
e
m
o
d
i
f
i
cat
i
o
n
s
i
n
s
p
eci
f
i
ed
an
at
o
m
i
ca
l
f
eat
u
r
es
o
f
t
h
e f
et
u
s
,
t
ech
n
i
q
u
es
l
i
k
e
ex
p
an
d
ed
al
p
h
a
-
f
e
to
p
r
o
te
in
(
A
F
P
)
s
c
r
e
e
n
in
g
te
s
t,
th
e
n
u
c
h
al
t
r
an
s
l
u
cen
c
y
t
es
t
,
an
d
s
u
p
p
l
e
m
en
t
ar
y
u
l
t
r
as
o
u
n
d
sc
r
een
s
w
h
i
c
h
ar
e av
ai
l
ab
l
e [
8
]
.
S
u
b
j
ect
s
w
er
e al
s
o
cl
ai
m
ed
t
h
at
t
h
e
n
o
s
e
s
o
f
ch
i
l
d
r
en
w
i
t
h
D
S
m
a
y
t
a
k
e i
n
a
n
und
e
r
-
de
v
e
l
ope
d n
a
s
a
l
bon
e
c
on
s
t
r
u
c
t
i
on
.
B
u
t
t
h
i
s
de
f
or
m
a
t
i
on
m
a
y
n
ot
be
i
de
n
t
i
f
i
e
d du
r
i
n
g
t
h
e ear
l
y
p
r
eg
n
an
c
y
s
ca
n
[
9
]
.
T
h
es
e s
u
r
v
e
y
s
co
n
f
i
r
m
t
h
a
t
t
h
e s
ci
e
n
t
i
f
i
c r
ef
l
ect
i
o
n
o
f
f
et
al
n
a
s
al
b
o
n
e co
n
t
r
i
b
u
t
es
t
o
t
h
e
p
er
f
o
r
m
a
n
ce o
f
t
h
e f
i
r
s
t
-
t
ri
m
e
s
t
e
r s
c
re
e
n
i
n
g
f
o
r D
S
[1
0
].
In
m
a
c
hi
ne
l
ea
r
ni
n
g
(
M
L
)
b
o
o
s
tin
g
is
a
n
a
t
tit
u
d
e
t
h
at
co
m
b
i
n
e
s
f
eeb
l
e a
n
d
er
r
o
n
eo
u
s
r
u
l
e
s
an
d
i
n
r
et
u
r
n
cr
ea
t
e a p
r
eci
s
e f
o
r
ecas
t
r
u
l
e.
T
h
e
m
o
s
t
w
i
d
el
y
u
s
ed
f
i
r
s
t
pr
a
c
t
i
c
a
l
boos
t
i
ng
a
l
g
or
i
t
hm
w
a
s
t
h
e
A
da
B
oos
t
a
l
g
or
i
t
hm
[
11]
.
A
da
boos
t
i
s
a
c
t
u
a
l
l
y a
m
a
c
h
i
n
e
-
l
ear
n
i
n
g
t
ech
n
i
q
u
e
f
o
r
f
ace d
et
ect
i
o
n
.
I
n
ad
d
i
t
i
o
n
,
t
h
e t
er
m
b
o
o
s
t
ed
r
ep
r
es
en
t
s
t
h
e cl
as
s
i
f
i
er
s
at
each
s
t
a
g
e,
w
h
i
c
h
i
s
b
u
ilt o
f
b
a
s
ic
c
la
s
s
if
ie
r
s
[
1
2
]
.
M
a
n
y
s
u
r
v
e
y
s
w
e
r
e
c
o
n
d
u
c
te
d
f
o
r
in
v
e
s
ti
g
a
ti
n
g
th
e
s
u
ita
b
i
lit
y
o
f
A
d
a
b
o
o
s
t in
f
a
c
e
r
e
c
og
n
i
t
i
on
a
n
d t
h
e
a
l
gor
i
t
hm
i
s
f
o
u
n
d t
o be
a
be
t
t
e
r
t
ool
[
1
3]
,
[
14]
.
I
n
th
is
a
lg
o
r
ith
m
,
t
h
e f
ace
r
es
t
r
u
ct
u
r
i
n
g
i
s
d
o
n
e b
y
m
o
v
i
n
g
t
h
e cas
cad
e d
et
ect
o
r
ac
r
o
s
s
t
h
e i
m
ag
e at
v
ar
i
o
u
s
s
cal
es
a
n
d
l
o
cat
i
o
n
s
.
W
h
en
co
m
p
ar
ed
w
i
t
h
t
h
e
n
u
m
b
er
o
f
r
eg
i
o
n
s
s
ca
n
n
ed
,
a d
i
s
t
i
n
ct
i
v
e i
m
a
g
e
h
as
o
n
l
y
a
m
o
d
es
t
t
u
r
n
o
f
f
ace p
ar
t
s
.
H
o
w
e
ve
r
,
i
n
no
n
-
f
ace r
eg
i
o
n
s
,
b
ecau
s
e o
f
t
h
e
u
n
t
i
m
el
y
t
er
m
i
n
at
i
o
n
o
f
t
h
e d
eci
s
i
o
n
p
r
o
ces
s
,
t
h
er
e ar
e ch
a
n
ce
s
t
o
g
et
o
n
l
y
a
s
m
al
l
n
u
m
b
er
o
f
s
t
a
g
es
o
f
t
h
e
cas
cad
e
t
h
at
a
r
e
es
t
i
m
at
ed
at
a
n
av
er
a
g
e
[
1
5
]
,
[
16
]
.
T
hus
,
t
h
e
r
es
t
r
ai
n
i
n
g
f
act
o
r
s
i
n
t
h
e s
p
ee
d
o
f
ev
al
u
at
i
o
n
ar
e t
h
e co
m
p
u
t
at
i
o
n
al
d
i
f
f
i
cu
l
t
y
a
n
d
t
h
e
d
e
n
i
a
l r
a
te
s
o
f
th
e
in
itia
l
p
o
in
ts
.
O
t
h
er
ch
al
l
en
g
es
t
h
at
m
a
y
co
n
f
r
o
n
t
d
u
r
i
n
g
t
h
e r
eo
r
g
an
i
zat
i
o
n
p
r
o
ces
s
ar
e t
h
e n
o
n
-
a
v
a
ila
b
ilit
y
o
f
f
in
e
i
m
a
g
es
,
t
h
e
g
ar
l
an
d
o
f
p
i
ct
u
r
e
s
an
d
o
t
h
er
i
m
ag
e q
u
a
l
i
t
y
m
at
t
er
s
.
S
o
,
f
o
r
g
en
er
at
i
n
g
m
o
r
e accu
r
at
e s
o
l
u
t
i
o
n
s
,
wh
a
t
i
s
r
e
qu
i
r
e
d i
s
a
n
e
f
f
i
c
i
e
n
t
da
t
a
s
e
t
.
I
n
t
h
i
s
w
or
k
,
a
n
e
nh
a
n
c
e
d A
da
B
oos
t
a
l
g
or
i
t
hm
i
s
pr
opos
e
d f
or
f
i
n
di
ng
th
e
d
i
f
f
e
r
e
n
t
m
a
g
n
i
tu
d
e
o
f
D
S
.
V
a
r
io
u
s
c
o
m
b
in
a
tio
n
s
o
f
in
p
u
t p
a
r
a
m
e
te
r
s
a
r
e
s
u
g
g
e
s
te
d
.
I
t is
n
o
te
d
th
a
t t
h
e
pr
op
os
e
d a
l
g
or
i
t
hm
y
i
e
l
ds
t
h
e
f
i
n
e
s
t
r
e
s
u
l
t
s
b
y
u
s
i
n
g
t
h
e
m
i
l
i
t
ar
y
cap
ab
i
l
i
t
i
es
o
f
t
h
e s
u
p
p
o
r
t
ed
p
ar
am
et
er
s
.
D
S
i
s
r
eco
g
n
i
zed
v
i
s
u
al
l
y
b
y
s
eei
n
g
a
s
t
h
e
u
l
t
r
as
o
n
o
g
r
a
m
i
m
a
g
e
o
f
t
h
e
f
et
u
s
f
o
r
t
h
e
o
ccu
r
r
en
ce o
f
n
a
s
al
b
o
n
e
(
N
B
)
.
O
p
t
i
cal
i
d
en
t
i
f
i
cat
i
o
n
o
f
N
B
m
a
y
co
n
f
u
s
e t
h
e
s
o
n
o
g
r
ap
h
er
s
an
d
l
ead
s
t
o
er
r
o
r
b
ecau
s
e o
f
t
h
e
m
o
d
es
t
s
iz
e
o
f
N
B
in
th
e
f
ir
s
t tr
i
m
e
s
t
e
r
.
E
q
u
a
ll
y
,
it is
v
i
s
u
a
liz
e
d
in
th
e
lite
r
a
tu
r
e
; t
h
e
d
e
f
ic
ie
n
c
y
o
f
n
a
s
a
l b
o
n
e
is
a
f
u
n
d
a
m
en
t
a
l
i
n
d
ex
t
o
d
et
ect
D
S
at
t
h
e ear
l
y
p
h
as
e o
f
p
r
eg
n
an
c
y
.
S
p
eck
l
e
n
o
i
s
e er
o
d
es
t
h
e ex
cel
l
e
n
ce o
f
t
h
e
ul
t
r
a
s
o
u
nd
i
m
a
g
es
.
T
h
er
ef
o
r
e,
t
h
e r
eco
g
n
i
t
i
o
n
o
f
N
B
b
eco
m
e
s
v
er
y
co
m
p
l
e
x
i
n
t
h
e f
i
r
s
t
t
r
i
m
es
t
er
[
1
7
]
.
F
i
g
u
r
e 1
s
h
o
w
s
t
h
e F
aci
al
d
i
s
t
i
n
ct
i
v
en
e
s
s
o
f
a
f
et
u
s
u
s
ed
t
o
an
al
y
ze D
S
.
Fi
g
ur
e
1
.
F
a
c
ia
l d
is
tin
c
tiv
e
n
e
s
s
o
f
a f
et
u
s
u
s
ed
t
o
an
al
y
ze d
ow
n s
y
n
dr
o
m
e
T
h
i
s
p
ap
er
p
r
o
p
o
s
es
an
al
g
o
r
i
t
h
m
t
o
d
et
ect
t
h
e c
h
a
n
ce
f
o
r
D
S
t
h
r
o
u
g
h
t
h
e
an
al
y
s
i
s
o
f
f
et
u
s
f
ac
i
a
l
ch
ar
act
er
s
.
L
o
cal
i
zat
i
o
n
o
f
f
a
ci
al
p
o
i
n
t
s
i
s
d
o
n
e a
n
d
t
h
e i
n
t
er
can
t
h
al
d
i
s
t
a
n
ce i
s
cal
c
u
l
at
ed
.
Cl
a
ss
i
f
i
e
d
t
h
e
c
ol
l
e
c
t
e
d i
nf
or
m
a
t
i
on i
nt
o t
w
o
ba
s
e
d on
t
h
e
a
g
e
.
U
s
i
n
g
t
h
e
s
e
cal
cu
l
at
i
o
n
s
,
t
h
e p
r
o
p
o
s
ed
al
g
o
r
i
t
h
m
d
et
ect
s
t
h
e
ch
an
ce
f
o
r
D
S
.
T
h
e r
e
m
ai
n
i
n
g
p
ar
t
s
o
f
t
h
e co
m
p
o
s
i
t
i
o
n
ar
e a
r
r
an
g
ed
as
f
o
l
l
o
w
s
.
P
ar
t
I
I
p
r
e
s
en
t
s
t
h
e
m
at
er
i
al
s
an
d
m
et
h
o
d
s
,
s
ect
i
o
n
3
i
s
t
he
r
e
s
ul
t
a
nd
d
i
s
c
u
s
s
i
o
ns
a
nd
f
i
na
l
l
y
,
s
e
c
t
i
o
n
4
c
o
nc
l
ud
e
s t
h
e
st
u
d
y
.
2.
M
AT
E
RI
AL
S
AND M
E
T
H
O
DS
2.
1.
A
da
pt
i
v
e
bo
o
s
t
i
ng
(
A
da
B
oos
t
)
al
gor
i
t
h
m
A
n
Ad
a
B
o
o
s
t a
lg
o
r
ith
m
is
a
t
o
o
l to
u
p
g
r
a
d
e
th
e
a
c
c
u
r
a
c
y
o
f
w
e
a
k
le
a
r
n
i
n
g
a
lg
o
r
ith
m
s
.
I
t is
a
M
L
a
lg
o
r
ith
m
in
tr
o
d
u
c
e
d
b
y
F
r
e
un
d a
n
d
S
c
h
a
pi
r
e
[
17]
.
I
t
c
a
n be
i
m
pl
e
m
e
n
t
e
d b
y
r
unn
i
n
g t
h
e
w
e
a
k
l
e
a
r
n
i
ng
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t
J
E
l
e
c
&
C
o
m
p
E
n
g
I
SSN
:
2088
-
8708
D
ow
n s
y
ndr
om
e
de
t
e
c
t
i
on us
i
ng m
odi
f
i
e
d
ada
boos
t
al
g
or
i
t
hm
(
V
i
nc
y
D
e
v
i
V
.
K
.
)
4283
a
lg
o
r
ith
m
m
u
ltip
le
ti
m
e
s
,
s
u
c
h
t
h
a
t t
h
e
d
a
ta
is
s
lig
h
tl
y
m
o
d
if
ie
d
i
n
e
a
c
h
r
u
n
a
n
d
th
e
n
c
o
m
b
i
n
e
th
e
p
r
ep
o
s
i
t
i
o
n
s
.
I
t
can
b
e n
a
m
e
d
as
a m
e
t
a
-
a
lg
o
r
it
h
m
s
i
n
c
e
i
t is
u
ti
liz
e
d
to
in
c
r
e
a
s
e
t
he
f
unc
t
i
o
ni
ng o
f
o
t
he
r
s
ch
o
l
ar
s
.
I
n
g
en
er
al
,
t
h
er
e
ar
e t
w
o
t
y
p
e
s
o
f
cl
as
s
i
f
i
er
s
i
n
l
ear
n
i
n
g
al
g
o
r
i
t
h
m
s
;
s
t
r
o
n
g
an
d
w
ea
k
cl
a
s
s
i
f
i
er
s
.
A
s
tr
o
n
g
c
la
s
s
if
ic
a
tio
n
a
l
g
o
r
ith
m
u
s
e
s
m
e
t
h
ods
lik
e
ar
t
i
f
i
ci
al
n
eu
r
al
n
et
w
o
r
k
(
A
NN)
a
n
d
s
u
ppor
t
v
e
c
t
or
m
ach
i
n
e
(S
V
M
).
H
o
w
e
v
er
,
a
w
eak
cl
a
s
s
i
f
i
cat
i
o
n
al
g
o
r
i
t
h
m
e
m
p
l
o
y
s
m
et
h
o
d
s
s
u
c
h
as
a
b
a
y
e
s
i
a
n ne
t
w
o
r
k
,
a
nd
d
e
c
is
io
n
t
r
ee.
A
d
a
B
o
o
st
c
l
a
s
si
f
i
e
r
i
s
sh
o
w
n
i
n
F
i
g
u
r
e
2
.
T
h
e
Ad
a
B
o
o
s
t a
lg
o
r
ith
m
i
s
a
d
a
p
tiv
e
in
n
a
tu
r
e
.
T
h
i
s
i
s
b
ecau
s
e i
f
t
h
e p
r
ev
i
o
u
s
cl
as
s
i
f
i
er
h
a
s
m
i
s
cl
as
s
i
f
i
ed
an
y
i
n
s
t
an
ce
s
,
an
d
s
o
t
h
o
s
e i
n
s
t
a
n
ces
w
er
e
r
e
s
tr
u
c
t
u
r
e
d
in
to
th
e
s
u
c
c
e
e
d
in
g
c
la
s
s
i
f
ie
r
.
T
h
is
a
lg
o
r
ith
m
is
a
ls
o
r
e
a
ll
y
s
e
n
s
it
iv
e
t
o
d
is
tu
r
b
a
n
c
e
.
T
h
is
al
g
o
r
i
t
h
m
s
t
ar
t
s
b
y
al
l
o
cat
i
n
g
t
h
e
s
a
m
e
w
e
i
g
h
t
s
t
o
al
l
o
ccu
r
r
en
ces
i
n
t
h
e t
r
ai
n
i
n
g
d
a
t
a.
T
h
en
i
t
f
o
r
m
s
a
cl
as
s
i
f
i
er
u
s
i
n
g
t
h
e l
ear
n
i
n
g
al
g
o
r
i
t
h
m
f
o
r
t
h
e d
at
a a
n
d
t
h
e
w
ei
g
h
t
s
w
i
l
l
g
et
u
p
d
at
ed
b
as
e
d
o
n
t
h
e cl
as
s
i
f
i
er
’
s
r
e
s
p
o
ns
e
.
T
he
e
a
s
y
i
n
s
t
a
nc
e
s
ge
t
l
o
w
w
e
i
gh
t
s
a
nd
hi
gh
w
e
i
ght
s
a
r
e
gi
ve
n
f
o
r
ha
r
d
c
a
s
e
s
.
F
o
r
t
he
w
e
i
g
ht
e
d
d
at
a,
a
cl
as
s
i
f
i
er
i
s
a
s
s
i
g
n
ed
i
n
t
h
e
f
o
l
l
o
w
i
n
g
ite
r
a
tio
n
.
A
l
l
th
e
s
e
s
te
p
s
a
i
m
f
o
r
a
b
e
tte
r
c
la
s
s
i
f
ic
a
tio
n
.
T
h
e
i
n
s
t
a
n
ce
w
ei
g
h
t
s
ar
e r
es
t
r
u
ct
u
r
ed
b
as
ed
o
n
t
h
e r
eact
i
o
n
o
f
t
h
e n
e
w
cl
as
s
i
f
i
er
.
T
h
er
e ar
e t
w
o
p
r
o
b
ab
i
l
i
t
i
es
,
ei
t
h
er
t
h
e ea
s
i
er
i
n
s
t
a
n
ce,
m
a
y
b
eco
m
e
eas
i
er
o
r
b
eco
m
e
h
ar
d
er
.
T
h
e s
eco
n
d
i
s
,
t
h
e
h
ar
d
er
i
n
s
t
an
ce
b
eco
m
e
s
h
ar
d
er
o
r
eas
i
er
.
T
h
e
y
ar
e r
en
o
r
m
al
i
zed
a
f
t
er
t
h
e
w
ei
g
h
t
u
p
d
at
es
.
A
s
t
h
e l
a
s
t
s
t
a
i
r
,
t
h
e as
s
u
m
ed
v
al
u
e i
s
es
t
i
m
at
ed
[
1
8
]
,
[
19]
.
Fi
g
ur
e
2.
A
da
B
oos
t
c
la
s
s
if
ie
r
A
l
g
o
r
ith
m
a)
St
a
r
t
b)
F
o
r
m
tr
a
in
i
n
g
s
a
m
p
le
s
e
ts
,
i
n
it
ia
liz
e
s
a
m
p
le
w
e
i
g
ht
s
,
a
nd
i
t
e
r
a
t
i
o
ns
.
T
he
w
e
i
g
ht
m
w
ji
1
=
; f
o
r
j
=
1
t
o
N
; N
is
th
e
n
u
m
b
e
r
o
f
ite
r
a
tio
n
s
.
(1
)
c)
C
al
cu
l
at
e
t
h
e n
o
r
m
al
i
zed
w
ei
g
h
t
s
∑
=
m
i
j
i
w
i
w
i
p
)
(
)
(
)
(
(2
)
d)
C
al
cu
l
at
e t
h
e
h
y
p
o
t
h
es
i
s
(
j
h
)
v
al
u
e u
s
i
n
g
N
aï
v
e B
a
y
es
p
r
o
b
ab
i
l
i
t
y
e)
C
al
cu
l
at
e t
h
e er
r
o
r
.
)]
(
)[
(
i
i
j
j
i
j
y
x
h
i
p
e
≠
=
∑
(3
)
f)
I
f e
g
>
1
/2
,
; N
=
j
-
1;
g
o t
o s
t
e
p 9
j
j
j
e
e
−
=
1
φ
(4
)
g)
C
al
cu
l
at
e t
h
e
w
ei
g
h
t
co
ef
f
i
ci
e
n
t
,
h)
C
o
m
p
u
t
e
ne
w
w
e
i
g
ht
s
us
i
n
g t
he
c
o
e
f
f
i
c
i
e
nt
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN
:
20
88
-
8708
In
t
J
E
l
e
c
&
C
o
m
p
E
n
g
,
V
o
l.
11
, N
o
.
5
,
O
c
t
obe
r
20
2
1
:
4
2
8
1
-
4
2
8
8
4284
)]
(
[
1
1
)
(
)
(
i
i
j
y
x
h
j
j
j
i
w
i
w
−
−
+
=
φ
(5
)
i)
F
i
n
a
l
hy
pot
h
e
s
i
s
)]
(
[
1
l
og
)
(
1
y
x
h
x
h
j
L
j
h
o
=
=
∑
=
φ
(6
)
j)
S
t
op.
2
.2
.
N
aï
ve
b
a
y
e
’
s c
l
a
s
si
f
i
e
r
In
m
a
c
hi
ne
l
e
a
r
ni
ng,
a p
r
o
b
ab
i
l
i
s
t
i
c cl
as
s
i
f
i
er
b
as
ed
o
n
b
a
ye
’
s
[
2
0
]
t
h
eo
r
em
i
s
t
h
e n
aï
v
e
b
a
ye
s
cl
as
s
i
f
i
er
s
.
T
h
e p
r
o
v
i
s
i
o
n
al
an
d
m
i
n
i
m
al
p
r
o
b
ab
i
l
i
t
i
es
o
f
t
w
o
ar
b
i
t
r
ar
y
ev
en
t
s
ar
e r
el
at
ed
u
s
i
n
g
b
ay
es
t
h
e
o
r
em
.
G
en
er
al
l
y
,
i
t
i
s
u
s
ed
t
o
cal
cu
l
at
e t
h
e s
u
b
s
eq
u
en
t
p
r
o
b
ab
i
l
i
t
i
es
.
T
h
es
e cl
as
s
i
f
i
er
s
ar
e i
n
d
ep
en
d
en
t
f
ea
t
u
r
e
m
ode
l
s
an
d
l
ear
n
i
n
g
s
u
c
h
cl
as
s
i
f
i
er
s
i
s
s
i
m
p
l
i
f
i
ed
i
f
t
h
e i
n
d
i
v
i
d
u
al
p
ar
a
m
et
er
s
ar
e
co
n
s
i
d
er
ed
.
W
h
en
co
m
p
ar
ed
w
i
t
h
o
t
h
er
t
ech
n
i
q
u
es
,
r
eg
ar
d
l
es
s
o
f
t
h
i
s
n
a
i
v
e
h
y
p
o
t
h
es
i
s
,
t
h
i
s
cl
as
s
i
f
i
er
i
s
r
e
m
ar
k
ab
l
y
e
f
f
ect
i
v
e i
n
p
r
act
i
ce,
i
n
cl
u
d
i
n
g
f
o
r
ap
p
l
i
cat
i
o
n
s
l
i
k
e a
m
ed
i
cal
d
i
ag
n
o
s
i
s
,
t
ex
t
cl
as
s
i
f
i
cat
i
o
n
,
an
d
s
y
s
t
e
m
s
p
er
f
o
r
m
a
n
ce
a
d
m
in
i
s
tr
a
tio
n
.
T
h
e
c
o
n
d
itio
n
a
l p
r
o
b
a
b
ilit
y
is
p
r
e
s
e
n
te
d
e
q
u
a
ll
y
,
∑
=
'
'
)
(
)
(
)
(
)
(
)
(
'
k
k
k
k
k
k
C
P
C
x
P
C
P
C
x
P
x
C
P
(7
)
T
he
na
ï
ve
b
a
y
e
s c
l
a
s
s
i
f
i
e
r
s
es
t
i
m
at
e t
h
at
t
h
e
v
al
u
e o
f
t
h
e s
p
eci
f
i
c
f
eat
u
r
e o
f
a cl
as
s
i
s
i
s
o
l
at
ed
f
r
o
m
t
he
va
l
ue
o
f
a
n
y
o
t
h
er
f
eat
u
r
e.
S
o
,
)
(
)
(
1
k
d
j
j
k
C
x
P
C
x
P
∏
=
=
(8
)
w
h
e
r
e
,
C
=
{
C
1
,
C
2
,
.
.
.
.
.
C
K
}
is
th
e
s
e
t o
f
c
la
s
s
la
b
e
ls
a
n
d
P
(
x
|
C
)
is
c
o
n
d
itio
n
a
l p
r
o
b
a
b
ility
.
2
.3
.
B
ac
k
p
r
op
agat
i
on
n
eu
ra
l
n
et
w
o
rk
(
B
P
NN)
A
n
e
u
r
al
n
et
w
o
r
k
[
2
1
]
i
s
d
ef
i
n
ed
as
a co
l
l
ect
i
o
n
o
f
as
s
o
ci
at
ed
i
n
p
u
t
-
o
u
tp
u
t e
le
m
e
n
ts
s
u
c
h
t
h
at
eac
h
c
on
n
e
c
t
i
o
n
h
a
s
a
w
e
i
gh
t
l
i
nk
e
d
w
i
t
h i
t
.
I
t
i
s
a
ppl
i
e
d t
o
m
a
ke
pr
e
di
c
t
i
v
e
m
ode
l
s
f
r
o
m
bi
g
da
t
a
ba
s
e
s
a
n
d
t
h
e
y
m
i
m
i
c t
h
e
h
u
m
a
n
n
eu
r
al
s
y
s
t
e
m
.
F
o
r
t
r
ai
n
i
n
g
f
eat
u
r
e
s
et
s
ex
t
r
act
ed
f
r
o
m
t
h
e
cl
as
s
i
f
i
er
,
m
u
l
t
i
-
l
a
y
e
r
B
P
NNs
ar
e u
s
ed
.
C
o
m
m
o
n
l
y
,
a
m
u
l
t
i
-
la
y
er
n
et
w
o
r
k
co
n
s
i
s
t
s
o
f
t
h
r
ee l
a
y
er
s
o
f
n
er
v
e cel
l
s
.
T
h
e
y
a
r
e t
h
e i
n
p
u
t
l
a
y
er
,
a
hi
d
d
e
n l
a
ye
r
,
a
nd
o
ut
p
ut
l
a
y
e
r
.
T
he
c
o
nne
c
t
i
o
ns
i
n
B
P
N
N
a
r
e
s
uc
h
t
ha
t
,
e
a
c
h
no
d
e
i
n t
he
i
np
u
t
l
a
ye
r
i
s
co
n
n
ect
ed
t
o
t
h
e n
o
d
es
o
f
t
h
e h
i
d
d
en
l
a
y
er
.
A
l
so
,
h
i
d
d
en
l
a
y
er
n
o
d
es
ar
e
r
e
la
te
d
to
th
e
o
u
tp
u
t la
y
e
r
.
T
h
e
r
e
is
no
d
i
r
e
c
t
l
i
n
k b
e
t
w
e
e
n t
he
n
o
d
e
s
i
n a
n i
nd
i
vi
d
ua
l
l
a
ye
r
a
nd
w
i
t
h e
ve
r
y c
o
n
ne
c
t
i
o
n;
t
he
r
e
s
ho
ul
d
b
e
a
n
as
s
o
ci
at
ed
w
ei
g
h
t
i
n
g
f
act
o
r
[
2
2
]
.
A
t
t
h
e
t
i
m
e o
f
t
r
ai
n
i
n
g
,
t
h
e
w
ei
g
h
t
s
ar
e
m
o
d
i
f
i
ed
u
s
i
n
g
t
h
e B
P
al
g
o
r
i
t
h
m
a
n
d
t
he
p
r
o
ces
s
i
s
cal
l
ed
l
ear
n
i
n
g
.
I
n
t
h
i
s
s
t
u
d
y
,
t
h
e s
u
p
er
v
i
s
ed
l
e
ar
n
i
n
g
m
et
h
o
d
i
s
ad
o
p
t
ed
.
T
h
e
A
N
N
t
r
ai
n
i
n
g
p
ar
a
m
et
er
s
ar
e
p
r
es
en
t
ed
i
n
T
ab
l
e
1
.
T
h
e
d
i
f
f
er
en
ce
b
et
w
een
t
ar
g
et
an
d
act
u
al
r
es
p
o
n
s
e i
s
t
er
m
ed
as
t
h
e
s
y
s
t
e
m
er
r
o
r
.
D
u
r
i
n
g
t
h
e
t
r
ai
n
i
n
g
p
r
o
ces
s
,
t
h
i
s
e
r
r
or
i
s
ba
c
k
pr
opa
g
a
t
e
d t
o t
h
e
p
r
eced
i
n
g
l
a
y
er
,
i
.
e.
t
h
e
h
i
d
d
en
l
a
y
er
.
T
h
e
er
r
o
r
i
s
t
h
en
cal
cu
l
at
ed
f
o
r
each
el
e
m
e
n
t
i
n
t
h
i
s
l
a
y
er
N
.
I
n
t
h
e
s
a
m
e
m
a
n
n
er
,
er
r
o
r
at
each
n
o
d
e o
f
a p
r
ev
i
o
u
s
l
y
h
i
d
d
en
l
a
y
er
t
h
at
i
s
N
-
1
is
a
ls
o
d
e
te
r
m
in
e
d
.
B
y
u
s
i
n
g
t
h
is
m
e
t
h
od t
o
r
ed
u
ce t
h
e er
r
o
r
t
o
t
h
e ex
p
ect
ed
l
ev
el
.
T
ab
l
e 1
.
A
N
N
tr
a
i
n
in
g
p
a
r
a
m
e
te
r
s
N
o
.
o
f
I
n
p
u
t
s
N
o.
of
o
u
t
p
u
t
s
N
o.
of
H
i
d
d
e
n
L
a
y
e
r
N
o.
of
H
i
d
d
e
n
N
e
u
r
on
s
N
o
.
o
f
it
e
r
a
tio
n
s
7
1
1
8
5
0
0
L
e
ar
n
i
n
g
R
at
e
M
o
m
en
t
u
m
R
at
e
A
c
t
iv
a
tio
n
F
u
n
c
t
io
n
T
r
a
i
n
i
n
g
M
e
t
h
o
d
E
rro
r
0
.
0
5
0
.5
S
i
gm
o
i
d
B
a
c
k
p
r
op
a
g
a
t
i
o
n
0
.
0
0
1
2.
4.
P
r
op
os
e
d
A
d
aB
oos
t
al
g
or
i
t
hm
T
h
e
Ad
a
-
B
oos
t
a
lg
o
r
ith
m
h
e
r
e u
s
ed
e
m
p
l
o
y
es
t
h
e
na
ï
ve
-
b
a
ye
s
cl
as
s
i
f
i
er
as
a
w
ea
k
l
e
ar
n
er
.
T
h
e
i
n
i
t
i
al
i
zat
i
o
n
o
f
t
r
ai
n
i
n
g
d
at
a a
n
d
w
ei
g
h
t
s
ar
e b
as
ed
o
n
t
h
e c
l
as
s
i
f
i
er
.
T
h
e
w
ei
g
h
t
s
ar
e eq
u
a
l
t
o
t
h
e
f
r
act
i
o
n
o
f
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t
J
E
l
e
c
&
C
o
m
p
E
n
g
I
SSN
:
2088
-
8708
D
ow
n s
y
ndr
om
e
de
t
e
c
t
i
on us
i
ng m
odi
f
i
e
d
ada
boos
t
al
g
or
i
t
hm
(
V
i
nc
y
D
e
v
i
V
.
K
.
)
4285
t
h
e t
o
t
al
n
u
m
b
er
o
f
cl
as
s
i
f
i
er
s
.
U
s
i
n
g
t
h
e s
u
p
er
v
i
s
ed
l
ear
n
i
n
g
m
et
h
o
d
,
t
h
es
e cl
a
s
s
i
f
i
er
s
can
b
e ef
f
ect
i
v
el
y
d
ev
el
o
p
ed
.
A
n
ad
v
an
t
a
g
e o
f
t
h
e
na
i
ve
b
a
y
e
s
c
la
s
s
i
f
ie
r
is
th
a
t it r
e
q
u
ir
e
s
o
n
l
y
a
li
ttle
a
m
o
u
n
t o
f
tr
a
i
n
i
n
g
i
nf
or
m
a
t
i
on f
or
as
s
es
s
i
n
g
t
h
e
co
m
p
o
n
en
t
s
n
eed
ed
f
o
r
s
o
r
t
i
n
g
.
A
f
t
er
each
i
t
er
at
i
o
n
n
e
w
h
y
p
o
t
h
es
i
s
h
j
i
s
ge
ne
r
a
t
e
d
.
A
p
p
l
y
i
ng t
hi
s
h
y
p
o
t
he
s
i
s
t
he
w
e
i
g
ht
an
d
er
r
o
r
a
r
e cal
cu
l
at
ed
an
d
t
h
e s
y
s
t
e
m
u
p
d
at
es
t
h
e
w
ei
g
h
t
s
b
as
ed
o
n
t
h
e er
r
o
r
v
al
u
e.
)]
(
[
1
1
)
(
)
(
i
i
j
y
x
h
j
j
j
i
w
i
w
−
−
+
=
β
(9
)
T
h
i
s
p
r
o
ced
u
r
e
i
s
r
ep
ea
t
ed
t
i
l
l
t
h
e i
t
er
a
t
i
o
n
s
a
r
e f
i
n
i
s
h
ed
a
n
d
t
h
e f
i
n
a
l
v
a
l
u
e i
s
co
m
p
u
t
ed
u
t
i
l
i
zi
n
g
t
h
e
r
u
l
e,
)]
(
[
1
l
og
)
(
1
y
x
h
x
h
j
L
j
h
o
=
=
∑
=
φ
(
10)
2.
5
.
F
ea
t
u
re
ex
t
ra
c
t
i
o
n
I
n
t
h
i
s
p
r
o
p
o
s
ed
m
et
h
o
d
,
t
h
e
f
aci
al
l
a
n
d
m
ar
k
s
ar
e d
et
ect
ed
an
d
t
h
e
n
a
g
eo
m
et
r
i
c d
e
s
cr
i
p
t
o
r
i
s
b
u
i
l
t
u
s
i
n
g
i
t
.
T
h
e g
eo
m
et
r
i
c r
ep
r
es
en
t
at
i
o
n
i
s
s
u
ch
t
h
at
,
i
t
u
s
es
2
-
D
f
a
c
ia
l f
id
u
c
ia
l p
o
in
ts
.
T
h
a
t is
,
t
w
o
p
o
in
ts
P
1
an
d
P
2
ar
e ch
o
s
en
f
r
o
m
t
h
e ce
n
t
er
of
t
h
e
e
y
e
br
o
w
s
,
f
or
t
h
e
gl
a
be
l
l
a
,
on
e
poi
n
t
-
P
3,
f
ou
r
poi
n
t
s
P
4,
P
5,
P
7,
a
n
d
P
8 f
or
t
h
e
i
nn
e
r
a
n
d ou
t
e
r
c
or
n
e
r
of
t
h
e
e
y
e
s
,
on
e
poi
n
t
P
6 f
or
t
h
e
r
oot
of
t
h
e
n
os
e
,
t
w
o p
oi
n
t
s
P
10 a
n
d P
1
1,
f
or
A
l
a
r
s
s
i
de
w
a
l
l
s
,
on
e
poi
n
t
P
9 f
or
t
h
e
s
u
pr
a
t
i
p,
on
e
poi
n
t
P
12
f
or
t
h
e
c
ol
um
e
l
l
a
,
t
h
e
poi
n
t
s
P
14 a
n
d P
1
5,
f
or
t
h
e
m
ou
t
h
c
or
n
e
r
s
a
n
d t
w
o
poi
n
t
s
P
13 a
n
d P
16 f
or
t
h
e
t
op a
n
d bot
t
om
of
t
h
e
u
ppe
r
l
i
p a
n
d l
o
w
e
r
l
i
p
r
es
p
ect
i
v
el
y
.
F
i
g
u
r
e
3
s
h
o
w
s
h
o
w
t
h
e
lo
c
a
liz
a
tio
n
o
f
t
h
e
1
6
f
a
c
ia
l p
o
in
ts
is
d
o
n
e
.
F
r
o
m
t
h
e
m
e
n
t
i
o
n
e
d 16
poi
n
t
s
,
f
o
u
r
t
een
d
i
s
t
a
n
ces
ar
e i
d
en
t
i
f
i
ed
an
d
t
ak
en
o
u
t
as
s
h
o
w
n
i
n
F
i
g
ur
e
3.
A
l
l
t
h
es
e
d
i
s
t
an
ce
s
ar
e
n
o
r
m
al
i
zed
t
o
t
h
e
w
i
d
t
h
o
f
t
h
e
f
ac
e
t
o
g
u
ar
an
t
ee
t
h
at
t
h
e
ch
ar
a
ct
er
i
s
t
i
cs
ar
e
s
cal
e
-
i
n
v
a
r
i
a
n
t
T
h
e
di
s
t
a
n
c
e
s
,
s
a
y
,
d
1,
d2,
d3,
d6,
d8,
d10
,
d1
1,
d
12,
a
n
d
d14
s
y
m
bol
i
ze t
h
e s
t
a
n
d
ar
d
v
al
u
es
o
f
t
h
e t
w
o
m
i
r
r
o
r
ed
d
i
s
t
an
ces
o
n
b
o
t
h
s
i
d
es
o
f
t
h
e
f
ace.
T
o
cal
cu
l
at
e t
h
e d
i
s
t
a
n
ce d
1
1
,
t
h
e m
ee
t
i
n
g
j
u
n
ct
i
o
n
s
o
f
t
h
e l
i
n
e
a
m
o
n
g
th
e
p
o
in
ts
o
n
t
h
e
to
p
o
f
th
e
u
p
p
e
r
lip
a
n
d
b
o
tto
m
o
f
th
e
lo
w
e
r
lip
,
a
n
d
th
e
lin
e
b
e
t
w
e
e
n
th
e
le
f
t a
n
d
ri
ght
c
o
r
ne
r
s
o
f
t
he
m
o
u
t
h a
r
e
us
e
d
.
T
hus
,
t
he
f
o
l
l
o
w
i
n
g c
ha
r
a
c
t
e
r
i
s
t
i
c
ve
c
t
o
r
ha
s
f
o
ur
t
e
e
n d
i
m
e
ns
i
o
ns
.
F
o
r
r
eco
g
n
i
zi
n
g
D
S
,
i
t
i
s
s
i
g
n
i
f
i
can
t
t
o
co
n
s
i
d
er
t
h
e i
n
t
er
can
t
h
a
l
s
p
ace d
4
,
i
.
e.
,
s
p
ace
a
m
o
n
g
t
h
e
i
n
s
i
d
e r
eces
s
es
o
f
t
h
e e
y
es
.
A
n
ad
d
i
t
i
o
n
i
n
t
h
e i
n
t
er
ca
n
t
h
al
d
i
s
t
an
ce cal
l
ed
t
el
eca
n
t
h
u
s
i
s
f
r
eq
u
en
t
l
y
o
b
s
er
v
ed
i
n
t
h
e p
eo
p
l
e
w
i
t
h
D
S
.
T
h
e p
r
es
en
t
s
o
f
t
i
n
y
p
al
p
eb
r
al
f
i
s
s
u
r
es
,
w
h
i
c
h
i
s
t
h
e d
i
s
t
a
n
ce b
et
w
een
t
h
e l
at
er
al
a
n
d
m
ed
i
al
can
t
h
u
s
o
f
t
h
e o
p
t
i
cs
i
s
an
o
t
h
er
co
m
m
o
n
f
eat
u
r
e.
T
h
i
s
as
p
ect
i
s
cap
t
u
r
ed
b
y
s
p
ace
d3.
P
e
opl
e
w
i
t
h
D
S
co
m
m
o
n
l
y
i
n
d
i
cat
e t
h
e s
y
m
p
t
o
m
o
f
co
m
p
r
es
s
ed
n
o
s
e an
d
d
i
s
t
an
ce
s
d
5
t
o
d
8
s
t
an
d
f
o
r
t
h
i
s
p
ar
t
i
cu
l
ar
f
eat
u
r
e.
L
a
s
t
l
y
,
t
h
e
pr
e
s
e
n
c
e
of
a
s
m
a
l
l
m
ou
t
h
,
w
hi
c
h
i
s
a
s
y
m
pt
o
m
o
f
D
S
,
i
s
c
a
pt
u
r
e
d by
di
s
t
a
n
c
e
s
d10 t
o
d13
.
Fi
g
ur
e
3
.
L
o
c
a
liz
a
tio
n
o
f
f
a
c
ia
l p
o
in
ts
3.
R
ES
U
LT
S
A
ND
D
I
SC
U
SS
I
O
N
F
id
u
c
ia
l p
o
in
ts
ar
e
p
o
in
ts
t
h
at
ar
e u
s
ed
as
p
o
in
ts
o
f
r
ef
er
en
ce o
r
m
eas
u
r
e;
h
en
ce d
e
t
er
m
i
n
i
n
g
t
he
f
id
u
c
ia
l p
o
in
ts
is
th
e
e
le
m
e
n
ta
r
y
s
te
p
to
d
is
ti
n
g
u
is
h
a
f
a
ce.
S
o
m
e o
f
t
h
e
m
ai
n
f
id
u
c
ia
l p
o
in
ts
ar
e t
h
e e
y
es
,
l
i
p
ed
g
es
,
n
o
s
e,
a
nd
c
hi
n.
I
n t
hi
s
s
t
ud
y
,
s
o
m
e
i
m
a
g
es
f
r
o
m
t
h
e D
ar
t
m
o
u
t
h
D
at
ab
as
e o
f
C
h
i
l
d
r
en
’
s
F
ace
s
a
r
e
s
el
ect
ed
,
i
n
cl
u
s
i
v
e o
f
m
al
es
an
d
f
e
m
al
e i
m
a
g
es
t
h
at
w
er
e s
h
o
w
i
n
g
v
er
y
d
i
s
t
i
n
ct
f
aci
al
ex
p
r
es
s
i
o
n
s
.
S
o
m
e
i
m
a
g
es
co
l
l
ect
ed
b
y
F
er
r
y
et
.
al
w
i
t
h
D
S
ar
e al
s
o
ad
m
i
t
t
ed
.
T
h
e i
m
ag
e
s
w
er
e act
u
al
l
y
ca
p
t
u
r
ed
i
m
p
ul
s
i
ve
l
y
f
r
o
m
u
n
r
el
at
ed
an
g
l
es
,
b
ack
g
r
o
u
n
d
s
,
an
d
d
i
s
t
a
n
ces
,
s
cr
een
i
n
g
u
n
r
el
i
ab
l
e
h
ead
p
o
s
es
,
f
aci
al
ex
p
r
es
s
i
o
n
s
,
an
d
o
ccl
u
s
i
o
n
s
.
A
s
t
h
e
n
ex
t
s
t
ep
,
t
h
e i
m
ag
e i
s
t
es
t
ed
an
d
v
er
i
f
i
ed
an
d
t
h
en
co
m
p
ar
ed
t
o
t
h
e
s
o
l
u
t
i
o
n
w
i
t
h
o
t
h
er
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN
:
20
88
-
8708
In
t
J
E
l
e
c
&
C
o
m
p
E
n
g
,
V
o
l.
11
, N
o
.
5
,
O
c
t
obe
r
20
2
1
:
4
2
8
1
-
4
2
8
8
4286
p
r
o
m
i
n
en
t
cl
as
s
i
f
i
er
s
.
T
h
e p
r
o
pos
e
d m
e
t
h
odol
ogy
m
a
i
n
l
y
c
o
m
pos
e
d of
f
o
u
r
s
t
a
g
e
s
:
i
)
f
aci
al
d
et
ect
i
o
n
,
ii)
f
e
a
tu
r
e
s
e
x
tr
a
c
tio
n
,
iii)
f
eat
u
r
es
r
ed
u
ct
i
o
n
,
an
d
i
v
)
cl
as
s
i
f
i
cat
i
o
n
.
I
t
can
b
e
r
em
ar
k
ed
t
h
at
f
r
o
m
t
h
e
s
t
u
d
y
th
a
t a
s
th
e
n
u
m
b
e
r
o
f
ite
r
a
tio
n
s
in
c
r
e
a
s
e
d
,
th
e
r
e
is
a
d
i
m
in
u
t
io
n
in
th
e
e
r
r
o
r
v
a
lu
e
.
B
P
N
N
is
u
s
e
d
to
tr
a
in
th
e
c
la
s
s
i
f
ie
d
p
a
r
a
m
e
te
r
s
a
n
d
M
A
T
L
A
B
is
t
h
e
p
r
o
g
r
a
m
u
til
iz
e
d
f
o
r
tr
a
in
in
g
a
n
d
te
s
ti
n
g
.
F
ig
u
r
e
4
s
h
o
w
s
th
e
p
er
f
o
r
m
a
n
ce o
f
t
h
e
A
d
aB
o
o
s
t
al
g
o
r
i
t
h
m
a
nd
T
a
bl
e
2 de
p
i
c
t
s
t
h
e
e
v
a
l
u
a
t
i
on
of
v
a
r
i
ous
m
e
t
h
ods
f
or
D
S
id
e
n
ti
f
ic
a
tio
n
acco
r
d
i
n
g
t
o
co
r
r
ect
n
es
s
r
at
e
s
(
i
n
p
er
cen
t
)
.
F
o
l
l
o
w
i
n
g
,
w
e co
m
p
ar
e t
h
e a
n
t
i
ci
p
at
ed
r
es
p
o
n
s
e o
f
A
d
a
-
B
P
N
N
w
i
t
h o
t
he
r
p
r
e
s
e
nt
e
d
m
e
t
ho
d
s
.
T
he
g
i
v
e
n
s
y
s
t
e
m
h
as
e
n
h
an
ced
p
er
f
o
r
m
a
n
ce t
h
an
t
h
e l
at
er
m
o
d
el
.
T
h
e accu
r
acy
o
f
t
h
e p
l
an
n
ed
t
ec
h
n
i
q
u
e i
s
s
up
e
r
i
o
r
t
ha
n t
he
a
ddi
t
i
on
a
l
m
e
t
h
ods
.
T
h
i
s
i
s
t
h
e
m
a
i
n pu
r
pos
e
t
o
u
s
e
B
P
f
or
pa
r
a
m
e
t
e
r
opt
i
m
i
z
a
t
i
o
n
a
n
d a
s
a
w
ea
k
cl
as
s
i
f
i
er
;
t
h
i
s
s
y
s
t
e
m
i
s
h
y
b
r
i
d
w
i
t
h
A
d
aB
o
o
s
t
t
ech
n
o
l
o
g
y
.
E
n
h
a
n
ced
p
r
ed
i
ct
i
o
n
accu
r
at
e
n
es
s
h
as
p
r
o
v
i
d
ed
b
y
a s
t
r
o
n
g
er
cl
a
s
s
i
f
i
er
.
T
h
e accu
r
ac
y
o
f
t
h
e s
y
s
t
e
m
i
s
c
a
l
c
ul
a
t
e
d
us
i
ng t
he
e
xp
r
e
s
s
i
o
n,
FN
FP
TN
TP
TN
TP
A
ccu
ra
cy
+
+
+
+
=
(
11
)
F
or
t
r
a
i
n
i
ng
pu
r
pos
e
70%
of
da
t
a
i
s
us
e
d,
f
or
t
e
s
t
i
ng
15%
,
a
n
d f
or
v
a
l
i
da
t
i
on
15%
.
F
or
t
h
e
t
r
a
i
ni
n
g
7
-
1
-
1
,
th
e
A
N
N
s
tr
u
c
t
u
r
e
is
a
p
p
lie
d
.
F
o
r
th
e
p
r
e
d
ic
tio
n
,
th
e
A
d
a
-
B
P
N
N
m
o
d
el
u
s
es
t
h
e B
P
N
N
o
f
f
eeb
l
e
c
l
a
s
s
i
f
i
e
r
s
t
o c
o
m
pi
l
e
t
o s
t
r
ong
.
T
h
e
n
e
u
r
a
l
n
e
t
w
or
k t
ool
box
i
n
M
A
T
L
A
B
i
s
us
e
d t
o c
on
s
t
r
u
c
t
t
h
e
m
ode
l
s
.
T
h
e
m
ean
er
r
o
r
r
at
e o
f
a s
t
r
o
n
g
cl
a
s
s
i
f
i
er
i
s
3
.
4
%
an
d
f
o
r
t
h
e
w
e
ak
er
cl
as
s
i
f
i
er
,
i
t
i
s
5
.
6
%
.
T
h
e co
m
p
et
e
n
ce o
f
t
h
e
p
r
o
j
ect
ed
s
y
s
t
e
m
i
s
m
o
r
e e
m
i
n
en
t
t
h
an
t
h
e n
o
r
m
al
B
P
m
o
d
el
an
d
o
t
h
er
ex
i
s
t
i
n
g
e
x
a
m
p
l
e
s
.
T
h
e b
es
t
v
al
i
d
at
i
o
n
p
er
f
o
r
m
a
n
ce an
d
t
h
e r
eg
r
es
s
i
o
n
an
al
y
s
i
s
ar
e ex
p
r
es
s
ed
i
n
F
i
g
ur
e
5 a
n
d
F
i
gur
e
6
(
s
ee i
n
ap
p
en
d
i
x
)
r
es
p
ect
i
v
el
y
.
Fi
g
ur
e
4.
A
da
boos
t
pe
r
f
or
m
a
nc
e
Fi
g
ur
e
5
.
N
N
p
e
r
f
o
r
m
a
nc
e
gr
a
p
h (
us
i
n
g M
AT
L
A
B
)
T
ab
l
e 2
.
A
ccu
r
ac
y
c
o
m
pa
r
i
s
on
A
cce
s
s
S
c
h
em
e
N
u
m
b
er
o
f
i
m
a
g
es
A
ccu
r
acy
(
%
)
Zh
a
o
e
t a
l
.
[
2
3]
S
VM
4
8
9
7
.
90
K
r
u
s
z
k
a
e
t
a
l.
[
2
4]
S
VM
6
5
9
4
.
30
B
u
r
c
i
n
a
n
d
V
a
s
i
f
[
25
]
LB
P
s
1
0
7
9
5
.
30
E
r
o
gu
l
e
t
a
l.
[
2
6]
E
B
GM
8
6
6
8
.
70
P
r
o
p
os
e
d
M
o
d
e
l
A
d
a
b
o
os
t
+
B
P
8
2
9
8
.
67
4.
CO
NCL
U
S
I
O
N
A
f
eed
-
f
o
r
w
ar
d
ar
t
i
f
i
ci
al
n
e
u
r
al
n
et
w
o
r
k
b
as
ed
o
n
a
b
o
o
s
t
i
n
g
t
ech
n
i
q
u
e f
o
r
t
h
e p
r
e
m
at
u
r
e d
et
ect
i
o
n
o
f
DS
i
s
de
m
ons
t
r
a
t
e
d.
F
or
t
h
i
s
pu
r
pos
e
-
d
e
s
i
g
ne
d
a
s
i
n
gl
e
l
a
ye
r
o
f
h
i
d
d
e
n ne
ur
o
ns
a
nd
a
s
i
ngl
e
o
ut
p
ut
ne
ur
o
n
n
et
w
o
r
k
.
T
h
er
e i
s
n
o
p
r
o
p
er
t
r
eat
m
e
n
t
f
o
r
D
o
w
n
s
y
n
d
r
o
m
e;
s
o
i
t
s
ear
l
y
d
et
ect
i
o
n
i
s
r
eal
l
y
i
n
d
i
s
p
en
s
ab
l
e an
d
n
eces
s
ar
y
.
T
h
e
p
r
o
p
o
s
ed
t
ech
n
i
q
u
e
i
s
b
as
ed
o
n
f
a
c
ia
l
id
e
n
t
if
ic
a
tio
n
a
n
d
lo
c
a
liz
a
tio
n
o
f
f
a
c
ia
l
p
o
in
ts
o
f
t
h
e
f
et
u
s
.
B
y
co
m
p
u
t
i
n
g
t
h
e l
en
g
t
h
b
et
w
ee
n
t
h
e
s
e p
o
i
n
t
s
,
t
h
e p
r
o
p
o
s
ed
al
g
o
r
i
t
h
m
p
r
ed
i
ct
s
t
h
e ch
an
ce f
o
r
D
o
w
n
s
y
n
d
r
o
m
e.
F
aci
al
i
d
en
t
i
f
i
cat
i
o
n
i
s
p
er
f
o
r
m
ed
w
i
t
h
t
h
e h
e
l
p
o
f
A
d
aB
o
o
s
t
an
d
n
eu
r
al
n
et
w
o
r
k
tr
a
in
in
g
i
s
p
e
r
f
o
r
m
e
d
w
it
h
th
e
h
e
lp
o
f
M
A
T
L
A
B
.
F
u
r
t
h
e
r
,
it is
p
o
s
s
ib
le
to
in
c
r
e
a
s
e
t
h
e accu
r
ac
y
b
y
i
n
cr
eas
i
n
g
t
h
e
n
u
m
b
e
r
o
f
tr
a
in
in
g
ite
r
a
tio
n
s
a
n
d
s
e
le
c
ti
n
g
m
o
r
e
s
e
n
s
iti
v
e
f
a
c
ia
l lo
c
a
liz
a
tio
n
p
o
in
ts
.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t
J
E
l
e
c
&
C
o
m
p
E
n
g
I
SSN
:
2088
-
8708
D
ow
n s
y
ndr
om
e
de
t
e
c
t
i
on us
i
ng m
odi
f
i
e
d
ada
boos
t
al
g
or
i
t
hm
(
V
i
nc
y
D
e
v
i
V
.
K
.
)
4287
AP
P
E
NDI
X
Fi
g
ur
e
6
.
NN p
e
r
f
o
r
m
a
n
c
e
-
r
e
gr
e
s
s
i
o
n
a
na
l
ys
i
s
R
EF
ER
EN
C
ES
[
1]
B
.
H
.
C
on
no
l
l
y
,
S
.
B
.
M
or
g
a
n,
F
.
F
.
R
u
s
s
el
l
,
an
d
W
.
L
.
F
ul
l
i
t
o
n,
“
A
L
ong
i
t
udi
na
l
S
t
udy
of
C
hi
l
dr
e
n w
i
t
h D
ow
n
S
y
ndr
om
e
W
ho E
x
pe
r
i
e
nc
e
d
E
a
r
l
y
I
nt
e
r
v
e
nt
i
on P
r
og
r
a
m
m
i
ng
,
”
P
h
ys
T
h
er
, v
o
l
. 7
3
,
n
o
. 3
,
pp.
1
70
–
1
79
,
19
93
,
doi
:
10
.
10
93
/
pt
j
/
7
3.
3.
1
70
.
[
2]
A
.
M
itta
l,
H.
Ga
u
r
,
a
nd
M
.
M
i
s
hr
a
,
“
D
e
t
e
c
t
i
on of
D
o
w
n S
y
ndr
om
e
U
s
i
ng
D
e
e
p F
a
c
i
a
l
R
e
c
ogni
t
i
o
n
,
”
i
n
B.
C
ha
ud
hur
i
,
M.
N
ak
ag
a
w
a,
P.
K
ha
nna
,
a
nd
S.
K
u
m
a
r (E
d
s
)
,
P
r
oc
e
e
di
ngs
of
3r
d I
nt
e
r
n
at
i
on
al
C
onf
e
r
e
nc
e
o
n
C
om
put
e
r
V
i
s
i
o
n an
d
I
m
age
P
r
o
c
e
s
s
i
ng
A
dv
anc
e
s
i
n
I
n
t
e
l
l
i
ge
n
t
S
y
s
t
e
m
s
and C
om
p
ut
i
ng
,
vo
l
.
10
22
,
pp.
1
19
-
13
0,
202
0
,
d
o
i:
10.
10
07/
97
8
-
98
1
-
32
-
9
088
-
4_
11
.
[
3]
V.
D
i
ma
,
A.
I
gn
a
t
,
a
nd
C.
R
us
u,
“
I
de
nt
i
f
y
i
ng D
ow
n
s
y
ndr
om
e
C
a
s
e
s
b
y
C
o
m
bi
ne
d U
s
e
o
f
F
a
c
e
R
e
c
og
ni
t
i
on
M
e
t
ho
ds
,”
i
n
V.
B
al
as
,
L.
J
ai
n
,
a
nd
M.
B
al
a
s (
E
d
s)
,
S
of
t
C
om
pu
t
i
ng A
ppl
i
c
at
i
o
ns
Ad
v
a
n
c
e
s
in
I
n
te
llig
e
n
t
S
y
s
te
m
s
and
C
om
put
i
ng
, S
p
r
i
n
g
e
r
, C
h
a
m
,
vo
l
.
63
4
,
pp
.
4
72
-
48
2
,
20
18
,
doi
:
10.
10
07/
97
8
-
3
-
3
19
-
62
52
4
-
9
_3
5.
[
4]
Q
.
Z
h
ao
,
K
.
R
o
s
en
b
au
m
,
K
.
O
k
ad
a,
D
.
J
.
Z
an
d
,
R
.
S
ze,
M
.
S
u
m
m
ar
et
al
.
,
“
A
ut
om
a
t
e
d D
ow
n
S
y
ndr
om
e
D
e
t
e
c
t
i
on
us
i
ng
F
a
c
i
a
l
P
h
ot
og
r
a
phs
,
”
3
5t
h
A
nnu
al
I
nt
e
r
n
at
i
o
na
l
C
onf
e
r
e
nc
e
of
t
he
I
E
E
E
E
M
B
S
,
Osa
k
a
,
Ja
p
a
n
,
Ju
l
.
20
13
,
pp.
36
70
-
36
73
,
d
o
i
:
1
0.
1
10
9/
E
M
B
C
.
20
13.
66
10
33
9
.
[
5]
H.
S
.
L
o
o
s,
D.
W
i
eczo
r
ek
,
R.
P
.
W
ü
rt
z
,
C.
v
on
de
r
M
a
l
s
bur
g
,
a
nd
B
.
Ho
r
st
h
e
m
k
e
,
“
C
o
m
put
e
r
-
ba
s
e
d r
e
c
og
ni
t
i
on
of
d
ys
m
o
r
p
h
i
c f
ac
es
,
”
E
ur
ope
an
J
our
n
al
of
H
um
an
G
e
ne
t
i
c
s
, v
o
l
. 1
1
,
no.
8,
pp
.
55
5
–
56
0,
20
03
,
doi
:
1
0.
1
03
8/
s
j
.
e
j
hg
.
52
00
99
7.
[
6]
S
.
Bo
e
h
ri
n
g
e
r,
M
.
G
ue
nt
he
r
,
S
.
S
i
n
i
ge
r
o
va
,
R
. P
.
W
u
r
t
z
,
B.
H
o
r
s
t
he
m
ke
,
a
nd D
.
W
i
ec
zo
r
ek
,
“
A
ut
om
a
t
e
d s
y
ndr
o
me
de
t
e
c
t
i
on i
n a
s
e
t
of
c
l
i
ni
c
a
l
f
a
c
i
a
l
phot
og
r
a
phs
,
”
A
m
e
r
i
c
an J
our
n
al
of
M
e
di
c
al
G
e
ne
t
i
c
s
P
ar
t
A
,
vol
.
15
5,
n
o.
9,
pp.
21
61
–
2
16
9,
20
11
,
d
o
i
:
1
0.
1
00
2/
a
j
m
g
.
a
.
34157
.
[
7]
S
.
S
a
ra
y
d
e
m
i
r,
N.
T
a
ş
pı
na
r
,
O.
E
r
o
ğ
u
l
,
H.
K
a
y
s
e
r
ili,
a
nd N
.
D
i
n
çk
an
.,
“
D
ow
n S
y
ndr
om
e
D
i
a
g
nos
i
s
B
a
s
e
d on
G
ab
o
r
W
av
el
et
T
r
an
s
f
o
r
m
,
”
J
our
nal
of
M
e
di
c
al
Sy
s
t
e
m
s
, v
o
l
. 3
6
,
no
.
5,
pp
.
32
05
–
3
21
3,
20
12
,
d
o
i:
10
.
10
07
/
s
10
91
6
-
011
-
98
11
-
1.
[
8]
T
.
A
.
A
nj
i
t
a
nd
S
.
R
i
s
hi
da
s
,
“
I
d
e
nt
i
f
i
c
a
t
i
on of
na
s
a
l
bo
ne
f
or
t
he
e
a
r
l
y
de
t
e
c
t
i
on of
D
ow
n s
y
ndr
om
e
us
i
ng
B
a
c
k
P
ro
p
ag
at
i
o
n
N
eu
r
al
N
et
w
o
r
k
,
”
20
11 I
nt
e
r
n
at
i
o
na
l
C
onf
e
r
e
nc
e
o
n
C
om
m
uni
c
a
t
i
o
ns
an
d
Si
g
nal
P
r
oc
e
s
s
i
ng
,
C
al
i
cu
t
,
201
1,
pp
.
1
36
–
14
0
,
d
oi
:
1
0.
1
10
9/
I
C
C
S
P
.
2
01
1.
57
39
28
6.
[
9]
C
. L
a
r
o
s
e
, P
. M
a
s
s
a
c
, Y
. H
i
l
l
i
o
n
, J
.
P
. B
e
r
n
a
r
d
,
a
nd
Y
.
V
i
lle
,
“
C
om
pa
r
i
s
on of
f
e
t
a
l
na
s
a
l
bone
a
s
s
es
s
m
en
t
b
y
u
ltr
a
s
o
u
n
d
a
t ll
-
14
w
e
e
ks
a
nd by
pos
t
m
or
t
e
m
X
-
r
a
y
in
tr
is
o
m
y
2
1
: a
pr
os
pe
c
t
i
v
e
obs
e
r
v
a
t
i
ona
l
s
t
u
d
y
,
”
U
l
t
r
a
s
o
u
nd
i
n O
bs
t
e
t
r
i
c
s
a
nd G
y
ne
c
ol
o
gy
,
J
ohn W
i
l
e
y
&
S
ons
,
L
t
d,
v
ol
.
22,
n
o
. 1
,
pp
.
27
–
30,
20
03
,
doi
:
10
.
10
02
/
uo
g.
16
9
.
[
1
0]
K
.
O
.
K
ag
an
,
S
.
C
i
cer
o
,
I
.
S
t
a
b
oul
i
do
u,
D
.
W
r
i
g
ht
,
a
nd
K.
H.
Ni
c
o
l
a
i
d
e
s,
“
'
P
et
al
n
as
al
b
o
n
e i
n
s
cr
een
i
n
g
f
o
r
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN
:
20
88
-
8708
In
t
J
E
l
e
c
&
C
o
m
p
E
n
g
,
V
o
l.
11
, N
o
.
5
,
O
c
t
obe
r
20
2
1
:
4
2
8
1
-
4
2
8
8
4288
t
r
i
s
om
i
e
s
21,
18
a
nd
13
a
nd T
ur
ne
r
s
y
ndr
o
m
e at
L
L
-
13 w
e
e
k
s
of
ge
s
t
a
t
i
on
,
”
Ul
t
ra
so
u
n
d
i
n
Ob
st
e
t
ri
c
s a
n
d
G
y
ne
c
ol
ogy
,
J
oh
n W
i
l
e
y
&
S
ons
,
L
t
d,
v
ol
.
33
,
p
p.
25
9
–
26
4,
20
04
,
doi
:
10
.
10
02
/
uo
g.
63
18
.
[
1
1]
J
.
S
o
ch
m
an
an
d
J
.
M
al
as
,
“
A
d
aB
o
o
s
t
w
i
t
h
t
o
t
al
l
y
co
r
r
ect
i
v
e u
p
d
at
es
f
o
r
f
as
t
f
ac
e d
et
ect
i
o
n
,
”
P
r
oc
e
e
di
ngs
o
f
S
i
x
th
I
E
E
E
I
nt
e
r
na
t
i
o
nal
C
o
nf
e
r
e
nc
e
on A
ut
om
at
i
c
F
a
c
e
an
d G
e
s
t
ur
e
R
e
c
ogni
t
i
on
,
S
e
oul
,
S
o
ut
h
K
or
e
a
,
2004
,
pp.
44
5
–
45
0,
do
i
:
1
0.
11
09/
A
F
G
R
.
20
04.
13
01
57
3.
[
1
2]
C.
S
.
R
a
m
,
“
R
e
c
og
ni
z
i
ng
F
a
c
e
E
m
ot
i
on of
D
ow
n S
y
ndr
om
e
C
hi
l
dr
e
n us
i
ng
V
i
ol
a
J
o
ne
T
e
c
hni
que
,
”
I
nt
e
r
nat
i
o
na
l
J
our
n
al
o
f
C
om
put
e
r
Sc
i
e
nc
e
T
r
e
nds
an
d T
e
c
hn
ol
ogy
(
I
J
C
ST
)
,
v
ol
.
7,
no.
2,
p
p.
93
-
10
0,
20
19.
[
1
3]
M
. G
.
Kr
i
sh
n
a
an
d
A
.
S
r
in
iv
a
s
u
lu
,
“F
ace D
et
ect
i
o
n
S
y
s
t
e
m
O
n
A
d
aB
o
o
s
t
A
l
g
o
r
i
t
h
m
U
s
i
n
g
H
aar
C
l
as
s
i
f
i
er
s
,
”
I
nt
e
r
nat
i
o
nal
J
our
n
al
of
M
o
de
r
n E
ngi
ne
e
r
i
ng R
e
s
e
ar
c
h,
v
ol
.
2,
n
o.
5,
pp.
3
35
6
–
35
60,
2
01
2.
[
1
4]
J
. Y
. R
. C
o
r
n
e
j
o
, H
.
P
e
d
r
i
n
i
, A
. M
a
c
h
a
d
o
-
L
i
m
a
, a
n
d
F
. D
. L
.
d
o
s S
a
n
t
o
s
Nu
n
e
s.
,
“
D
ow
n s
y
ndr
om
e
d
et
ect
i
o
n
b
as
ed
o
n
f
aci
al
f
e
at
u
r
es
u
s
i
n
g
a g
eo
m
e
t
r
i
c d
es
cr
i
p
t
o
r
,
”
J
our
nal
of
m
e
di
c
al
i
m
a
gi
n
g (
B
e
l
l
i
ng
ham
,
W
as
h.
)
,
vo
l
.
4
,
n
o.
4,
201
7
, A
r
t
. N
o
.
04
40
08
,
doi
:
1
0.
11
17/
1.
J
M
I
.
4.
4.
0
44
00
8
.
[
1
5]
R
.
L
i
e
nha
r
t
,
A
.
K
ur
a
nov
,
a
nd
P
.
V
a
di
m
,
“E
m
p
i
r
i
cal
an
al
y
s
i
s
o
f
d
et
ect
i
o
n
c
as
cad
es
o
f
b
o
o
s
t
ed
cl
as
s
i
f
i
er
s
f
o
r
r
ap
i
d
o
b
j
ect
d
et
ect
i
o
n
,
”
i
n
J
oi
nt
P
at
t
e
r
n R
e
c
ogn
i
t
i
o
n Sy
m
p
os
i
um
,
D
A
G
M
2003:
P
a
tte
r
n
Re
c
o
g
n
i
tio
n
,
pp.
2
97
-
3
04,
S
ep
.
200
3
,
d
o
i:
10.
10
07/
97
8
-
3
-
5
40
-
452
4
3
-
0_
39
.
[
1
6]
P
.
V
i
ol
a
a
nd M
.
J
o
ne
s
,
“R
o
b
u
s
t
r
eal
-
t
i
m
e o
b
j
ect
s
d
et
ect
i
o
n
,
”
i
n
I
nt
e
r
nat
i
on
al
j
our
n
al
of
c
om
put
e
r
v
i
s
i
on
,
vo
l
.
4,
no.
34
-
47
,
20
01
.
[
1
7]
S
. P
.
A
r
j
una
n
a
nd M
.
C
.
T
h
o
m
as
,
“
A
R
e
v
i
e
w
o
f
U
l
t
r
a
s
ound I
m
a
g
i
ng
T
e
c
hni
que
s
f
or
t
he
D
e
t
e
c
t
i
on of
D
o
w
n
S
y
ndr
om
e
,
”
I
R
B
M
,
2019
.
do
i
:
10
.
101
6/
j
.
i
r
bm
.
20
19.
10
.
0
04.
[
1
8]
Y
.
F
r
e
u
nd
a
nd
R.
E
.
S
c
ha
p
i
r
e
,
“
A
S
hor
t
I
nt
r
o
duc
t
i
on t
o B
o
os
t
i
ng
,
”
J
our
na
l
of
J
a
pa
ne
s
e
Soc
i
e
t
y
f
or
A
r
t
i
f
i
c
i
al
I
n
te
llig
e
n
c
e
,
v
ol
.
14,
n
o.
5,
pp.
7
7
1
–
78
0,
19
99
.
[
1
9]
R.
R.
C
h
illa
r
ig
e
,
S
.
D
i
s
t
ef
an
o
,
a
n
d
S.
S.
R
aw
at
,
“A
d
v
an
ce
s
i
n
C
o
m
p
u
t
at
i
o
n
al
I
n
t
el
l
i
g
en
ce an
d
I
n
f
o
r
m
at
i
cs
,
”
L
ect
u
r
e
N
ot
e
s
i
n N
e
t
w
or
k
s
an
d Sy
s
t
e
m
s
,
2
020
,
do
i
:
10
.
10
07
/
97
8
-
9
81
-
15
-
333
8
-
9.
[
2
0]
X
.
Li
u
,
J
.
B
a
o
, Y
.
J
i
a
n
g,
Z
.
K
e,
a
n
d
C
.
W
a
n
g
,
“
C
o
l
l
eg
e g
r
ad
u
at
es
'
e
m
p
l
o
y
m
en
t
pr
e
di
c
t
i
o
n ba
s
e
d o
n
P
r
i
nc
i
pa
l
C
om
pone
nt
A
na
l
y
s
i
s
a
nd t
he
c
om
bi
ne
d a
da
pt
i
v
e
boos
t
i
ng
a
nd t
he
ba
c
k
pr
opa
g
a
t
i
on ne
ur
a
l
ne
t
w
o
r
k
a
lg
o
r
ith
m
,
”
201
5
11t
h I
nt
e
r
n
at
i
on
al
C
onf
e
r
e
nc
e
on N
at
ur
al
C
om
put
at
i
o
n (
I
C
N
C
)
,
20
15
,
pp.
11
33
-
11
37
,
doi
:
1
0.
1
10
9/
i
c
nc
.
20
15
.
7
37
81
51.
[
2
1]
V
.
M
.
J
oy
a
nd S
.
K
r
i
s
hna
k
um
a
r
,
“
O
pt
i
m
a
l
D
e
s
i
g
n o
f
P
ow
e
r
S
c
he
dul
i
ng
us
i
ng
A
r
t
i
f
i
c
i
a
l
N
e
ur
a
l
N
e
t
w
or
k i
n A
n
I
s
o
l
at
ed
P
o
w
er
S
y
s
t
e
m
,
”
I
nt
.
J
.
o
f
P
ur
e
an
d A
pp
l
i
e
d M
at
he
m
at
i
c
s
,
v
ol
.
11
8,
no.
8,
p
p.
28
9
–
29
4,
20
18
.
[
2
2]
V
.
M
.
J
oy
a
nd S
.
K
r
i
s
hna
k
um
a
r
,
“
O
pt
i
m
a
l
de
s
i
g
n of
a
da
pt
i
v
e
pow
e
r
s
c
he
dul
i
ng
us
i
ng
m
odi
f
i
e
d a
nt
c
ol
o
ny
o
p
tim
iz
a
tio
n
a
lg
o
r
ith
m
,
”
I
nt
e
r
nat
i
on
al
J
our
nal
o
f
E
l
e
c
t
r
i
c
al
an
d C
om
put
e
r
E
n
gi
ne
e
r
i
n
g (
I
J
E
C
E
)
,
v
ol
.
10,
n
o.
1
,
pp.
73
8
–
74
5,
20
20
.
[
2
3]
Q.
Z
h
a
o
, K
.
R
os
e
n
ba
um
,
R
.
S
ze,
D
.
Z
a
n
d,
M.
S
u
m
m
ar
,
an
d
M
. G
.
L
i
ng
ur
a
r
u,
“
D
ow
n s
y
ndr
om
e
de
t
e
c
t
i
on f
r
om
f
a
c
i
a
l
phot
og
r
a
phs
us
i
ng
m
a
c
hi
ne
l
e
a
r
ni
ng
t
e
c
hni
q
ue
s
,
”
M
e
di
c
al
I
m
agi
n
g 2
01
3:
C
om
put
e
r
-
A
i
de
d D
i
ag
nos
i
s
,
20
13.
doi
:
10
.
11
17
/
12
.
2
00
72
67
.
[
2
4]
P.
K
r
u
s
zk
a,
A
.
R
.
P
o
r
r
a
s,
A
.
K.
S
obe
r
i
ng
,
F
.
A
.
I
k
ol
o,
S
.
L
a
Q
u
a,
V
.
S
ho
t
e
l
e
r
s
u
k
et
a
l
.
,
“
D
ow
n s
y
ndr
om
e
i
n
di
v
e
r
s
e
pop
ul
a
t
i
ons
,
”
A
m
er
i
ca
n
Jo
u
r
n
a
l
o
f
M
ed
i
ca
l
G
en
et
i
cs
P
a
r
t
A
.
,
v
o
l.
17
3
,
no.
1,
pp
.
42
–
53
,
20
17
,
doi
:
1
0.
1
00
2/
a
j
m
g
.
a
.
v
173.
1
.
[
2
5]
B
u
r
çi
n
K
.
an
d
V
as
i
f
N
.
V
.
,
“
D
ow
n s
y
ndr
om
e
r
e
c
og
ni
t
i
on
us
i
ng
l
oc
a
l
bi
na
r
y
pa
t
t
e
r
ns
a
nd s
t
a
t
i
s
t
i
c
a
l
e
v
a
l
ua
t
i
on of
t
he
s
ys
t
e
m
,
”
E
x
pe
r
t
Sy
s
t
.
A
ppl
.
v
ol
.
3
8,
n
o.
7,
p
p.
86
90
–
86
95
,
20
11
.
d
o
i:
1
0.
1
01
6/
j
.
e
s
w
a
.
201
1.
0
1.
07
6
.
[
2
6]
O.
E
r
o
g
u
l
, M
.
E
.
S
i
pa
hi
,
Y
.
T
unc
a
an
d
S
.
V
ur
uc
u,
“R
eco
g
n
i
t
i
o
n
of
D
ow
n s
y
ndr
om
e
s
us
i
ng
i
m
a
ge
a
na
l
y
s
i
s
,
”
i
n
1
4t
h
N
at
i
o
nal
B
i
om
e
di
c
al
E
n
gi
ne
e
r
i
n
g
M
e
e
t
i
ng
,
pp
.
1
-
4
,
2
00
9.
d
o
i:
10.
1
109
/
B
I
Y
O
M
U
T
.
200
9.
51
30
32
2
.
B
I
O
G
RAP
H
I
ES
O
F
AUT
H
O
RS
Vi
n
c
y
D
e
v
i
V.
K
.
i
s
w
or
k
i
ng
a
s
A
s
s
i
s
t
a
nt
P
r
of
e
s
s
or
i
n t
he
de
pa
r
t
m
e
nt
of
C
om
put
e
r
A
p
p
lic
a
tio
n
s
,
S
r
ee N
ar
a
y
an
a
G
u
r
uk
ul
a
m
C
ol
l
e
g
e
of
E
ng
i
ne
e
r
i
ng
,
K
a
da
y
i
r
uppu,
E
r
na
k
ul
a
m
.
S
h
e
ha
s
12
y
ear
s
o
f
t
each
i
n
g
e
x
p
er
i
en
ce
.
S
h
e i
s
d
o
i
n
g
h
er
r
es
ear
ch
i
n
M
ed
i
cal
I
m
ag
e
pr
oc
e
s
s
i
ng
.
H
er
ar
ea o
f
i
nt
e
r
e
s
t
i
nc
l
ude
C
om
put
e
r
N
e
t
w
or
k
i
ng
,
C
l
ou
d C
om
put
i
ng
a
nd
B
i
g
d
at
a A
n
al
y
s
i
s
.
R
aj
e
s
h
R
.
i
s
w
or
k
i
ng
a
s
A
s
s
oc
i
a
t
e
P
r
of
e
s
s
or
of
C
om
put
e
r
S
c
i
en
ce i
n
t
h
e D
ep
ar
t
m
en
t
o
f
C
o
m
p
u
t
er
S
ci
en
ce,
C
H
R
I
S
T
(
D
ee
m
ed
t
o
b
e U
n
i
v
er
s
i
t
y
)
B
an
g
al
o
r
e,
I
n
d
i
a.
D
r
.
R
aj
es
h
r
es
ear
ch
i
nt
e
r
e
s
t
s
a
r
e
i
n t
he
a
r
e
a
s
of
D
a
t
a
S
t
r
uc
t
ur
e
s
a
n
d A
na
l
y
s
i
s
of
A
l
gor
i
t
hm
s
.
H
e
ha
s
publ
i
s
he
d
33
pa
pe
r
s
i
n v
a
r
i
o
us
J
our
na
l
s
a
nd
C
o
nf
e
r
e
nc
e
s
.
D
r
.
R
aj
es
h
i
s a
l
so
s
e
r
vi
ng
as
M
a
na
g
i
ng
E
di
t
or
,
L
e
a
d
G
ue
s
t
E
di
t
or
,
A
s
s
oc
i
a
t
e
E
di
t
or
,
E
di
t
o
r
i
a
l
b
oa
r
d m
e
m
be
r
a
nd T
e
c
hni
c
al
C
o
m
m
i
t
t
ee
m
e
m
b
er
o
f
v
ar
i
o
u
s
N
at
i
o
n
al
an
d
I
n
t
er
n
at
i
o
n
al
C
o
n
f
er
en
ces
an
d
J
o
u
r
n
al
s
.
H
e h
as
r
ec
ei
v
ed
V
een
u
s
I
nt
e
r
na
t
i
o
na
l
F
ou
nda
t
i
on’
s
O
ut
s
t
a
ndi
ng
F
a
c
ul
t
y
a
w
a
r
d
a
nd
D
ew
an
g
M
eh
t
a E
d
u
cat
i
o
n
L
ead
er
s
h
i
p
A
w
a
r
d
to
h
is
c
r
e
d
it.
Evaluation Warning : The document was created with Spire.PDF for Python.