I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Art
if
icia
l In
t
ellig
ence
(
I
J
-
AI
)
Vo
l.
14
,
No
.
2
,
A
p
r
il
20
25
,
p
p
.
1
3
0
2
~
1
3
1
0
I
SS
N:
2
2
5
2
-
8
9
3
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijai.v
14
.i
2
.
p
p
1
3
0
2
-
1
3
1
0
1302
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
Depressi
o
n
de
tec
t
io
n t
hro
ug
h t
ra
nsfo
rmers
-
ba
sed
e
mo
tion
recog
nition in m
u
ltiva
ria
te
time
ser
ies facia
l data
K
enj
o
v
a
n Na
ng
g
a
la
1
,
G
re
g
o
rius
Na
t
a
na
el
E
lwireha
rdj
a
2,
3
,
B
ens
P
a
rda
m
ea
n
1,
2
1
M
a
s
t
e
r
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
S
c
h
o
o
l
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
B
i
n
a
N
u
sa
n
t
a
r
a
U
n
i
v
e
r
s
i
t
y
,
Ja
k
a
r
t
a
,
I
n
d
o
n
e
s
i
a
2
B
i
o
i
n
f
o
r
m
a
t
i
c
s
a
n
d
D
a
t
a
S
c
i
e
n
c
e
R
e
s
e
a
r
c
h
C
e
n
t
e
r
,
B
i
n
a
N
u
sa
n
t
a
r
a
U
n
i
v
e
r
si
t
y
,
J
a
k
a
r
t
a
,
I
n
d
o
n
e
s
i
a
3
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
S
c
h
o
o
l
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
B
i
n
a
N
u
sa
n
t
a
r
a
U
n
i
v
e
r
si
t
y
,
J
a
k
a
r
t
a
,
I
n
d
o
n
e
s
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
No
v
1
2
,
2
0
2
3
R
ev
is
ed
No
v
1
6
,
2
0
2
4
Acc
ep
ted
No
v
2
4
,
2
0
2
4
G
lo
b
a
ll
y
,
t
h
e
p
re
v
a
len
c
e
o
f
m
e
n
t
a
l
h
e
a
lt
h
d
iso
r
d
e
rs,
p
a
rti
c
u
larly
d
e
p
re
ss
io
n
,
h
a
s
b
e
c
o
m
e
a
p
re
ss
in
g
issu
e
.
Ea
rly
d
e
tec
ti
o
n
a
n
d
in
terv
e
n
ti
o
n
a
r
e
v
it
a
l
to
m
it
ig
a
te
th
e
p
r
o
fo
u
n
d
imp
a
c
t
o
f
d
e
p
re
ss
io
n
o
n
i
n
d
iv
i
d
u
a
ls
a
n
d
so
c
iety
.
Lev
e
ra
g
in
g
tran
sfo
rm
e
r
m
o
d
e
ls,
re
n
o
wn
e
d
fo
r
th
e
ir
e
x
c
e
ll
e
n
c
e
in
n
a
t
u
ra
l
lan
g
u
a
g
e
p
ro
c
e
ss
in
g
a
n
d
ti
m
e
se
ries
tas
k
s,
we
e
x
p
lo
re
th
e
ir
a
p
p
l
ica
ti
o
n
in
d
e
p
re
ss
io
n
d
e
tec
ti
o
n
u
si
n
g
m
u
lt
i
v
a
riate
ti
m
e
se
ries
(M
TS
)
d
a
ta
f
ro
m
fa
c
ial
e
x
p
re
ss
io
n
s.
Tran
sfo
rm
e
r
m
o
d
e
ls
e
x
c
e
l
in
se
q
u
e
n
ti
a
l
d
a
ta
p
r
o
c
e
ss
in
g
b
u
t
re
m
a
in
re
lativ
e
ly
u
n
e
x
p
l
o
re
d
in
f
a
c
ial
e
x
p
re
ss
io
n
a
n
a
ly
sis.
T
h
is
stu
d
y
a
im
s
to
c
o
m
p
a
re
t
ra
n
sfo
rm
e
r
m
o
d
e
ls
a
p
p
li
e
d
to
first
-
o
rd
e
r
ti
m
e
d
e
ri
v
a
ti
v
e
d
a
ta
with
trad
it
io
n
a
l
m
e
th
o
d
s.
We
u
se
th
e
d
istres
s
a
n
a
l
y
sis
i
n
terv
ie
w
c
o
rp
u
s
wiz
a
rd
o
f
o
z
(DA
IC
-
WOZ
)
d
a
tas
e
t
a
n
d
e
v
a
lu
a
te
m
o
d
e
ls
with
m
e
a
n
a
b
so
l
u
te
e
rro
r
(M
AE)
a
n
d
r
o
o
t
m
e
a
n
sq
u
a
re
d
e
rro
r
(RM
S
E)
m
e
tri
c
s.
Re
su
lt
s
sh
o
w
th
a
t
t
ra
n
sfo
rm
e
r
m
o
d
e
ls
o
n
fir
st
d
e
riv
a
ti
v
e
s
o
u
tp
e
rf
o
rm
o
th
e
r
s
with
a
n
M
AE
o
f
4
.
4
2
a
n
d
R
M
S
E
o
f
5
.
4
2
.
W
h
il
e
tran
sf
o
rm
e
r
m
o
d
e
ls
o
n
ra
w
d
a
ta
su
rp
a
ss
XG
Bo
o
st
i
n
RM
S
E
,
t
h
e
y
fa
ll
sh
o
rt
o
f
LS
TM
+
tran
sfo
rm
e
r
with
a
n
M
AE
o
f
5
.
4
1
a
n
d
RM
S
E
o
f
6
.
0
2
.
P
re
p
r
o
c
e
ss
in
g
th
r
o
u
g
h
d
iff
e
re
n
ti
a
ti
o
n
e
n
h
a
n
c
e
s
t
ra
n
sfo
rm
e
r
m
o
d
e
ls'
a
b
i
li
ty
to
c
a
p
tu
re
tem
p
o
ra
l
p
a
tt
e
rn
s,
p
ro
m
isin
g
imp
ro
v
e
d
d
e
p
re
ss
io
n
d
e
tec
ti
o
n
a
c
c
u
ra
c
y
.
K
ey
w
o
r
d
s
:
Dee
p
lear
n
in
g
Facial
b
eh
av
io
r
an
al
y
s
is
Facial
d
ata
Ma
jo
r
d
ep
r
ess
iv
e
d
is
o
r
d
er
T
r
an
s
f
o
r
m
e
r
s
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Ken
jo
v
an
Nan
g
g
ala
Ma
s
ter
o
f
C
o
m
p
u
ter
Scien
ce
,
Sch
o
o
l o
f
C
o
m
p
u
ter
Scien
ce
,
B
in
a
Nu
s
an
tar
a
Un
iv
er
s
ity
J
ak
ar
ta
1
1
4
8
0
,
I
n
d
o
n
esia
E
m
ail: k
en
jo
v
an
.
n
an
g
g
ala@b
in
u
s
.
ac
.
id
1.
I
NT
RO
D
UCT
I
O
N
Me
n
tal
h
ea
lth
d
is
o
r
d
er
s
,
p
ar
t
icu
lar
ly
d
e
p
r
ess
io
n
,
h
a
v
e
e
v
o
lv
ed
in
to
a
s
ig
n
if
ica
n
t
g
lo
b
al
co
n
ce
r
n
,
im
p
ac
tin
g
m
illi
o
n
s
o
f
in
d
iv
id
u
als
wo
r
ld
wid
e
in
r
ec
e
n
t
y
ea
r
s
[
1
]
,
[
2
]
.
E
v
e
n
if
it
is
alr
ea
d
y
q
u
ite
s
ev
er
e,
it
ca
n
lead
to
ca
s
es
o
f
s
u
icid
e
[
3
]
.
T
h
e
cr
itical
r
o
le
o
f
ea
r
ly
d
etec
tio
n
an
d
in
ter
v
e
n
tio
n
in
allev
i
atin
g
th
e
p
r
o
f
o
u
n
d
co
n
s
eq
u
en
ce
s
o
f
d
ep
r
ess
io
n
o
n
in
d
i
v
id
u
als
an
d
s
o
ciety
at
la
r
g
e
is
wid
ely
ac
k
n
o
wled
g
ed
[
4
]
,
[
5
]
.
I
n
th
e
er
a
o
f
ar
tific
ial
in
tellig
en
ce
an
d
d
ee
p
lear
n
in
g
,
r
esear
ch
e
r
s
ar
e
ac
ti
v
ely
ex
p
lo
r
in
g
in
n
o
v
ativ
e
a
p
p
r
o
ac
h
es
to
e
n
h
an
ce
th
e
ac
cu
r
ac
y
a
n
d
e
f
f
icien
cy
o
f
d
ep
r
ess
io
n
d
etec
tio
n
[
6
]
.
T
h
is
s
tu
d
y
asp
ir
es
to
co
n
tr
ib
u
te
to
th
e
ex
p
an
d
in
g
b
o
d
y
o
f
k
n
o
wled
g
e
in
t
h
e
f
ield
o
f
m
e
n
tal
h
ea
lth
ass
es
s
m
en
t,
with
a
s
p
ec
if
ic
f
o
cu
s
o
n
th
e
d
etec
tio
n
o
f
d
e
p
r
es
s
io
n
.
T
o
ac
co
m
p
lis
h
th
is
,
we
h
ar
n
ess
t
r
an
s
f
o
r
m
er
m
o
d
els,
a
d
ee
p
n
eu
r
al
n
etwo
r
k
ar
ch
itectu
r
e
r
en
o
wn
e
d
f
o
r
it
s
ex
ce
p
tio
n
al
p
er
f
o
r
m
a
n
ce
in
a
v
ar
iety
o
f
n
atu
r
al
lan
g
u
ag
e
p
r
o
ce
s
s
in
g
(
NL
P)
an
d
tim
e
s
er
ies
d
ata
task
s
[
7
]
,
[
8
]
.
T
r
an
s
f
o
r
m
er
s
m
o
d
el
em
p
lo
y
atten
tio
n
m
ec
h
an
is
m
s
an
d
s
elf
-
atten
tio
n
lay
er
s
to
g
r
asp
th
e
co
n
n
ec
tio
n
s
b
etwe
en
wo
r
d
s
a
n
d
p
h
r
ase
s
in
a
p
r
o
v
id
ed
tex
t
[
9
]
.
W
h
ile
t
r
an
s
f
o
r
m
er
m
o
d
e
ls
h
av
e
g
ar
n
er
e
d
co
n
s
id
er
a
b
l
e
atten
tio
n
f
o
r
t
h
eir
ef
f
ec
tiv
e
n
ess
in
p
r
o
ce
s
s
in
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Dep
r
ess
io
n
d
etec
tio
n
th
r
o
u
g
h
tr
a
n
s
fo
r
mer
s
-
b
a
s
ed
emo
tio
n
r
ec
o
g
n
itio
n
i
n
…
(
K
en
jo
va
n
N
a
n
g
g
a
la
)
1303
s
eq
u
en
tial
d
ata,
th
ei
r
ap
p
licatio
n
in
t
h
e
d
o
m
ain
o
f
f
ac
ial
e
x
p
r
ess
io
n
an
aly
s
is
,
p
a
r
ticu
lar
l
y
in
th
e
co
n
tex
t
o
f
m
u
ltiv
ar
iate
tim
e
s
er
ies
(
MT
S)
d
ata,
h
as r
em
ai
n
ed
r
elativ
ely
u
n
ch
ar
te
d
ter
r
ito
r
y
[
1
0
]
.
T
h
is
r
esear
ch
is
g
r
o
u
n
d
e
d
i
n
th
e
f
o
llo
win
g
ce
n
tr
al
r
ese
ar
ch
q
u
esti
o
n
:
wh
y
d
o
t
h
e
r
eg
r
ess
io
n
o
u
tco
m
es
r
esu
ltin
g
f
r
o
m
th
e
u
tili
za
tio
n
o
f
a
t
r
an
s
f
o
r
m
er
m
o
d
el
o
n
f
i
r
s
t
-
o
r
d
er
tim
e
d
er
iv
ativ
e
d
ata
o
f
f
ac
ial
b
eh
av
io
r
in
d
ep
r
ess
io
n
d
etec
tio
n
co
m
p
a
r
e
to
th
o
s
e
o
f
th
e
b
a
s
ic
m
eth
o
d
?
T
h
e
r
esear
ch
en
d
ea
v
o
r
en
co
m
p
ass
es
an
ex
p
l
o
r
atio
n
o
f
th
e
p
o
ten
ti
al
o
f
t
r
a
n
s
f
o
r
m
e
r
m
o
d
els
in
th
e
r
ea
lm
o
f
d
ep
r
ess
io
n
d
ete
ctio
n
th
r
o
u
g
h
th
e
an
aly
s
is
o
f
MT
S
d
ata
d
er
iv
ed
f
r
o
m
f
ac
ial
ex
p
r
ess
io
n
s
[
1
1
]
.
T
h
e
ch
o
ice
o
f
u
tili
zin
g
a
t
r
a
n
s
f
o
r
m
er
m
o
d
el
f
o
r
th
is
r
esear
ch
s
tem
s
f
r
o
m
its
ab
ilit
y
to
ef
f
ec
tiv
ely
ca
p
tu
r
e
c
o
m
p
lex
s
eq
u
e
n
tial
p
atter
n
s
an
d
d
ep
e
n
d
en
cies
i
n
MTS
d
ata,
m
ak
in
g
it
a
p
r
o
m
i
s
in
g
ap
p
r
o
ac
h
f
o
r
th
e
task
o
f
d
ep
r
ess
io
n
d
etec
tio
n
.
Fu
r
th
er
m
o
r
e,
o
u
r
aim
is
to
ev
alu
ate
th
e
p
er
f
o
r
m
an
ce
o
f
t
r
an
s
f
o
r
m
er
m
o
d
els in
co
m
p
a
r
is
o
n
to
b
aselin
e
m
o
d
els,
with
a
p
ar
ticu
lar
em
p
h
asis
o
n
th
eir
ab
ilit
y
to
ca
p
t
u
r
e
lo
n
g
-
ter
m
tem
p
o
r
al
d
ep
e
n
d
en
cies
[
1
2
]
.
Ad
d
itio
n
ally
,
th
is
s
tu
d
y
s
ee
k
s
to
ass
e
s
s
th
e
s
u
cc
ess
o
f
th
e
t
r
an
s
f
o
r
m
e
r
-
b
a
s
ed
ap
p
r
o
ac
h
to
d
ep
r
ess
io
n
d
etec
tio
n
as
a
v
alu
ab
le
to
o
l
f
o
r
ea
r
ly
in
ter
v
en
tio
n
an
d
m
en
tal
h
ea
lth
s
u
p
p
o
r
t.
2.
RE
L
AT
E
D
WO
RK
S
Sev
er
al
n
o
tab
le
s
tu
d
ies
in
th
e
d
o
m
ain
o
f
d
ep
r
ess
io
n
d
etec
tio
n
th
r
o
u
g
h
f
ac
ial
a
n
aly
s
is
h
av
e
p
av
ed
th
e
way
f
o
r
in
n
o
v
ativ
e
ap
p
r
o
ac
h
es.
Gav
r
iles
cu
an
d
Vizir
ea
n
u
[
1
3
]
u
tili
ze
d
th
e
f
ac
ial
ac
ti
o
n
co
d
i
n
g
s
y
s
tem
(
FAC
S)
to
d
is
ce
r
n
d
e
p
r
ess
io
n
,
a
n
x
iety
,
an
d
s
tr
ess
(
DASS)
lev
els,
ac
h
iev
in
g
im
p
r
e
s
s
iv
e
ac
cu
r
ac
ies
o
f
8
7
.
2
%
f
o
r
d
ep
r
ess
io
n
,
7
7
.
9
%
f
o
r
an
x
iety
,
an
d
9
0
.
2
%
f
o
r
s
tr
ess
with
a
u
n
iq
u
e
th
r
e
e
-
lay
er
ar
ch
itectu
r
e
.
Mu
za
m
m
el
et
a
l.
[
1
4
]
d
el
v
ed
in
to
m
ajo
r
d
e
p
r
ess
iv
e
d
is
o
r
d
er
(
MD
D)
d
etec
tio
n
,
em
p
lo
y
i
n
g
lo
n
g
s
h
o
r
t
-
ter
m
m
em
o
r
y
(
L
STM
)
an
d
co
n
v
o
l
u
tio
n
al
n
eu
r
al
n
etwo
r
k
(
C
NN
)
,
with
s
lig
h
tly
im
p
r
o
v
ed
a
cc
u
r
ac
y
o
f
6
6
.
2
5
%
co
m
p
ar
ed
to
6
5
.
6
0
%
f
o
r
b
i
n
ar
y
ca
s
es
o
f
d
ep
r
ess
io
n
.
Gr
im
m
et
a
l.
[
1
5
]
in
tr
o
d
u
ce
d
th
e
p
atien
t
h
ea
lth
q
u
esti
o
n
n
air
e
(
PHQ
-
V)
an
d
g
en
er
alize
d
an
x
iety
d
is
o
r
d
er
(
GAD
-
V)
,
r
ep
lacin
g
PHQ
-
9
an
d
GAD
-
7
,
em
p
lo
y
in
g
th
r
ee
tr
an
s
f
o
r
m
er
b
lo
c
k
s
,
lead
in
g
to
im
p
r
o
v
ed
r
esu
lts
.
Su
n
e
t
a
l.
[
1
6
]
cr
af
ted
t
h
e
d
ee
p
f
ea
tu
r
e
f
u
s
io
n
n
etwo
r
k
(
DFFN),
ac
h
iev
in
g
s
u
p
er
io
r
r
e
s
u
lts
in
d
ep
r
ess
io
n
d
etec
tio
n
f
r
o
m
a
f
u
s
io
n
o
f
tex
t,
au
d
io
,
an
d
v
id
eo
m
o
d
alities
with
a
p
r
ec
is
io
n
s
co
r
e
o
f
0
.
9
1
.
Su
n
et
a
l.
[
1
7
]
h
ar
n
ess
ed
th
e
tr
an
s
f
o
r
m
er
n
etwo
r
k
to
d
etec
t
MD
D,
s
u
r
p
ass
in
g
b
aselin
e
m
eth
o
d
s
with
a
co
n
co
r
d
an
ce
co
r
r
elatio
n
c
o
ef
f
icien
t
(
C
C
C
)
s
co
r
e
o
f
0
.
7
3
3
.
R
asip
u
r
am
et
a
l.
[
1
8
]
c
o
m
b
in
ed
C
NN
an
d
b
id
ir
ec
tio
n
al
lo
n
g
s
h
o
r
t
-
te
r
m
m
e
m
o
r
y
(
B
iLST
M
)
lay
er
s
f
o
r
MD
D
d
etec
tio
n
,
ex
ce
llin
g
in
ad
d
r
ess
in
g
im
b
alan
ce
d
d
ata
,
g
en
d
er
b
ias,
an
d
s
m
all
-
s
ca
le
d
ataset
ch
allen
g
es.
T
ig
g
a
an
d
Gar
g
[
1
9
]
ex
p
l
o
r
e
d
elec
tr
o
en
ce
p
h
alo
g
r
a
m
(
E
E
G)
s
ig
n
als
with
an
atten
tio
n
-
b
ased
g
ated
r
ec
u
r
r
e
n
t
u
n
its
tr
an
s
f
o
r
m
e
r
(
AttGR
UT
)
tim
e
s
er
ies
n
eu
r
al
n
etwo
r
k
,
co
n
s
is
ten
tly
o
u
tp
er
f
o
r
m
in
g
o
th
e
r
tim
e
-
s
er
ies
m
o
d
els.
T
iwar
y
et
a
l.
[
2
0
]
a
ch
iev
ed
th
e
h
ig
h
est
ac
cu
r
ac
y
o
f
8
2
%
f
o
r
MD
D
d
etec
tio
n
u
s
in
g
a
d
ee
p
C
NN
m
o
d
el
co
u
p
led
with
th
e
DASS
an
d
FAC
S.
Sh
an
g
g
u
an
et
a
l
.
[
2
1
]
ac
h
iev
ed
a
n
ac
c
u
r
ac
y
o
f
7
4
.
7
%
an
d
a
r
ec
all
o
f
7
4
.
5
%
with
th
e
atten
tio
n
-
b
ased
d
ee
p
d
o
m
ain
m
atch
in
g
in
s
tan
ce
lear
n
in
g
(
ADDM
I
L
)
m
o
d
el
f
o
r
MD
D
d
etec
tio
n
.
Gu
o
et
a
l.
[
2
2
]
co
m
b
in
ed
t
h
e
tem
p
o
r
al
d
ilated
co
n
v
o
lu
tio
n
n
etwo
r
k
(
T
DC
N)
b
r
a
n
ch
es,
ex
ce
llin
g
in
ter
m
s
o
f
ac
cu
r
ac
y
,
r
ec
all,
an
d
F1
-
s
co
r
e,
r
an
k
in
g
s
ec
o
n
d
b
es
t
f
o
r
p
r
ec
is
io
n
in
au
to
m
atic
d
ep
r
ess
io
n
d
etec
tio
n
.
T
h
ese
s
tu
d
ies
co
llectiv
ely
co
n
tr
ib
u
te
to
t
h
e
p
u
r
s
u
it
o
f
ac
cu
r
ate
d
ep
r
ess
io
n
d
etec
tio
n
m
eth
o
d
s
,
s
ig
n
alin
g
p
r
o
m
is
in
g
p
o
ten
tial f
o
r
th
e
f
iel
d
.
T
h
e
l
i
te
r
a
t
u
r
e
r
e
v
i
e
w
p
r
o
v
i
d
ed
v
a
l
u
a
b
l
e
i
n
s
i
g
h
ts
f
o
r
y
o
u
r
r
e
g
r
e
s
s
i
o
n
-
b
a
s
e
d
r
es
e
a
r
c
h
o
n
d
e
p
r
e
s
s
i
o
n
d
e
t
e
c
t
i
o
n
wi
t
h
i
n
M
T
S
d
a
t
a
u
s
in
g
a
t
r
a
n
s
f
o
r
m
e
r
m
o
d
e
l
.
T
h
e
im
p
l
e
m
e
n
t
a
t
i
o
n
o
f
t
h
e
t
r
a
n
s
f
o
r
m
e
r
m
o
d
e
l
,
i
n
s
p
i
r
e
d
b
y
s
t
u
d
i
e
s
l
i
k
e
S
u
n
e
t
a
l
.
[
1
7
]
,
r
e
m
a
i
n
s
a
p
r
o
m
i
s
i
n
g
a
p
p
r
o
a
c
h
f
o
r
m
o
d
e
l
i
n
g
a
n
d
r
e
g
r
e
s
s
i
n
g
d
e
p
r
e
s
s
i
o
n
s
e
v
e
r
it
y
l
e
v
e
l
s
b
a
s
e
d
o
n
v
i
d
e
o
-
b
as
e
d
MT
S
.
M
o
r
e
o
v
e
r
,
t
h
e
e
x
p
e
r
ie
n
c
e
s
o
f
h
a
n
d
l
i
n
g
d
a
t
a
i
m
b
a
l
a
n
ce
a
n
d
b
i
a
s
,
as
a
d
d
r
e
s
s
e
d
in
[
1
8
]
,
[
2
3
]
,
c
a
n
b
e
i
n
v
a
l
u
a
b
l
e
i
n
e
n
s
u
r
i
n
g
t
h
e
r
o
b
u
s
t
n
e
s
s
o
f
y
o
u
r
m
o
d
e
l
w
h
e
n
w
o
r
k
i
n
g
e
x
c
l
u
s
i
v
e
l
y
wi
t
h
v
i
d
eo
d
a
t
a
.
F
u
r
t
h
e
r
m
o
r
e
,
a
n
a
v
e
n
u
e
f
o
r
e
n
h
a
n
c
i
n
g
t
h
e
m
o
d
e
l
'
s
p
e
r
f
o
r
m
a
n
c
e
l
i
e
s
i
n
i
ts
a
d
a
p
t
at
i
o
n
a
n
d
f
i
n
e
-
t
u
n
i
n
g
f
o
r
v
i
d
e
o
d
a
t
a
,
w
i
t
h
t
h
e
o
b
j
e
c
t
i
v
e
o
f
m
i
n
i
m
i
z
i
n
g
m
e
a
n
a
b
s
o
l
u
t
e
er
r
o
r
(
M
A
E
)
a
n
d
r
o
o
t
m
e
a
n
s
q
u
a
r
e
d
e
r
r
o
r
(
R
M
S
E
)
,
a
s
i
n
s
p
i
r
e
d
b
y
n
o
t
a
b
l
e
a
c
h
i
e
v
em
e
n
t
s
i
n
t
h
e
f
i
e
l
d
o
f
d
e
p
r
e
s
s
i
o
n
d
e
t
e
c
t
i
o
n
as
d
o
c
u
m
e
n
t
e
d
i
n
t
h
e
l
i
t
e
r
at
u
r
e
.
3.
M
E
T
H
O
D
I
n
th
is
s
tu
d
y
,
we
em
p
lo
y
tr
an
s
f
o
r
m
er
m
o
d
els
to
d
etec
t
d
ep
r
e
s
s
io
n
f
r
o
m
v
id
eo
d
ata,
with
a
p
ar
ticu
lar
f
o
cu
s
o
n
MT
S
f
ac
ial
b
eh
a
v
io
r
d
ata.
T
h
e
c
h
o
ice
o
f
tr
an
s
f
o
r
m
er
s
is
m
o
tiv
ated
b
y
th
eir
p
r
o
v
en
ef
f
ec
tiv
e
n
ess
in
h
an
d
lin
g
MT
S
d
ata,
a
ch
ar
ac
t
er
is
tic
n
o
t
co
m
m
o
n
ly
e
x
p
lo
r
e
d
in
th
e
co
n
te
x
t
o
f
f
ac
ial
ex
p
r
ess
io
n
an
aly
s
is
f
o
r
d
ep
r
ess
io
n
d
etec
tio
n
.
C
o
m
p
ar
ed
to
r
ec
u
r
r
en
t
n
e
u
r
al
n
etwo
r
k
s
(
R
NNs)
[
2
4
]
,
tr
an
s
f
o
r
m
er
s
o
f
f
er
ad
v
an
tag
es
i
n
ca
p
tu
r
in
g
lo
n
g
-
r
an
g
e
tem
p
o
r
a
l
d
ep
en
d
e
n
cies
[
2
5
]
,
m
ak
in
g
th
em
s
u
itab
le
f
o
r
an
al
y
zin
g
f
ac
ial
b
eh
av
io
r
o
v
e
r
ex
ten
d
ed
d
u
r
atio
n
s
.
T
h
is
s
tu
d
y
alig
n
s
with
p
r
ev
io
u
s
r
esear
ch
b
y
R
asip
u
r
am
et
a
l
.
[
1
8
]
th
at
s
u
cc
ess
f
u
lly
u
tili
ze
d
tr
an
s
f
o
r
m
er
s
in
m
e
n
ta
l
h
ea
lth
d
is
o
r
d
er
d
etec
tio
n
,
p
r
im
ar
ily
in
au
d
i
o
an
d
te
x
t
m
o
d
a
liti
es.
Sti
ll,
it
aim
s
to
ex
ten
d
th
eir
a
p
p
licatio
n
to
th
e
u
n
d
er
e
x
p
lo
r
ed
r
ea
lm
o
f
tim
e
-
s
er
ies
f
ac
ial
ex
p
r
ess
io
n
d
a
ta.
B
y
ca
p
italizin
g
o
n
th
e
p
ar
allel
p
r
o
ce
s
s
in
g
c
ap
ab
ilit
ies
o
f
tr
an
s
f
o
r
m
er
s
,
th
is
s
tu
d
y
s
ee
k
s
to
en
h
an
ce
th
e
ef
f
icien
cy
an
d
ac
cu
r
ac
y
o
f
d
e
p
r
ess
io
n
d
etec
tio
n
m
eth
o
d
s
.
T
h
e
en
h
an
ce
d
ef
f
icien
cy
p
r
im
ar
ily
r
elate
s
to
th
e
m
o
d
el'
s
ab
ili
ty
to
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
1
4
,
No
.
2
,
Ap
r
il 2
0
2
5
:
1
3
0
2
-
1
3
1
0
1304
p
r
o
ce
s
s
d
ata
in
p
ar
allel,
wh
ich
ca
n
lead
to
f
aster
in
f
er
en
ce
tim
es
an
d
p
o
ten
tially
r
ed
u
ce
co
m
p
u
tatio
n
al
r
eso
u
r
ce
s
r
eq
u
ir
ed
f
o
r
d
e
p
r
ess
io
n
d
etec
tio
n
.
I
n
Fig
u
r
e
1
,
it
is
illu
s
tr
ated
t
h
at
th
is
p
ap
er
co
n
s
is
ts
o
f
s
ev
er
al
s
eq
u
en
tial
s
tep
s
.
T
h
ese
i
n
clu
d
e
d
ata
ac
q
u
is
itio
n
,
wh
ich
in
v
o
lv
es
u
tili
zin
g
th
e
d
is
tr
ess
an
aly
s
is
in
ter
v
iew
co
r
p
u
s
wiza
r
d
o
f
o
z
(
DAI
C
-
W
OZ
)
d
ataset,
f
o
llo
wed
b
y
p
r
ep
r
o
ce
s
s
in
g
th
e
o
b
tain
ed
d
ata.
Su
b
s
eq
u
en
tly
,
th
e
p
r
o
ce
s
s
en
co
m
p
ass
es
tr
ain
in
g
an
d
p
ar
am
eter
tu
n
in
g
,
m
o
d
el
e
v
alu
atio
n
,
an
d
u
ltima
tely
cu
lm
in
at
es in
p
u
b
lis
h
in
g
t
h
e
r
esear
ch
f
in
d
in
g
s
.
Fig
u
r
e
1
.
R
esear
ch
f
lo
w
3
.
1
.
Da
t
a
s
et
T
h
e
d
ataset
e
m
p
lo
y
e
d
f
o
r
th
i
s
r
esear
ch
is
s
o
u
r
ce
d
f
r
o
m
th
e
DAI
C
-
W
OZ
.
T
h
is
co
r
p
u
s
c
o
n
s
is
ts
o
f
clin
ical
in
ter
v
iews
co
n
d
u
cted
b
y
a
v
ir
tu
al
an
im
ate
d
in
ter
v
ie
wer
,
d
esig
n
ed
f
o
r
d
iag
n
o
s
in
g
p
s
y
ch
o
lo
g
ical
s
tr
ess
d
is
o
r
d
er
s
,
in
clu
d
in
g
d
ep
r
ess
io
n
.
T
h
e
d
ataset
p
r
o
v
id
es
a
r
ich
s
o
u
r
ce
o
f
r
ea
l
-
wo
r
ld
d
ata
f
o
r
a
n
aly
zin
g
an
d
u
n
d
er
s
tan
d
i
n
g
v
a
r
io
u
s
asp
ec
ts
o
f
th
ese
d
is
o
r
d
e
r
s
.
Utilizin
g
s
u
ch
d
ata
f
ac
ilit
ates
co
m
p
r
eh
en
s
iv
e
r
esear
ch
an
d
in
s
ig
h
ts
in
to
th
e
co
m
p
lex
ities
o
f
p
s
y
ch
o
lo
g
ical
d
is
tr
ess
.
T
h
e
v
id
eo
d
ata
in
th
is
d
ataset
co
m
p
r
is
es
in
f
o
r
m
atio
n
r
elate
d
to
f
ac
ial
ac
tio
n
u
n
its
th
at
ca
n
b
e
s
ee
in
Fig
u
r
es
2
an
d
3
,
f
o
r
th
e
u
p
p
er
an
d
lo
wer
f
ac
e,
r
esp
ec
tiv
ely
.
T
h
e
r
esear
ch
f
lo
w
en
c
o
m
p
ass
es
d
ata
ac
q
u
is
itio
n
,
p
r
ep
r
o
ce
s
s
in
g
,
f
ea
tu
r
e
s
elec
tio
n
u
s
in
g
Pear
s
o
n
c
o
r
r
elatio
n
(
PC
)
,
m
o
d
el
tr
ain
i
n
g
a
n
d
p
ar
am
eter
tu
n
in
g
,
an
d
f
in
ally
,
m
o
d
el
ev
alu
atio
n
.
T
h
e
p
r
im
ar
y
ev
alu
atio
n
m
etr
ics
in
clu
d
e
MA
E
a
n
d
R
MSE
to
ass
ess
th
e
m
o
d
el's
p
er
f
o
r
m
an
ce
in
p
r
ed
ictin
g
d
ep
r
ess
io
n
lev
els b
ased
o
n
f
ac
ial
b
eh
av
io
r
d
ata.
Fig
u
r
e
2
.
Up
p
er
f
ac
e
f
ac
ial
ac
t
io
n
u
n
it
[2
6
]
Fig
u
r
e
3
.
L
o
wer
f
ac
e
f
ac
ial
ac
tio
n
u
n
it
[2
6
]
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Dep
r
ess
io
n
d
etec
tio
n
th
r
o
u
g
h
tr
a
n
s
fo
r
mer
s
-
b
a
s
ed
emo
tio
n
r
ec
o
g
n
itio
n
i
n
…
(
K
en
jo
va
n
N
a
n
g
g
a
la
)
1305
3
.
2
.
Da
t
a
p
re
pro
ce
s
s
ing
B
ef
o
r
e
d
elv
in
g
in
to
th
e
p
r
ep
r
o
ce
s
s
in
g
s
tep
s
,
it
's
cr
u
cial
to
u
n
d
er
s
tan
d
th
e
p
i
v
o
tal
r
o
le
th
ey
p
lay
in
r
ef
in
in
g
r
aw
d
ata
f
o
r
an
aly
s
is
.
Pre
p
r
o
ce
s
s
in
g
en
co
m
p
ass
es
v
ar
io
u
s
tech
n
i
q
u
es
aim
e
d
at
en
h
an
cin
g
th
e
q
u
ality
an
d
r
eliab
ilit
y
o
f
th
e
d
ataset,
en
s
u
r
in
g
it
is
s
u
itab
le
f
o
r
f
u
r
th
er
an
aly
s
is
.
T
h
ese
s
tep
s
ty
p
ically
in
v
o
lv
e
d
o
wn
lo
ad
in
g
th
e
d
ataset,
ag
g
r
eg
atio
n
,
cr
o
p
p
in
g
,
s
p
litt
in
g
an
d
m
ea
n
ca
lcu
latio
n
t
h
at
ca
n
b
e
s
ee
n
in
Fig
u
r
e
4
.
B
y
ex
ec
u
tin
g
p
r
e
p
r
o
ce
s
s
in
g
d
ilig
en
tly
,
r
esear
ch
e
r
s
ca
n
m
itig
ate
p
o
ten
tial
b
iases
,
ad
d
r
ess
m
is
s
in
g
v
alu
es,
an
d
s
tan
d
ar
d
ize
d
ata
f
o
r
m
ats,
th
e
r
eb
y
lay
i
n
g
a
s
o
lid
f
o
u
n
d
atio
n
f
o
r
s
u
b
s
eq
u
en
t
an
aly
s
es.
Hen
c
e,
a
c
o
m
p
r
eh
en
s
iv
e
u
n
d
er
s
tan
d
i
n
g
o
f
p
r
ep
r
o
ce
s
s
in
g
tech
n
iq
u
es
is
in
d
is
p
en
s
ab
le
f
o
r
ex
t
r
ac
tin
g
m
ea
n
in
g
f
u
l
in
s
i
g
h
ts
f
r
o
m
th
e
d
ata
an
d
m
ak
i
n
g
in
f
o
r
m
e
d
d
ec
is
io
n
s
b
ased
o
n
an
aly
tical
o
u
tco
m
e
s
.
Fig
u
r
e
4
.
Data
p
r
ep
r
o
ce
s
s
in
g
f
lo
w
R
ef
er
to
Fig
u
r
e
4
,
we
ca
n
s
e
e
th
at
f
o
r
th
e
p
r
e
p
r
o
ce
s
s
in
g
s
tag
e,
af
ter
co
n
s
o
lid
atin
g
tex
t
f
iles
f
r
o
m
ea
ch
p
ar
ticip
a
n
t
in
to
a
d
ataf
r
am
e
with
v
i
d
eo
r
ec
o
r
d
i
n
g
c
o
o
r
d
in
ates,
t
h
e
r
esear
ch
e
r
p
r
o
ce
ed
s
to
cr
o
p
th
e
tim
estam
p
(
s
tar
t_
tim
e,
s
to
p
_
tim
e)
b
ased
o
n
t
h
e
au
d
i
o
r
ec
o
r
d
in
g
d
ata
in
t
h
e
tr
an
s
cr
ip
t
f
ile
.
T
h
e
p
u
r
p
o
s
e
is
to
alig
n
th
e
tim
estam
p
s
in
t
h
e
c
o
m
b
in
ed
d
ataf
r
a
m
e
with
th
o
s
e
in
th
e
a
u
d
io
r
ec
o
r
d
in
g
to
eli
m
in
ate
ir
r
elev
an
t
o
r
u
n
in
f
o
r
m
ativ
e
tim
estam
p
s
.
I
n
th
is
s
tu
d
y
,
all
d
ataset
f
ea
tu
r
es
ar
e
u
tili
ze
d
ex
ce
p
t
f
o
r
h
is
to
g
r
am
o
f
o
r
ien
ted
g
r
ad
ien
ts
(
HOG)
f
ea
tu
r
es.
T
h
e
r
esu
ltin
g
cr
o
p
p
ed
d
ata
is
th
en
ex
p
o
r
ted
to
co
m
m
a
-
s
ep
ar
a
ted
v
alu
es
(
C
SV
)
.
Fu
r
th
er
m
o
r
e
,
a
f
ir
s
t
d
e
r
iv
ati
v
e
o
f
t
h
e
cr
o
p
p
ed
d
ata
is
co
m
p
u
ted
,
r
ep
r
esen
tin
g
th
e
d
if
f
er
en
ce
b
etwe
en
co
n
s
ec
u
tiv
e
f
r
am
es,
an
d
th
ese
d
atasets
ar
e
also
ex
p
o
r
ted
to
C
SV.
T
h
e
f
ir
s
t
d
er
iv
ativ
e
aim
s
to
in
v
esti
g
ate
wh
eth
er
th
e
r
ate
o
f
ch
an
g
e
in
lan
d
m
ar
k
p
o
s
itio
n
s
o
r
s
p
ec
if
ic
f
ea
tu
r
es
s
ig
n
if
ica
n
tly
d
if
f
er
s
b
etwe
en
n
o
r
m
al
in
d
iv
id
u
als
an
d
th
o
s
e
with
d
e
p
r
ess
io
n
.
Af
ter
war
d
,
th
e
m
ea
n
v
alu
es
f
o
r
ea
ch
d
ataset
ar
e
ca
lcu
lated
f
o
r
f
ea
tu
r
e
s
elec
tio
n
,
an
d
o
n
ly
th
ese
m
ea
n
v
alu
es
ar
e
u
s
ed
f
o
r
th
e
s
elec
tio
n
p
r
o
ce
s
s
.
Su
b
s
eq
u
e
n
tly
,
t
h
e
s
elec
ted
f
ea
tu
r
es
ar
e
u
s
ed
in
th
e
MT
S
d
ata
f
o
r
m
o
d
elin
g
.
Fo
llo
win
g
d
ata
p
r
ep
r
o
ce
s
s
in
g
,
f
ea
tu
r
e
s
elec
tio
n
is
p
er
f
o
r
m
ed
u
s
in
g
PC
b
ased
o
n
f
ea
tu
r
es
with
s
i
g
n
if
ican
t
p
-
v
alu
es,
co
n
s
id
er
in
g
b
o
t
h
th
e
r
aw
a
n
d
f
ir
s
t
d
er
i
v
ativ
e
(
m
ea
n
)
d
ata.
Utilizin
g
PC
f
o
r
f
ea
tu
r
e
s
elec
tio
n
h
elp
s
id
en
tify
f
ea
tu
r
es
with
s
tr
o
n
g
co
r
r
elatio
n
s
,
r
ed
u
cin
g
d
ataset
d
im
en
s
io
n
s
,
en
h
an
cin
g
co
m
p
u
tatio
n
al
ef
f
icien
c
y
,
a
n
d
im
p
r
o
v
in
g
m
o
d
el
p
er
f
o
r
m
a
n
ce
b
y
r
etain
in
g
th
e
m
o
s
t
in
f
o
r
m
ativ
e
a
n
d
r
ele
v
an
t f
ea
tu
r
es in
d
ata
an
aly
s
is
.
Fo
llo
win
g
th
e
d
escr
ib
ed
d
at
a
p
r
ep
r
o
ce
s
s
in
g
s
tep
s
,
f
ea
tu
r
e
s
elec
tio
n
was
p
er
f
o
r
m
ed
u
s
in
g
PC
,
co
n
s
id
er
in
g
f
ea
tu
r
es
with
p
-
v
a
lu
es
<0
.
1
f
o
r
b
o
t
h
th
e
r
aw
a
n
d
f
ir
s
t
d
er
iv
ativ
e
(
m
ea
n
)
d
ata.
T
h
is
m
eth
o
d
h
elp
s
id
en
tify
f
ea
t
u
r
es
ex
h
i
b
itin
g
s
tr
o
n
g
c
o
r
r
elatio
n
s
,
lea
d
in
g
to
a
r
ed
u
ctio
n
in
d
ataset
d
im
e
n
s
io
n
s
.
C
o
n
s
eq
u
en
tly
,
it
en
h
an
ce
s
c
o
m
p
u
tatio
n
al
ef
f
icien
cy
a
n
d
m
o
d
el
p
e
r
f
o
r
m
an
ce
b
y
r
etain
in
g
t
h
e
m
o
s
t
in
f
o
r
m
ativ
e
an
d
r
ele
v
an
t
f
ea
tu
r
es
f
o
r
f
u
r
th
e
r
an
aly
s
is
.
Su
b
s
eq
u
en
tly
,
th
e
r
esu
lts
o
f
t
h
e
f
ea
tu
r
e
s
elec
tio
n
,
in
clu
d
in
g
th
e
p
-
v
al
u
e
o
f
ea
ch
f
ea
tu
r
e,
ar
e
p
r
esen
ted
f
o
llo
win
g
.
T
h
e
d
ata
p
r
esen
ted
in
T
a
b
le
1
h
ig
h
lig
h
ts
th
e
o
u
tco
m
es
o
f
f
ea
tu
r
e
s
elec
tio
n
b
ased
o
n
a
c
r
iter
io
n
o
f
p
-
v
alu
es b
elo
w
0
.
1
,
alb
eit
s
till
ap
p
r
o
ac
h
in
g
0
.
0
5
.
No
tab
ly
,
A
U1
2
_
r
ex
h
ib
its
a
p
-
v
alu
e
o
f
0
.
0
1
2
,
r
ep
r
esen
tin
g
a
s
ig
n
if
ican
t
f
ea
tu
r
e
ass
o
ciate
d
with
th
e
lip
ar
ea
.
Ad
d
itio
n
ally
,
y
_
h
0
an
d
y
_
h
1
,
with
p
-
v
alu
es
o
f
0
.
0
6
9
an
d
0
.
0
6
1
r
esp
ec
tiv
ely
,
a
r
e
id
en
tifi
ed
as c
r
u
cial
co
m
p
o
n
e
n
ts
o
f
e
y
e
g
az
e
with
in
th
e
d
ataset.
T
ab
le
1
.
R
esu
lts
o
f
f
ea
tu
r
e
s
el
ec
tio
n
F
e
a
t
u
r
e
s
P
-
V
a
l
u
e
A
U
1
2
_
r
0
.
0
1
2
y
_
h
0
0
.
0
6
9
y
_
h
1
0
.
0
6
1
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
1
4
,
No
.
2
,
Ap
r
il 2
0
2
5
:
1
3
0
2
-
1
3
1
0
1306
3
.
2
.
M
o
del
a
rc
hite
ct
ure
T
h
e
m
o
d
el'
s
ar
ch
itectu
r
e
s
er
v
es
as
th
e
b
lu
ep
r
in
t
th
at
o
u
tlin
es
its
s
tr
u
ctu
r
e,
in
clu
d
in
g
th
e
ar
r
an
g
em
en
t
o
f
lay
er
s
,
co
n
n
ec
tio
n
s
,
a
n
d
alg
o
r
ith
m
s
em
p
lo
y
ed
f
o
r
p
r
o
ce
s
s
in
g
d
ata.
Un
d
er
s
tan
d
in
g
th
e
ar
ch
itectu
r
e
p
r
o
v
id
es
in
s
ig
h
ts
in
to
h
o
w
th
e
m
o
d
el
lear
n
s
f
r
o
m
in
p
u
t
d
ata,
m
a
k
es
p
r
ed
ictio
n
s
,
an
d
ad
a
p
ts
to
v
ar
y
in
g
co
m
p
le
x
ities
.
Ad
d
itio
n
ally
,
it
s
h
ed
s
lig
h
t
o
n
th
e
m
o
d
el'
s
ca
p
ab
ilit
ies,
li
m
itatio
n
s
,
an
d
p
o
ten
tial
f
o
r
o
p
tim
izatio
n
.
Hen
ce
,
ex
p
lo
r
in
g
th
e
ar
ch
itectu
r
e
is
p
iv
o
tal
f
o
r
c
o
m
p
r
eh
en
d
in
g
th
e
m
o
d
el'
s
in
n
er
wo
r
k
in
g
s
an
d
its
ef
f
icac
y
in
ad
d
r
ess
in
g
th
e
r
esear
ch
o
b
jectiv
es.
I
n
th
is
r
esear
ch
,
we
u
s
e
tr
an
s
f
o
r
m
er
as
th
e
m
o
d
el
an
d
th
e
ar
ch
itectu
r
e
o
f
th
e
m
o
d
el
ca
n
b
e
s
ee
n
in
Fig
u
r
e
5
.
Fig
u
r
e
5
.
T
r
an
s
f
o
r
m
e
r
m
o
d
el
ar
ch
itectu
r
e
Fro
m
Fig
u
r
e
5
,
we
ca
n
s
ee
t
h
at
th
is
s
tu
d
y
u
tili
ze
s
a
tr
an
s
f
o
r
m
er
m
o
d
el
ar
ch
itectu
r
e
th
at
in
clu
d
es
p
o
s
itio
n
en
co
d
in
g
,
s
elf
-
atten
tio
n
lay
er
s
,
f
ee
d
-
f
o
r
war
d
n
e
u
r
a
l
n
etwo
r
k
lay
er
s
,
an
d
lin
ea
r
ac
tiv
atio
n
f
u
n
ctio
n
s
f
o
r
d
ep
r
ess
io
n
lev
el
p
r
ed
ictio
n
s
.
Hy
p
er
p
ar
a
m
eter
tu
n
in
g
,
p
er
f
o
r
m
e
d
th
r
o
u
g
h
g
r
id
s
ea
r
ch
,
allo
ws
f
o
r
o
p
tim
al
m
o
d
el
co
n
f
ig
u
r
atio
n
.
T
h
e
r
e
s
ea
r
ch
lev
er
ag
es
i
n
s
ig
h
ts
f
r
o
m
p
r
e
v
io
u
s
wo
r
k
in
th
e
d
o
m
ain
o
f
d
e
p
r
ess
io
n
d
etec
tio
n
wh
ile
e
x
ten
d
in
g
th
e
ap
p
licatio
n
o
f
tr
an
s
f
o
r
m
er
s
t
o
im
p
r
o
v
e
th
e
ac
c
u
r
ac
y
an
d
e
f
f
ec
tiv
en
ess
o
f
th
e
an
aly
s
is
o
f
MT
S
f
ac
ial
b
eh
a
v
io
r
d
ata,
u
ltima
tely
co
n
tr
ib
u
tin
g
to
th
e
d
ev
elo
p
m
en
t
o
f
m
o
r
e
r
eliab
le
an
d
ef
f
icien
t
d
ep
r
ess
io
n
d
etec
tio
n
m
eth
o
d
s
.
T
h
e
f
o
llo
win
g
h
y
p
er
p
ar
am
eter
s
wer
e
ex
p
lo
r
e
d
d
u
r
in
g
th
e
g
r
id
s
ea
r
ch
p
r
o
ce
s
s
:
-
L
ea
r
n
in
g
r
ate
: [
0
.
0
5
,
0
.
0
1
,
0
.
0
0
1
]
-
Neu
r
o
n
s
in
ea
ch
lay
er
: [
3
2
,
6
4
,
1
2
8
,
2
5
6
]
-
Nu
m
b
er
o
f
atten
tio
n
h
ea
d
s
: [
1
,
2
,
4
,
8
]
-
L
in
ea
r
u
n
its
: [
5
1
2
,
1
0
2
4
,
2
0
4
8
,
4
0
9
6
]
-
Dim
en
s
io
n
o
f
th
e
f
ee
d
-
f
o
r
war
d
(
DFF):
[
6
4
,
1
2
8
,
2
5
6
]
-
Dr
o
p
o
u
t
r
ate
: [
0
.
1
,
0
.
2
,
0
.
3
,
0
.
4
]
Fo
r
th
e
b
est p
ar
am
ete
r
s
ettin
g
s
u
s
ed
in
th
is
s
tu
d
y
ar
e
as f
o
llo
ws:
-
L
ea
r
n
in
g
r
ate
: 0
.
0
5
-
Neu
r
o
n
s
in
ea
ch
lay
er
: 1
2
8
-
Nu
m
b
er
o
f
atten
tio
n
h
ea
d
s
: 2
-
L
in
ea
r
u
n
its
: 2
0
4
8
-
DFF:
6
4
-
Dr
o
p
o
u
t
r
ate
: 0
.
1
3
.
3
.
M
o
del
e
v
a
lua
t
i
o
n
I
n
ev
al
u
atin
g
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
p
r
o
p
o
s
ed
tr
an
s
f
o
r
m
er
m
o
d
el
o
n
f
ir
s
t
d
er
i
v
ativ
e
v
id
e
o
d
ata,
we
em
p
lo
y
two
p
r
im
ar
y
m
etr
ics:
MA
E
an
d
R
MSE
.
T
h
ese
m
e
tr
ics
ar
e
ch
o
s
en
f
o
r
th
eir
ab
ilit
y
to
q
u
an
tify
th
e
av
er
ag
e
m
ag
n
itu
d
e
o
f
th
e
er
r
o
r
s
in
a
s
et
o
f
p
r
ed
ictio
n
s
with
o
u
t
co
n
s
id
er
in
g
th
eir
d
ir
ec
tio
n
.
T
h
ese
m
etr
ics
a
r
e
well
-
s
u
ited
f
o
r
o
u
r
s
tu
d
y
as
th
ey
o
f
f
e
r
a
clea
r
an
d
s
tr
aig
h
tf
o
r
war
d
in
ter
p
r
etatio
n
o
f
m
o
d
el
p
er
f
o
r
m
an
ce
.
MA
E
p
r
o
v
id
es
a
n
atu
r
al
a
n
d
ea
s
ily
in
ter
p
r
etab
le
m
ea
s
u
r
e
o
f
av
er
ag
e
er
r
o
r
m
a
g
n
itu
d
e
,
wh
ile
R
MSE
g
iv
es
a
h
ig
h
er
weig
h
t
to
lar
g
er
e
r
r
o
r
s
,
wh
i
ch
ca
n
b
e
p
ar
ticu
lar
ly
r
elev
an
t
in
s
ce
n
ar
i
o
s
wh
er
e
lar
g
e
er
r
o
r
s
ar
e
m
o
r
e
d
etr
im
en
tal
th
an
s
m
aller
o
n
es.
=
1
∑
=
1
|
−
̂
|
=
√
=
√
1
∑
=
1
(
−
̂
)
2
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Dep
r
ess
io
n
d
etec
tio
n
th
r
o
u
g
h
tr
a
n
s
fo
r
mer
s
-
b
a
s
ed
emo
tio
n
r
ec
o
g
n
itio
n
i
n
…
(
K
en
jo
va
n
N
a
n
g
g
a
la
)
1307
w
h
er
e:
̂
=
,
=
,
an
d
=
/
.
T
h
e
n
o
v
elty
o
f
o
u
r
r
esear
ch
lies
in
th
e
im
p
lem
en
tatio
n
o
f
tr
an
s
f
o
r
m
er
m
o
d
els
to
p
r
o
ce
s
s
f
ir
s
t
d
er
iv
ativ
e
d
ata
ex
tr
ac
ted
f
r
o
m
v
id
eo
s
eq
u
en
ce
s
.
T
h
is
ap
p
licatio
n
is
in
n
o
v
ativ
e,
a
s
tr
an
s
f
o
r
m
er
s
ar
e
p
r
ed
o
m
i
n
an
tly
u
s
ed
i
n
NL
P
a
n
d
h
av
e
o
n
ly
r
ec
e
n
tly
b
ee
n
e
x
p
lo
r
ed
in
th
e
co
n
tex
t
o
f
v
i
d
e
o
d
ata
an
aly
s
is
.
B
y
ap
p
ly
in
g
th
e
tr
an
s
f
o
r
m
er
'
s
s
elf
-
atten
tio
n
m
ec
h
an
is
m
s
to
th
e
t
em
p
o
r
al
d
y
n
am
ics
ca
p
tu
r
ed
in
th
e
f
ir
s
t
d
e
r
iv
ativ
e
o
f
v
id
eo
d
ata,
o
u
r
m
o
d
el
ai
m
s
to
e
n
h
an
ce
th
e
lear
n
in
g
o
f
tem
p
o
r
al
p
atter
n
s
th
at
a
r
e
cr
u
cial
f
o
r
ac
cu
r
ate
p
r
ed
ictio
n
s
.
B
y
c
o
m
p
a
r
in
g
o
u
r
tr
an
s
f
o
r
m
er
m
o
d
el'
s
p
er
f
o
r
m
an
ce
o
n
f
ir
s
t
d
e
r
iv
ativ
e
v
id
eo
d
ata
ag
ai
n
s
t
th
ese
b
aselin
es
u
s
in
g
MA
E
an
d
R
MSE
,
we
aim
to
d
em
o
n
s
tr
ate
its
ef
f
icac
y
an
d
th
e
p
o
ten
tial
ad
v
an
tag
es
o
f
o
u
r
ap
p
r
o
ac
h
f
o
r
v
id
e
o
-
b
ased
a
p
p
licatio
n
s
.
T
h
is
co
m
p
ar
is
o
n
allo
ws
u
s
to
estab
lis
h
th
e
tr
an
s
f
o
r
m
er
m
o
d
el'
s
r
o
b
u
s
tn
ess
an
d
ac
cu
r
ac
y
,
f
u
r
t
h
er
co
n
tr
ib
u
tin
g
to
th
e
d
o
m
ai
n
o
f
v
id
e
o
d
ata
an
aly
s
is
an
d
ex
p
an
d
i
n
g
th
e
u
tili
ty
o
f
tr
an
s
f
o
r
m
er
ar
c
h
itectu
r
es b
ey
o
n
d
t
h
eir
tr
ad
itio
n
al
d
o
m
ai
n
s
.
4.
RE
SU
L
T
AND
DI
SCUS
SI
O
N
I
n
th
is
s
ec
tio
n
,
we
m
eticu
lo
u
s
ly
d
is
s
ec
t
th
e
f
in
d
in
g
s
o
b
tai
n
ed
f
r
o
m
o
u
r
r
esear
ch
m
eth
o
d
o
lo
g
y
an
d
d
elv
e
in
to
th
eir
im
p
licatio
n
s
.
T
h
r
o
u
g
h
a
c
o
m
p
r
e
h
en
s
iv
e
ex
am
in
atio
n
o
f
th
e
r
esu
lts
,
we
aim
to
u
n
r
a
v
el
th
e
u
n
d
er
ly
i
n
g
p
atter
n
s
,
tr
e
n
d
s
,
a
n
d
co
r
r
elatio
n
s
em
b
ed
d
e
d
wit
h
in
th
e
d
ata.
Fu
r
th
er
m
o
r
e,
we
en
g
ag
e
in
a
cr
itical
d
is
co
u
r
s
e
to
c
o
n
tex
tu
alize
o
u
r
f
in
d
i
n
g
s
with
in
th
e
ex
is
tin
g
b
o
d
y
o
f
k
n
o
wled
g
e
,
o
f
f
er
in
g
in
s
ig
h
ts
in
to
th
eir
s
ig
n
if
ican
ce
an
d
p
o
ten
tial
c
o
n
tr
ib
u
tio
n
s
to
th
e
f
ield
.
B
y
i
n
ter
twin
in
g
th
e
p
r
esen
tatio
n
o
f
r
e
s
u
lts
with
in
-
d
ep
t
h
d
is
cu
s
s
io
n
s
,
we
s
tr
iv
e
to
p
r
o
v
id
e
a
h
o
lis
tic
u
n
d
er
s
tan
d
i
n
g
o
f
th
e
r
esear
ch
o
u
tc
o
m
e
s
an
d
th
eir
b
r
o
ad
er
im
p
licatio
n
s
.
4
.
1
.
Resul
t
I
n
th
is
r
esear
ch
,
th
e
r
esear
ch
e
r
u
tili
ze
d
a
tr
an
s
f
o
r
m
er
m
o
d
e
l
o
n
f
ir
s
t
d
er
iv
ativ
e
d
ata
as
well
as
r
aw
d
ata,
wh
ich
ca
n
b
e
s
ee
n
in
Fig
u
r
e
6
.
Fig
u
r
e
6
(
a)
tr
ain
in
g
a
n
d
v
alid
atio
n
lo
s
s
tr
an
s
f
o
r
m
er
f
ir
s
t
d
er
iv
ativ
e
an
d
Fig
u
r
e
6
(
b
)
tr
ai
n
in
g
a
n
d
v
alid
atio
n
lo
s
s
tr
an
s
f
o
r
m
e
r
r
aw.
T
h
e
r
esu
lts
s
h
o
w
a
co
n
s
is
ten
t
d
ec
lin
e
in
b
o
th
d
ata
f
o
r
b
o
th
tr
ain
i
n
g
an
d
v
alid
a
tio
n
MA
E
lo
s
s
o
v
er
th
e
ep
o
ch
s
,
in
d
icatin
g
th
e
m
o
d
el'
s
lear
n
in
g
p
r
o
g
r
ess
.
Ho
wev
er
,
it
is
im
p
o
r
tan
t
to
n
o
te
th
at
th
e
m
o
d
el
h
as
n
o
t
y
et
co
n
v
er
g
e
d
to
its
o
p
tim
al
p
er
f
o
r
m
an
ce
.
W
h
ile
th
e
d
ec
lin
e
in
v
alid
atio
n
lo
s
s
alo
n
g
s
id
e
tr
ain
in
g
lo
s
s
s
u
g
g
ests
p
r
o
m
is
in
g
g
en
er
aliza
tio
n
an
d
a
l
ac
k
o
f
o
v
er
f
itti
n
g
,
th
e
tr
an
s
f
o
r
m
e
r
'
s
p
er
f
o
r
m
an
ce
o
n
th
e
f
ir
s
t
d
er
iv
ati
v
e
d
ata
is
s
till
s
h
o
win
g
r
o
o
m
f
o
r
im
p
r
o
v
em
en
t
in
te
r
m
s
o
f
MA
E
an
d
R
MSE
v
al
u
es.
T
h
e
tr
an
s
f
o
r
m
er
m
o
d
el
a
ch
iev
e
d
n
o
tab
le
r
esu
lts
with
a
n
M
AE
o
f
4
.
4
2
a
n
d
a
n
R
MSE
o
f
5
.
4
2
,
b
u
t
it
h
asn
'
t
r
ea
ch
ed
its
co
n
v
er
g
e
n
ce
p
o
i
n
t.
I
t
is
wo
r
th
m
en
tio
n
in
g
th
at
t
h
e
tr
an
s
f
o
r
m
e
r
with
th
e
f
ir
s
t
d
er
iv
ativ
e
m
o
d
el
o
u
tp
er
f
o
r
m
s
th
e
b
aselin
e
m
o
d
els.
On
th
e
o
th
er
h
an
d
,
th
e
tr
a
n
s
f
o
r
m
er
m
o
d
el
u
s
in
g
r
aw
d
ata
s
h
o
ws
b
etter
r
esu
lts
co
m
p
ar
ed
to
th
e
b
aselin
e
ex
tr
em
e
g
r
ad
ien
t
b
o
o
s
tin
g
(
XGBo
o
s
t
)
m
o
d
el
in
te
r
m
s
o
f
R
MSE
,
b
u
t
it
is
s
lig
h
tly
b
eh
in
d
th
e
b
aselin
e
L
STM
+
t
r
a
n
s
f
o
r
m
er
m
o
d
el,
with
an
M
AE
o
f
5
.
4
1
an
d
an
R
MSE
o
f
6
.
0
2
.
T
h
e
ev
al
u
atio
n
r
esu
lts
ar
e
p
r
esen
ted
in
T
a
b
le
2
.
(
a)
(
b
)
Fig
u
r
e
6
.
Plo
ttin
g
r
esu
lts
o
f
th
e
tr
ain
in
g
u
s
in
g
tr
an
s
f
o
r
m
er
m
o
d
el
o
n
d
ata
(
a)
f
ir
s
t d
er
iv
ativ
e
an
d
(
b
)
r
aw
T
h
e
r
esu
lts
d
ep
icted
in
th
e
p
lo
ts
f
o
r
th
e
tr
an
s
f
o
r
m
er
m
o
d
el
ap
p
lied
to
th
e
f
ir
s
t
d
er
iv
ativ
e
an
d
r
aw
d
ata
d
em
o
n
s
tr
ate
s
ig
n
if
ican
t
f
in
d
in
g
s
.
T
h
e
tr
a
n
s
f
o
r
m
e
r
m
o
d
el’
s
p
er
f
o
r
m
an
ce
o
n
th
e
f
ir
s
t
d
er
iv
ativ
e
d
ata
is
n
o
tab
ly
s
u
p
er
io
r
,
ac
h
iev
in
g
a
lo
wer
MA
E
o
f
4
.
4
2
,
wh
ich
s
u
r
p
ass
es
th
e
b
en
ch
m
ar
k
m
o
d
el
s
.
T
h
is
lo
wer
MA
E
s
u
g
g
ests
s
tr
o
n
g
p
r
ed
ictiv
e
ac
cu
r
ac
y
a
n
d
co
u
ld
in
d
icate
t
h
a
t
p
r
ep
r
o
ce
s
s
in
g
d
ata
th
r
o
u
g
h
d
if
f
er
en
tiatio
n
m
ay
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
1
4
,
No
.
2
,
Ap
r
il 2
0
2
5
:
1
3
0
2
-
1
3
1
0
1308
h
elp
th
e
tr
an
s
f
o
r
m
er
m
o
d
el
c
ap
tu
r
e
u
n
d
e
r
ly
in
g
tr
en
d
s
m
o
r
e
ef
f
ec
tiv
ely
.
Fo
r
th
e
r
aw
d
a
ta,
th
e
tr
a
n
s
f
o
r
m
e
r
m
o
d
el
s
till
o
u
tp
er
f
o
r
m
s
th
e
X
GB
o
o
s
t
b
aselin
e
in
ter
m
s
o
f
R
MSE
b
u
t
d
o
es
n
o
t
e
x
ce
ed
th
e
L
STM
+tr
an
s
f
o
r
m
er
m
o
d
el'
s
p
er
f
o
r
m
an
ce
.
T
ab
le
2
.
C
o
m
p
a
r
is
o
n
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
with
th
e
b
aselin
e
m
o
d
el
P
a
r
a
me
t
e
r
M
A
E
R
M
S
E
R
2
Tr
a
n
sf
o
r
mer
r
a
w
d
a
t
a
(
o
u
r
m
o
d
e
l
)
5
.
4
1
6
.
0
2
0
.
2
0
Tr
a
n
sf
o
r
mer
f
i
r
s
t
d
e
r
i
v
a
t
i
v
e
(
o
u
r
m
o
d
e
l
)
4
.
4
2
5
.
4
2
0
.
2
2
X
G
B
o
o
st
f
i
r
st
d
e
r
i
v
a
t
i
v
e
[
2
3
]
5
.
4
1
6
.
3
2
0
.
0
2
LSTM
+
t
r
a
n
sf
o
r
m
e
r
r
a
w
d
a
t
a
[
1
8
]
4
.
8
3
5
.
7
6
-
4
.
2
.
Dis
cus
s
io
n
T
h
is
f
in
d
in
g
co
u
l
d
s
u
g
g
est
th
at
wh
ile
th
e
tr
an
s
f
o
r
m
er
m
o
d
el
is
r
o
b
u
s
t,
th
e
in
co
r
p
o
r
atio
n
o
f
L
STM
in
to
th
e
a
r
ch
itectu
r
e
m
ig
h
t
c
ap
tu
r
e
tem
p
o
r
al
d
ep
en
d
en
cie
s
in
th
e
r
aw
d
ata
m
o
r
e
e
f
f
e
ctiv
ely
,
wh
ich
ar
e
p
er
h
ap
s
less
p
r
o
n
o
u
n
ce
d
af
te
r
th
e
d
if
f
er
en
tiatio
n
p
r
o
ce
s
s
.
Su
ch
a
r
esu
lt
s
u
g
g
ests
th
at
u
s
in
g
f
ir
s
t
d
er
iv
ativ
e
d
ata
in
th
is
co
n
tex
t
m
ig
h
t
ca
p
tu
r
e
r
elev
an
t
in
f
o
r
m
atio
n
th
at
allo
ws
th
e
tr
an
s
f
o
r
m
er
m
o
d
el
to
p
r
ed
ict
m
o
r
e
ac
cu
r
ately
co
m
p
a
r
ed
to
its
b
aselin
e
co
u
n
ter
p
a
r
ts
o
r
ev
en
th
e
r
aw
d
ata.
I
t a
ls
o
v
alid
ates th
e
ef
f
ec
tiv
en
ess
o
f
th
e
tr
an
s
f
o
r
m
er
ar
ch
itectu
r
e
in
h
a
n
d
lin
g
th
is
ty
p
e
o
f
tim
e
-
s
er
ies
d
ata,
p
o
ten
tially
p
r
o
v
id
in
g
in
s
ig
h
ts
th
at
co
u
ld
b
e
lev
er
ag
ed
i
n
o
th
e
r
s
im
ilar
ap
p
licatio
n
s
.
Ho
wev
er
,
it
wo
u
l
d
b
e
im
p
o
r
tan
t
t
o
lo
o
k
at
ad
d
itio
n
al
m
etr
ics
an
d
co
n
d
u
ct
f
u
r
th
e
r
test
s
,
s
u
ch
as c
r
o
s
s
-
v
alid
atio
n
,
to
c
o
n
f
ir
m
th
e
s
e
f
in
d
in
g
s
an
d
en
s
u
r
e
th
e
m
o
d
el'
s
r
o
b
u
s
tn
ess
.
A
d
ee
p
er
an
al
y
s
is
o
f
t
h
e
f
i
r
s
t
d
er
iv
ativ
e
d
ata
r
ev
ea
ls
th
at
th
e
m
o
d
el
ef
f
ec
ti
v
ely
ca
p
tu
r
es
p
atter
n
s
o
f
ch
an
g
e
with
in
th
e
v
id
eo
s
,
as
ev
id
en
ce
d
b
y
th
e
r
elativ
ely
lo
w
v
alu
es
o
f
MA
E
an
d
R
MSE
.
Ho
wev
er
,
th
e
s
till
r
elativ
ely
lo
w
R
2
v
alu
es
in
d
icate
th
at
th
e
m
o
d
el
u
s
in
g
f
ir
s
t
d
er
iv
ativ
e
d
ata
ca
n
e
x
p
lain
s
li
g
h
tly
m
o
r
e
v
a
r
iatio
n
in
th
e
tar
g
et
d
ata
co
m
p
ar
ed
to
th
e
m
o
d
el
u
s
in
g
r
aw
d
ata.
T
h
is
m
ay
b
e
d
u
e
to
th
e
f
ir
s
t
d
er
iv
ativ
e
p
r
o
ce
s
s
's
r
o
le
in
r
ed
u
cin
g
n
o
is
e
o
r
f
lu
ctu
a
tio
n
s
p
r
esen
t
in
th
e
r
aw
v
id
eo
d
ata.
B
y
elim
in
atin
g
o
r
r
ed
u
cin
g
s
m
all
o
r
f
lu
ctu
atin
g
asp
ec
ts
th
at
m
ay
n
o
t b
e
r
elev
a
n
t,
th
e
m
o
d
el
ca
n
f
o
cu
s
m
o
r
e
o
n
lar
g
e
r
an
d
m
o
r
e
s
ig
n
if
ican
t p
atter
n
s
o
f
ch
an
g
e
.
Ad
d
itio
n
ally
,
th
e
f
ir
s
t
d
er
iv
ativ
e
m
ay
b
e
m
o
r
e
e
f
f
ec
tiv
e
in
h
ig
h
lig
h
ti
n
g
d
y
n
a
m
ic
ch
an
g
es
in
d
ata
o
v
er
tim
e,
p
r
o
v
id
in
g
an
ad
v
a
n
tag
e
in
d
etec
tin
g
s
m
all
ch
an
g
es
o
r
tr
en
d
s
th
at
m
ay
b
e
o
v
er
lo
o
k
ed
b
y
m
o
d
els
u
s
in
g
r
aw
d
ata.
Me
an
wh
ile,
wh
en
u
s
in
g
r
aw
d
ata,
th
e
m
o
d
el
s
tr
u
g
g
les
to
ca
p
tu
r
e
p
atter
n
s
th
at
m
a
y
ex
i
s
t
in
v
id
eo
ch
an
g
es.
T
h
is
is
r
ef
lecte
d
in
t
h
e
h
ig
h
er
MA
E
an
d
R
MSE
v
alu
es,
in
d
icatin
g
th
at
th
e
m
o
d
el
h
as
a
h
ig
h
er
lev
el
o
f
er
r
o
r
in
p
r
ed
ictin
g
th
e
s
ev
e
r
ity
o
f
d
ep
r
ess
io
n
lev
els
f
r
o
m
u
n
p
r
o
ce
s
s
ed
v
id
e
o
d
ata.
Ho
wev
er
,
f
o
r
R
2
m
etr
ics,
b
o
th
d
ata
ty
p
es
y
ield
r
elativ
ely
lo
w
v
alu
es.
T
h
is
m
ay
b
e
d
u
e
t
o
lim
itatio
n
s
o
n
th
e
am
o
u
n
t
o
f
d
ata,
as
T
r
an
s
f
o
r
m
e
r
s
o
f
ten
r
eq
u
ir
e
lar
g
e
am
o
u
n
ts
o
f
d
ata
f
o
r
ef
f
ec
ti
v
e
tr
ain
in
g
.
I
f
th
e
d
ataset
u
s
ed
is
r
elativ
ely
s
m
al
l
o
r
n
o
t
s
u
f
f
icien
tly
r
e
p
r
esen
tati
v
e
o
f
t
h
e
v
ar
iatio
n
s
th
at
m
ay
o
cc
u
r
in
th
e
p
o
p
u
latio
n
,
th
en
th
e
m
o
d
el
m
ay
n
o
t
g
en
er
alize
well
to
n
ew
d
ata
.
5.
CO
NCLU
SI
O
N
I
n
co
n
clu
s
io
n
,
o
u
r
s
tu
d
y
ex
p
lo
r
es
th
e
p
o
ten
tial
o
f
t
r
an
s
f
o
r
m
er
m
o
d
els
in
d
etec
tin
g
s
ig
n
s
o
f
d
ep
r
ess
io
n
th
r
o
u
g
h
th
e
an
aly
s
is
o
f
v
id
eo
d
ata,
with
a
s
p
ec
if
ic
f
o
cu
s
o
n
f
ac
ial
b
eh
av
io
r
as
a
r
ep
r
esen
tatio
n
o
f
MT
S.
E
x
p
er
im
en
tal
r
esu
lts
d
em
o
n
s
tr
ate
th
at
th
e
ap
p
lica
tio
n
o
f
t
h
e
t
r
an
s
f
o
r
m
er
m
o
d
el
to
v
id
e
o
d
ata,
p
ar
ticu
lar
ly
a
f
ter
p
r
o
ce
s
s
in
g
with
th
e
f
ir
s
t
d
e
r
iv
ativ
e,
y
ield
s
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
co
m
p
a
r
ed
to
r
aw
d
ata.
T
h
e
t
r
an
s
f
o
r
m
er
m
o
d
el
ex
h
ib
its
im
p
r
o
v
e
d
p
er
f
o
r
m
an
ce
m
etr
ic
s
,
in
clu
d
in
g
MA
E
o
f
4
.
4
2
,
R
MSE
o
f
5
.
4
2
,
a
n
d
R
2
v
alu
e
o
f
0
.
2
2
,
o
u
tp
e
r
f
o
r
m
in
g
b
aselin
e
m
o
d
els
an
d
u
n
d
er
s
co
r
in
g
its
ca
p
ab
ilit
y
to
ca
p
tu
r
e
r
elev
an
t
te
m
p
o
r
a
l
p
atter
n
s
f
o
r
e
n
h
an
ce
d
p
r
ed
ict
iv
e
ac
cu
r
ac
y
.
W
h
ile
th
e
s
tu
d
y
co
n
tr
ib
u
tes
s
ig
n
if
ica
n
tly
to
u
n
d
er
s
tan
d
in
g
th
e
p
o
ten
tial
o
f
t
r
an
s
f
o
r
m
er
m
o
d
els
in
d
ep
r
ess
io
n
d
etec
tio
n
th
r
o
u
g
h
v
id
eo
d
ata,
a
d
d
r
ess
in
g
lim
itatio
n
s
s
u
ch
as
co
m
p
u
tatio
n
al
r
e
q
u
ir
em
e
n
ts
an
d
d
ataset
av
ailab
ilit
y
is
es
s
en
tial.
Op
p
o
r
tu
n
ities
f
o
r
f
u
t
u
r
e
r
esear
ch
lie
in
ex
p
an
d
i
n
g
d
atasets
to
im
p
r
o
v
e
m
o
d
el
g
e
n
er
aliza
tio
n
,
i
n
teg
r
atin
g
f
in
d
in
g
s
with
m
en
tal
h
ea
lth
p
latf
o
r
m
s
o
r
o
n
lin
e
co
u
n
s
elin
g
s
er
v
ices,
an
d
ex
p
lo
r
in
g
alter
n
ativ
e
ap
p
r
o
a
ch
es,
s
u
ch
as
f
ea
tu
r
e
en
g
in
ee
r
in
g
,
to
e
n
h
an
ce
th
e
m
o
d
el'
s
ca
p
ab
ilit
ies
f
u
r
th
er
.
Ho
wev
er
,
lim
itatio
n
s
o
f
th
e
s
tu
d
y
in
v
o
lv
e
t
h
e
tr
an
s
f
o
r
m
a
tio
n
o
f
v
id
e
o
d
ata,
wh
ich
m
ay
n
o
t
f
u
lly
en
co
m
p
ass
co
m
p
lex
n
o
n
-
v
er
b
al
asp
ec
ts
,
an
d
th
e
n
ee
d
f
o
r
lar
g
er
d
atasets
to
en
h
an
c
e
m
o
d
el
g
en
er
aliza
tio
n
.
T
o
m
i
tig
ate
th
ese
lim
itatio
n
s
,
ef
f
o
r
ts
ar
e
r
eq
u
ir
ed
to
o
b
tai
n
b
r
o
ad
er
d
atasets
f
o
r
im
p
r
o
v
e
d
m
o
d
el
g
en
er
aliza
ti
o
n
.
Fu
tu
r
e
r
esear
c
h
o
p
p
o
r
tu
n
ities
in
clu
d
e
ex
p
an
d
i
n
g
d
a
tasets
,
in
teg
r
atin
g
r
esear
ch
f
in
d
in
g
s
with
m
en
tal
h
ea
lth
p
latf
o
r
m
s
o
r
o
n
lin
e
co
u
n
s
elin
g
s
er
v
ices,
an
d
c
o
n
d
u
ctin
g
r
esear
ch
with
o
u
t
u
s
in
g
f
ea
tu
r
e
s
elec
tio
n
to
m
ea
s
u
r
e
th
e
i
m
p
ac
t
o
f
th
e
p
r
o
ce
s
s
an
d
m
ain
tain
all
av
ail
ab
le
f
ea
tu
r
es
in
t
h
e
d
ataset.
Ad
d
itio
n
ally
,
th
e
d
e
v
elo
p
m
en
t
o
f
th
is
r
esear
ch
ca
n
b
e
en
h
an
ce
d
th
r
o
u
g
h
f
ea
tu
r
e
en
g
in
ee
r
in
g
,
wh
er
e
s
p
ec
if
ic
an
d
r
elev
an
t
f
ea
tu
r
e
s
s
u
ch
as
in
ter
o
cu
lar
d
is
tan
ce
,
lip
h
eig
h
t,
an
d
lip
wid
th
ca
n
p
r
o
v
i
d
e
m
o
r
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Dep
r
ess
io
n
d
etec
tio
n
th
r
o
u
g
h
tr
a
n
s
fo
r
mer
s
-
b
a
s
ed
emo
tio
n
r
ec
o
g
n
itio
n
i
n
…
(
K
en
jo
va
n
N
a
n
g
g
a
la
)
1309
in
f
o
r
m
ativ
e
d
ata
r
ep
r
esen
tatio
n
s
.
Featu
r
e
en
g
in
ee
r
in
g
o
f
f
er
s
s
ig
n
if
ican
t
ad
v
an
ta
g
es
o
v
er
p
r
ev
io
u
s
ap
p
r
o
ac
h
es,
allo
win
g
f
o
r
a
d
ee
p
er
u
n
d
e
r
s
tan
d
in
g
o
f
d
ep
r
ess
io
n
-
r
elate
d
p
atter
n
s
an
d
p
o
te
n
tially
im
p
r
o
v
in
g
th
e
m
o
d
el'
s
ab
ilit
y
to
d
etec
t d
ep
r
ess
io
n
s
ig
n
s
.
RE
F
E
R
E
NC
E
S
[
1
]
C
.
S
u
,
Z.
X
u
,
J
.
P
a
t
h
a
k
,
a
n
d
F
.
W
a
n
g
,
“
D
e
e
p
l
e
a
r
n
i
n
g
i
n
m
e
n
t
a
l
h
e
a
l
t
h
o
u
t
c
o
me
r
e
se
a
r
c
h
:
a
s
c
o
p
i
n
g
r
e
v
i
e
w
,
”
T
ra
n
sl
a
t
i
o
n
a
l
Psy
c
h
i
a
t
r
y
,
v
o
l
.
1
0
,
n
o
.
1
,
A
p
r
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
3
8
/
s
4
1
3
9
8
-
020
-
0
7
8
0
-
3.
[
2
]
L.
B
a
e
r
a
n
d
M
.
A
.
B
l
a
i
s,
H
a
n
d
b
o
o
k
o
f
c
l
i
n
i
c
a
l
ra
t
i
n
g
sc
a
l
e
s
a
n
d
a
ss
e
s
sm
e
n
t
i
n
p
s
y
c
h
i
a
t
r
y
a
n
d
m
e
n
t
a
l
h
e
a
l
t
h
.
T
o
t
o
w
a
,
N
J:
H
u
ma
n
a
P
r
e
ss,
2
0
1
0
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
1
-
5
9
7
4
5
-
3
8
7
-
5.
[
3
]
G
.
N
.
E
l
w
i
r
e
h
a
r
d
j
a
,
M
.
I
sn
a
n
,
A
.
S
.
P
e
r
b
a
n
g
s
a
,
K
.
M
u
c
h
t
a
r
,
a
n
d
B
.
P
a
r
d
a
m
e
a
n
,
“
Tr
e
n
d
s,
o
p
p
o
r
t
u
n
i
t
i
e
s
,
a
n
d
c
h
a
l
l
e
n
g
e
s
i
n
d
e
t
e
c
t
i
n
g
d
e
p
r
e
ss
i
v
e
d
i
so
r
d
e
r
s
t
h
r
o
u
g
h
m
o
b
i
l
e
d
e
v
i
c
e
s:
a
r
e
v
i
e
w
,
”
i
n
2
0
2
3
2
n
d
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
C
o
m
p
u
t
e
r
S
y
st
e
m
,
I
n
f
o
rm
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
,
a
n
d
El
e
c
t
r
i
c
a
l
E
n
g
i
n
e
e
ri
n
g
(
C
O
S
I
T
E)
,
A
u
g
.
2
0
2
3
,
p
p
.
1
8
8
–
193
,
d
o
i
:
1
0
.
1
1
0
9
/
C
O
S
I
TE6
0
2
3
3
.
2
0
2
3
.
1
0
2
4
9
8
5
9
.
[
4
]
M
.
B
r
i
l
e
y
a
n
d
J.
-
P
.
Lé
p
i
n
e
,
“
Th
e
i
n
c
r
e
a
si
n
g
b
u
r
d
e
n
o
f
d
e
p
r
e
ssi
o
n
,
”
N
e
u
r
o
p
sy
c
h
i
a
t
ri
c
D
i
s
e
a
s
e
a
n
d
T
r
e
a
t
m
e
n
t
,
v
o
l
.
7
,
n
o
.
1
,
M
a
y
2
0
1
1
,
d
o
i
:
1
0
.
2
1
4
7
/
N
D
T.
S
1
9
6
1
7
.
[
5
]
A
.
M
u
r
r
u
e
t
a
l
.
,
“
T
h
e
i
m
p
l
i
c
a
t
i
o
n
s
o
f
h
y
p
e
r
s
o
mn
i
a
i
n
t
h
e
c
o
n
t
e
x
t
o
f
maj
o
r
d
e
p
r
e
ssi
o
n
:
r
e
su
l
t
s
f
r
o
m
a
l
a
r
g
e
,
i
n
t
e
r
n
a
t
i
o
n
a
l
,
o
b
s
e
r
v
a
t
i
o
n
a
l
st
u
d
y
,
”
Eu
r
o
p
e
a
n
N
e
u
r
o
p
sy
c
h
o
p
h
a
rm
a
c
o
l
o
g
y
,
v
o
l
.
2
9
,
n
o
.
4
,
p
p
.
4
7
1
–
4
8
1
,
A
p
r
.
2
0
1
9
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
u
r
o
n
e
u
r
o
.
2
0
1
9
.
0
2
.
0
1
1
.
[
6
]
A
.
M
.
N
e
z
u
,
K
.
S
.
M
c
C
l
u
r
e
,
a
n
d
C
.
M
.
N
e
z
u
,
“
Th
e
a
ssessm
e
n
t
o
f
d
e
p
r
e
s
s
i
o
n
,
”
i
n
T
re
a
t
i
n
g
D
e
p
ress
i
o
n
,
W
i
l
e
y
,
p
p
.
2
4
–
5
1
,
2
0
1
5
,
d
o
i
:
1
0
.
1
0
0
2
/
9
7
8
1
1
1
9
1
1
4
4
8
2
.
c
h
2
.
[
7
]
A
.
G
i
l
l
i
o
z
,
J.
C
a
sas,
E.
M
u
g
e
l
l
i
n
i
,
a
n
d
O
.
A
.
K
h
a
l
e
d
,
“
O
v
e
r
v
i
e
w
o
f
t
h
e
t
r
a
n
sf
o
r
mer
-
b
a
se
d
m
o
d
e
l
s
f
o
r
N
LP
t
a
sk
s
,
”
2
0
2
0
1
5
t
h
C
o
n
f
e
re
n
c
e
o
n
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
I
n
f
o
rm
a
t
i
o
n
S
y
s
t
e
m
s (F
e
d
C
S
I
S
)
,
2
0
2
0
,
p
p
.
1
7
9
-
1
8
3
,
d
o
i
:
1
0
.
1
5
4
3
9
/
2
0
2
0
F
2
0
.
.
[
8
]
S
.
K
h
a
n
,
M
.
N
a
se
e
r
,
M
.
H
a
y
a
t
,
S
.
W
.
Za
m
i
r
,
F
.
S
.
K
h
a
n
,
a
n
d
M
.
S
h
a
h
,
“
Tr
a
n
sf
o
r
m
e
r
s
i
n
v
i
s
i
o
n
:
a
s
u
r
v
e
y
,
”
AC
M
C
o
m
p
u
t
i
n
g
S
u
r
v
e
y
s
,
v
o
l
.
5
4
,
n
o
.
1
0
s,
p
p
.
1
–
4
1
,
J
a
n
.
2
0
2
2
,
d
o
i
:
1
0
.
1
1
4
5
/
3
5
0
5
2
4
4
.
[
9
]
K
.
N
a
n
g
g
a
l
a
,
G
.
N
.
El
w
i
r
e
h
a
r
d
j
a
,
a
n
d
B
.
P
a
r
d
a
m
e
a
n
,
“
S
y
st
e
ma
t
i
c
l
i
t
e
r
a
t
u
r
e
r
e
v
i
e
w
o
f
t
r
a
n
sf
o
r
mer
m
o
d
e
l
i
mp
l
e
me
n
t
a
t
i
o
n
s
i
n
d
e
t
e
c
t
i
n
g
d
e
p
r
e
s
si
o
n
,
”
i
n
2
0
2
3
6
t
h
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
f
C
o
m
p
u
t
e
r
a
n
d
I
n
f
o
rm
a
t
i
c
s
E
n
g
i
n
e
e
ri
n
g
(
I
C
2
I
E)
,
p
p
.
2
0
3
–
2
0
8
,
S
e
p
.
2
0
2
3
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
2
I
E6
0
5
4
7
.
2
0
2
3
.
1
0
3
3
1
4
4
8
.
[
1
0
]
T.
W
.
C
e
n
g
g
o
r
o
,
B
.
M
a
h
e
sw
o
r
o
,
A
.
B
u
d
i
a
r
t
o
,
J.
B
a
u
r
l
e
y
,
T
.
S
u
p
a
r
y
a
n
t
o
,
a
n
d
B
.
P
a
r
d
a
m
e
a
n
,
“
F
e
a
t
u
r
e
s
i
m
p
o
r
t
a
n
c
e
i
n
c
l
a
ss
i
f
i
c
a
t
i
o
n
m
o
d
e
l
s
f
o
r
c
o
l
o
r
e
c
t
a
l
c
a
n
c
e
r
c
a
s
e
s
p
h
e
n
o
t
y
p
e
i
n
I
n
d
o
n
e
si
a
,
”
Pr
o
c
e
d
i
a
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
1
5
7
,
p
p
.
3
1
3
–
3
2
0
,
2
0
1
9
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
p
r
o
c
s.
2
0
1
9
.
0
8
.
1
7
2
.
[
1
1
]
L.
K
o
n
g
,
J.
Y
u
,
D
.
Ta
n
g
,
Y
.
S
o
n
g
,
a
n
d
D
.
H
a
n
,
“
M
u
l
t
i
v
a
r
i
a
t
e
t
i
m
e
ser
i
e
s
a
n
o
m
a
l
y
d
e
t
e
c
t
i
o
n
w
i
t
h
g
e
n
e
r
a
t
i
v
e
a
d
v
e
r
sari
a
l
n
e
t
w
o
r
k
s
b
a
s
e
d
o
n
a
c
t
i
v
e
d
i
s
t
o
r
t
i
o
n
t
r
a
n
sf
o
r
mer,
”
I
EE
E
S
e
n
s
o
rs
J
o
u
r
n
a
l
,
v
o
l
.
2
3
,
n
o
.
9
,
p
p
.
9
6
5
8
–
9
6
6
8
,
M
a
y
2
0
2
3
,
d
o
i
:
1
0
.
1
1
0
9
/
JS
EN
.
2
0
2
3
.
3
2
6
0
5
6
3
.
[
1
2
]
F
.
Ze
n
g
,
M
.
C
h
e
n
,
C
.
Q
i
a
n
,
Y
.
W
a
n
g
,
Y
.
Z
h
o
u
,
a
n
d
W
.
Ta
n
g
,
“
M
u
l
t
i
v
a
r
i
a
t
e
t
i
m
e
seri
e
s
a
n
o
ma
l
y
d
e
t
e
c
t
i
o
n
w
i
t
h
a
d
v
e
r
sari
a
l
t
r
a
n
sf
o
r
mer
a
r
c
h
i
t
e
c
t
u
r
e
i
n
t
h
e
i
n
t
e
r
n
e
t
o
f
t
h
i
n
g
s,
”
Fu
t
u
re
G
e
n
e
r
a
t
i
o
n
C
o
m
p
u
t
e
r
S
y
s
t
e
m
s
,
v
o
l
.
1
4
4
,
p
p
.
2
4
4
–
2
5
5
,
Ju
l
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
f
u
t
u
r
e
.
2
0
2
3
.
0
2
.
0
1
5
.
[
1
3
]
M
.
G
a
v
r
i
l
e
s
c
u
a
n
d
N
.
V
i
z
i
r
e
a
n
u
,
“
P
r
e
d
i
c
t
i
n
g
d
e
p
r
e
ss
i
o
n
,
a
n
x
i
e
t
y
,
a
n
d
s
t
r
e
ss
l
e
v
e
l
s
f
r
o
m
v
i
d
e
o
s
u
si
n
g
t
h
e
f
a
c
i
a
l
a
c
t
i
o
n
c
o
d
i
n
g
sy
st
e
m,
”
S
e
n
so
rs
,
v
o
l
.
1
9
,
n
o
.
1
7
,
A
u
g
.
2
0
1
9
,
d
o
i
:
1
0
.
3
3
9
0
/
s
1
9
1
7
3
6
9
3
.
[
1
4
]
M
.
M
u
z
a
mm
e
l
,
H
.
S
a
l
a
m
,
a
n
d
A
.
O
t
h
man
i
,
“
E
n
d
-
to
-
e
n
d
m
u
l
t
i
m
o
d
a
l
c
l
i
n
i
c
a
l
d
e
p
r
e
ss
i
o
n
r
e
c
o
g
n
i
t
i
o
n
u
s
i
n
g
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
s:
a
c
o
m
p
a
r
a
t
i
v
e
a
n
a
l
y
si
s,
”
C
o
m
p
u
t
e
r
M
e
t
h
o
d
s
a
n
d
Pr
o
g
r
a
m
s
i
n
B
i
o
m
e
d
i
c
i
n
e
,
v
o
l
.
2
1
1
,
N
o
v
.
2
0
2
1
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
m
p
b
.
2
0
2
1
.
1
0
6
4
3
3
.
[
1
5
]
B
.
G
r
i
mm
,
B
.
Ta
l
b
o
t
,
a
n
d
L.
L
a
r
sen
,
“
P
H
Q
-
V
/
G
A
D
-
V
:
a
ssessm
e
n
t
s
t
o
i
d
e
n
t
i
f
y
s
i
g
n
a
l
s
o
f
d
e
p
r
e
ssi
o
n
a
n
d
a
n
x
i
e
t
y
f
r
o
m
p
a
t
i
e
n
t
v
i
d
e
o
r
e
s
p
o
n
ses,
”
A
p
p
l
i
e
d
S
c
i
e
n
c
e
s
,
v
o
l
.
1
2
,
n
o
.
1
8
,
S
e
p
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
a
p
p
1
2
1
8
9
1
5
0
.
[
1
6
]
G
.
S
u
n
,
S
.
Zh
a
o
,
B
.
Zo
u
,
a
n
d
Y
.
A
n
,
“
M
u
l
t
i
mo
d
a
l
d
e
p
r
e
ss
i
o
n
d
e
t
e
c
t
i
o
n
u
s
i
n
g
a
d
e
e
p
f
e
a
t
u
r
e
f
u
si
o
n
n
e
t
w
o
r
k
,
”
i
n
T
h
i
rd
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
C
o
m
m
u
n
i
c
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
(
I
C
C
S
C
T
2
0
2
2
)
,
D
e
c
.
2
0
2
2
,
d
o
i
:
1
0
.
1
1
1
7
/
1
2
.
2
6
6
2
6
2
0
.
[
1
7
]
H
.
S
u
n
e
t
a
l
.
,
“
M
u
l
t
i
-
mo
d
a
l
a
d
a
p
t
i
v
e
f
u
si
o
n
t
r
a
n
sf
o
r
mer
n
e
t
w
o
r
k
f
o
r
t
h
e
e
s
t
i
m
a
t
i
o
n
o
f
d
e
p
r
e
ssi
o
n
l
e
v
e
l
,
”
S
e
n
s
o
rs
,
v
o
l
.
2
1
,
n
o
.
1
4
,
Ju
l
.
2
0
2
1
,
d
o
i
:
1
0
.
3
3
9
0
/
s2
1
1
4
4
7
6
4
.
[
1
8
]
S
.
R
a
s
i
p
u
r
a
m
,
J
.
H
.
B
h
a
t
,
A
.
M
a
i
t
r
a
,
B
.
S
h
a
w
,
a
n
d
S
.
S
a
h
a
,
“
M
u
l
t
i
m
o
d
a
l
d
e
p
r
e
ssi
o
n
d
e
t
e
c
t
i
o
n
u
si
n
g
t
a
s
k
-
o
r
i
e
n
t
e
d
t
r
a
n
sf
o
r
mer
-
b
a
s
e
d
e
mb
e
d
d
i
n
g
,
”
i
n
2
0
2
2
I
EE
E
S
y
m
p
o
s
i
u
m
o
n
C
o
m
p
u
t
e
rs
a
n
d
C
o
m
m
u
n
i
c
a
t
i
o
n
s
(
I
S
C
C
)
,
p
p
.
1
–
4
,
Ju
n
.
2
0
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
I
S
C
C
5
5
5
2
8
.
2
0
2
2
.
9
9
1
3
0
4
4
.
[
1
9
]
N
.
P
.
Ti
g
g
a
a
n
d
S
.
G
a
r
g
,
“
Ef
f
i
c
a
c
y
o
f
n
o
v
e
l
a
t
t
e
n
t
i
o
n
-
b
a
se
d
g
a
t
e
d
r
e
c
u
r
r
e
n
t
u
n
i
t
s
t
r
a
n
sf
o
r
mer
f
o
r
d
e
p
r
e
ss
i
o
n
d
e
t
e
c
t
i
o
n
u
s
i
n
g
e
l
e
c
t
r
o
e
n
c
e
p
h
a
l
o
g
r
a
m
si
g
n
a
l
s,”
H
e
a
l
t
h
I
n
f
o
rm
a
t
i
o
n
S
c
i
e
n
c
e
a
n
d
S
y
st
e
m
s
,
v
o
l
.
1
1
,
n
o
.
1
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
3
7
5
5
-
0
2
2
-
0
0
2
0
5
-
8.
[
2
0
]
G
.
Ti
w
a
r
y
,
S
.
C
h
a
u
h
a
n
,
a
n
d
K
.
K
.
G
o
y
a
l
,
“
V
i
d
e
o
b
a
se
d
d
e
e
p
C
N
N
m
o
d
e
l
f
o
r
d
e
p
r
e
ssi
o
n
d
e
t
e
c
t
i
o
n
,
”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
n
Re
c
e
n
t
a
n
d
I
n
n
o
v
a
t
i
o
n
T
re
n
d
s
i
n
C
o
m
p
u
t
i
n
g
a
n
d
C
o
m
m
u
n
i
c
a
t
i
o
n
,
v
o
l
.
1
0
,
n
o
.
1
0
,
p
p
.
5
9
–
6
4
,
O
c
t
.
2
0
2
2
,
d
o
i
:
1
0
.
1
7
7
6
2
/
i
j
r
i
t
c
c
.
v
1
0
i
1
0
.
5
7
3
5
.
[
2
1
]
Z.
S
h
a
n
g
g
u
a
n
,
Z
.
Li
u
,
G
.
L
i
,
Q
.
C
h
e
n
,
Z.
D
i
n
g
,
a
n
d
B
.
H
u
,
“
D
u
a
l
-
st
r
e
a
m
m
u
l
t
i
p
l
e
i
n
st
a
n
c
e
l
e
a
r
n
i
n
g
f
o
r
d
e
p
r
e
ssi
o
n
d
e
t
e
c
t
i
o
n
w
i
t
h
f
a
c
i
a
l
e
x
p
r
e
ssi
o
n
v
i
d
e
o
s
,
”
I
EEE
T
r
a
n
s
a
c
t
i
o
n
s
o
n
N
e
u
r
a
l
S
y
s
t
e
m
s
a
n
d
Re
h
a
b
i
l
i
t
a
t
i
o
n
E
n
g
i
n
e
e
r
i
n
g
,
v
o
l
.
3
1
,
p
p
.
5
5
4
–
5
6
3
,
2
0
2
3
,
d
o
i
:
1
0
.
1
1
0
9
/
TN
S
R
E.
2
0
2
2
.
3
2
0
4
7
5
7
.
[
2
2
]
Y
.
G
u
o
,
C
.
Zh
u
,
S
.
H
a
o
,
a
n
d
R
.
H
o
n
g
,
“
A
u
t
o
m
a
t
i
c
d
e
p
r
e
ssi
o
n
d
e
t
e
c
t
i
o
n
v
i
a
l
e
a
r
n
i
n
g
a
n
d
f
u
si
n
g
f
e
a
t
u
r
e
s
f
r
o
m
v
i
s
u
a
l
c
u
e
s,”
I
EE
E
T
ra
n
s
a
c
t
i
o
n
s
o
n
C
o
m
p
u
t
a
t
i
o
n
a
l
S
o
c
i
a
l
S
y
st
e
m
s
,
v
o
l
.
1
0
,
n
o
.
5
,
p
p
.
2
8
0
6
–
2
8
1
3
,
O
c
t
.
2
0
2
3
,
d
o
i
:
1
0
.
1
1
0
9
/
T
C
S
S
.
2
0
2
2
.
3
2
0
2
3
1
6
.
[
2
3
]
B
.
N
.
R
u
ma
h
o
r
b
o
,
K
.
N
a
n
g
g
a
l
a
,
G
.
N
.
E
l
w
i
r
e
h
a
r
d
j
a
,
a
n
d
B
.
P
a
r
d
a
mea
n
,
“
A
n
a
l
y
z
i
n
g
i
mp
o
r
t
a
n
t
s
t
a
t
i
s
t
i
c
a
l
f
e
a
t
u
r
e
s
f
r
o
m
f
a
c
i
a
l
b
e
h
a
v
i
o
r
i
n
h
u
m
a
n
d
e
p
r
e
ss
i
o
n
u
si
n
g
X
G
B
o
o
st
,
”
C
o
m
m
u
n
i
c
a
t
i
o
n
s i
n
Ma
t
h
e
m
a
t
i
c
a
l
B
i
o
l
o
g
y
a
n
d
N
e
u
r
o
sc
i
e
n
c
e
,
v
o
l
.
2
0
2
3
,
p
p
.
1
-
2
4
,
2
0
2
3
,
d
o
i
:
1
0
.
2
8
9
1
9
/
c
m
b
n
/
7
9
1
6
.
[
2
4
]
P
.
Ta
n
g
,
Q
.
Z
h
a
n
g
,
a
n
d
X
.
Zh
a
n
g
,
“
A
r
e
c
u
r
r
e
n
t
n
e
u
r
a
l
n
e
t
w
o
r
k
b
a
s
e
d
g
e
n
e
r
a
t
i
v
e
a
d
v
e
r
s
a
r
i
a
l
n
e
t
w
o
r
k
f
o
r
l
o
n
g
mu
l
t
i
v
a
r
i
a
t
e
t
i
me
seri
e
s
f
o
r
e
c
a
s
t
i
n
g
,
”
i
n
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
2
0
2
3
A
C
M
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
M
u
l
t
i
m
e
d
i
a
R
e
t
r
i
e
v
a
l
,
p
p
.
1
8
1
–
1
8
9
,
Ju
n
.
2
0
2
3
,
d
o
i
:
1
0
.
1
1
4
5
/
3
5
9
1
1
0
6
.
3
5
9
2
3
0
6
.
[
2
5
]
K
.
M
u
c
h
t
a
r
,
F
.
R
a
h
m
a
n
,
T.
W
.
C
e
n
g
g
o
r
o
,
A
.
B
u
d
i
a
r
t
o
,
a
n
d
B
.
P
a
r
d
a
me
a
n
,
“
A
n
i
m
p
r
o
v
e
d
v
e
r
s
i
o
n
o
f
t
e
x
t
u
r
e
-
b
a
s
e
d
f
o
r
e
g
r
o
u
n
d
seg
m
e
n
t
a
t
i
o
n
:
b
l
o
c
k
-
b
a
se
d
a
d
a
p
t
i
v
e
se
g
me
n
t
e
r
,
”
Pr
o
c
e
d
i
a
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
1
3
5
,
p
p
.
5
7
9
–
5
8
6
,
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
p
r
o
c
s
.
2
0
1
8
.
0
8
.
2
2
8
.
[2
6
]
R
.
Z
h
i
,
M
.
L
i
u
,
a
n
d
D
.
Z
h
a
n
g
,
“
A
c
o
mp
r
e
h
e
n
si
v
e
su
r
v
e
y
o
n
a
u
t
o
ma
t
i
c
f
a
c
i
a
l
a
c
t
i
o
n
u
n
i
t
a
n
a
l
y
s
i
s,
”
T
h
e
V
i
su
a
l
C
o
m
p
u
t
e
r
,
v
o
l
.
3
6
,
n
o
.
5
,
p
p
.
1
0
6
7
–
1
0
9
3
,
M
a
y
2
0
2
0
,
d
o
i
:
1
0
.
1
0
0
7
/
s
0
0
3
7
1
-
0
1
9
-
0
1
7
0
7
-
5.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
1
4
,
No
.
2
,
Ap
r
il 2
0
2
5
:
1
3
0
2
-
1
3
1
0
1310
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
K
e
n
jo
v
a
n
Na
n
g
g
a
la
is
a
M
a
ste
r
o
f
Co
m
p
u
ter
S
c
ien
c
e
st
u
d
e
n
t
a
t
Bin
a
Nu
sa
n
tara
Un
iv
e
rsity
,
In
d
o
n
e
sia
.
He
e
n
ro
ll
e
d
in
2
0
1
9
a
n
d
is
a
c
ti
v
e
l
y
p
a
rti
c
ip
a
ti
n
g
in
t
h
e
u
n
iv
e
rsit
y
'
s
m
a
ste
r
trac
k
p
ro
g
ra
m
a
s
p
a
rt
o
f
h
is
a
c
a
d
e
m
ic
jo
u
r
n
e
y
to
wa
rd
o
b
tai
n
in
g
a
m
a
ste
r'
s
d
e
g
re
e
.
His
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
a
rti
fici
a
l
in
telli
g
e
n
c
e
,
m
a
c
h
in
e
lea
rn
in
g
,
d
e
e
p
lea
rn
in
g
,
a
n
d
m
e
n
tal
h
e
a
lt
h
.
He
c
a
n
b
e
c
o
n
tac
ted
at
e
m
a
il
:
k
e
n
j
o
v
a
n
.
n
a
n
g
g
a
la@
b
in
u
s.a
c
.
id
.
G
r
e
g
o
r
iu
s
Na
ta
n
a
e
l
E
lwire
h
a
r
d
ja
is
a
d
ju
n
c
t
re
se
a
rc
h
fro
m
Bio
in
f
o
rm
a
ti
c
s
&
Da
ta
S
c
ien
c
e
Re
se
a
rc
h
Ce
n
ter
a
n
d
NV
IDIA
-
BINU
S
Artifi
c
ial
In
telli
g
e
n
c
e
Re
se
a
rc
h
&
De
v
e
lo
p
m
e
n
t
Ce
n
ter
o
f
B
in
a
Nu
s
a
n
tara
Un
iv
e
rsity
,
a
n
d
a
c
e
rti
fied
in
stru
c
t
o
r
a
t
NV
IDIA
d
e
e
p
lea
rn
in
g
i
n
stit
u
te
.
He
c
o
m
p
lete
d
b
o
t
h
h
is
b
a
c
h
e
l
o
r
a
n
d
m
a
ste
r
lev
e
l
e
d
u
c
a
ti
o
n
i
n
Bin
a
Nu
sa
n
tara
Un
iv
e
rsity
.
His
re
se
a
rc
h
in
tere
sts
in
c
l
u
d
e
a
p
p
li
e
d
m
a
c
h
i
n
e
lea
rn
in
g
in
v
a
rio
u
s
field
s
in
c
lu
d
in
g
,
b
u
t
a
re
n
o
t
li
m
it
e
d
to
,
c
o
m
p
u
ter
v
isio
n
,
m
e
n
tal
h
e
a
lt
h
,
a
n
d
n
a
t
u
ra
l
la
n
g
u
a
g
e
p
ro
c
e
ss
in
g
.
He
c
a
n
b
e
c
o
n
tac
ted
at
e
m
a
il
:
g
re
g
o
r
iu
s.e
lwireh
a
rd
ja@
b
in
u
s.a
c
.
id
.
Be
n
s
Pa
r
d
a
m
e
a
n
h
a
s
o
v
e
r
fo
rt
y
y
e
a
rs
o
f
g
lo
b
a
l
e
x
p
e
rien
c
e
in
i
n
fo
rm
a
ti
o
n
tec
h
n
o
l
o
g
y
,
b
io
i
n
fo
rm
a
ti
c
s,
a
n
d
e
d
u
c
a
ti
o
n
.
His
p
r
o
fe
ss
io
n
a
l
e
x
p
e
rien
c
e
in
c
lu
d
e
s
b
e
in
g
a
p
ra
c
ti
ti
o
n
e
r,
re
se
a
rc
h
e
r,
c
o
n
su
lt
a
n
t,
e
n
trep
re
n
e
u
r,
a
n
d
lec
tu
re
r.
He
c
u
rre
n
tl
y
h
o
l
d
s
a
d
u
a
l
a
p
p
o
i
n
tme
n
t
a
s
Dire
c
to
r
o
f
Bi
o
i
n
fo
rm
a
ti
c
s
&
Da
ta
S
c
ien
c
e
Re
se
a
rc
h
Ce
n
ter
(BD
S
RC)
|
AI
Re
se
a
rc
h
&
De
v
e
lo
p
m
e
n
t
Ce
n
t
e
r
(AIRD
C)
,
a
n
d
P
r
o
fe
ss
o
r
o
f
Co
m
p
u
ter
S
c
ie
n
c
e
a
t
Bin
a
Nu
sa
n
tara
(BINU
S
)
Un
iv
e
rsity
in
Ja
k
a
rta,
In
d
o
n
e
sia
.
He
e
a
rn
e
d
a
d
o
c
t
o
ra
l
d
e
g
re
e
in
in
fo
rm
a
ti
c
s
re
se
a
rc
h
fro
m
U
n
iv
e
rsity
o
f
S
o
u
th
e
rn
Ca
li
fo
r
n
ia
(U
S
C),
a
s
we
ll
a
s
a
m
a
ste
r’s
d
e
g
re
e
in
c
o
m
p
u
ter
e
d
u
c
a
ti
o
n
a
n
d
a
b
a
c
h
e
lo
r’s
d
e
g
re
e
in
c
o
m
p
u
t
e
r
sc
ien
c
e
fro
m
C
a
l
ifo
rn
ia
S
tate
Un
i
v
e
rsity
a
t
Lo
s
An
g
e
les
(USA).
He
c
a
n
b
e
c
o
n
tac
ted
at
e
m
a
il
:
b
p
a
rd
a
m
e
a
n
@b
i
n
u
s.e
d
u
.
Evaluation Warning : The document was created with Spire.PDF for Python.