I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
p
ute
r
E
ng
in
ee
ring
(
I
J
E
CE
)
Vo
l.
10
,
No
.
6
,
Decem
b
er
2
0
2
0
,
p
p
.
6
2
8
3
~
6
2
9
1
I
SS
N:
2088
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v
10
i
6
.
pp
6
2
8
3
-
6
2
9
1
6283
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m/in
d
ex
.
p
h
p
/I
JE
C
E
Identifica
tion
o
f
i
ntersti
tial lung
di
sea
ses
using
deep
lea
rning
Nidh
i
n Ra
j
u,
Anita
H
.
B
.
,
P
e
t
er
Aug
us
t
ine
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter S
c
ien
c
e
,
CHRIST
(De
e
m
e
d
to
b
e
Un
iv
e
rsity
),
In
d
ia
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Dec
1
7
,
2
0
1
9
R
ev
i
s
ed
Ma
y
2
8
,
2020
A
cc
ep
ted
J
u
n
6
,
2
0
2
0
T
h
e
a
d
v
a
n
c
e
d
m
e
d
ica
l
im
a
g
in
g
p
ro
v
id
e
s
v
a
rio
u
s
a
d
v
a
n
tag
e
s
to
b
o
th
th
e
p
a
ti
e
n
ts
a
n
d
t
h
e
h
e
a
lt
h
c
a
re
p
ro
v
i
d
e
rs.
M
e
d
ica
l
Im
a
g
in
g
tru
ly
h
e
lp
s
th
e
d
o
c
t
o
r
t
o
d
e
term
in
e
th
e
in
c
o
n
v
e
n
ien
c
e
s
in
a
h
u
m
a
n
b
o
d
y
a
n
d
e
m
p
o
w
e
rs
th
e
m
to
m
a
k
e
b
e
tt
e
r
c
h
o
ice
s.
De
e
p
lea
rn
in
g
h
a
s
a
n
im
p
o
rtan
t
ro
le
i
n
th
e
m
e
d
ica
l
f
ield
e
sp
e
c
iall
y
f
o
r
m
e
d
ica
l
i
m
a
g
e
a
n
a
l
y
sis
to
d
a
y
.
It
is
a
n
a
d
v
a
n
c
e
d
tec
h
n
iq
u
e
in
th
e
m
a
c
h
in
e
lea
rn
in
g
c
o
n
c
e
p
t
w
h
ich
c
a
n
b
e
u
se
d
to
g
e
t
e
ff
ici
e
n
t
o
u
tp
u
t
th
a
n
u
si
n
g
a
n
y
o
th
e
r
p
re
v
io
u
s
tec
h
n
iq
u
e
s.
I
n
t
h
e
a
n
ti
c
ip
a
ted
w
o
rk
d
e
e
p
lea
rn
in
g
is
u
se
d
to
f
in
d
t
h
e
p
re
se
n
c
e
o
f
in
ters
ti
ti
a
l
lu
n
g
d
ise
a
se
s
(IL
D)
b
y
a
n
a
l
y
z
in
g
h
ig
h
-
re
so
lu
ti
o
n
c
o
m
p
u
ted
to
m
o
g
ra
p
h
y
(HRC
T
)
i
m
a
g
e
s
a
n
d
id
e
n
ti
fy
in
g
th
e
IL
D
c
a
te
g
o
ry
.
T
h
e
e
ff
icie
n
c
y
o
f
th
e
d
iag
n
o
sis
o
f
ILD
th
ro
u
g
h
c
li
n
ica
l
h
isto
ry
is
les
s
th
a
n
2
0
%
.
Cu
rre
n
tl
y
,
a
n
o
p
e
n
c
h
e
st
b
i
o
p
sy
is
th
e
b
e
st
w
a
y
o
f
c
o
n
f
ir
m
in
g
th
e
p
re
se
n
c
e
o
f
IL
D.
HRC
T
i
m
a
g
e
s
c
a
n
b
e
u
se
d
e
ffe
c
ti
v
e
l
y
to
a
v
o
id
o
p
e
n
c
h
e
s
t
b
io
p
sy
a
n
d
im
p
ro
v
e
a
c
c
u
ra
c
y
.
In
th
is
p
ro
p
o
se
d
w
o
rk
m
u
lt
i
-
lab
e
l
c
las
sifica
ti
o
n
is d
o
n
e
f
o
r
1
7
d
if
f
e
r
e
n
t
c
a
teg
o
ries
o
f
IL
D.
T
h
e
a
v
e
r
a
g
e
a
c
c
u
ra
c
y
o
f
9
5
%
is
o
b
tai
n
e
d
b
y
e
x
trac
ti
n
g
f
e
a
tu
re
s
w
it
h
th
e
h
e
l
p
o
f
a
c
o
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
(CNN
)
a
rc
h
it
e
c
tu
re
c
a
ll
e
d
S
m
a
ll
e
rV
GG
Ne
t.
K
ey
w
o
r
d
s
:
C
las
s
i
f
icatio
n
Dee
p
lear
n
in
g
Me
d
ical
i
m
a
g
e
an
al
y
s
i
s
Neu
r
al
n
et
w
o
r
k
Co
p
y
rig
h
t
©
2
0
2
0
In
stit
u
te o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
rin
g
a
n
d
S
c
ien
c
e
.
Al
l
rig
h
ts re
se
rv
e
d
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Nid
h
i
n
R
aj
u
,
Dep
ar
t
m
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
,
C
HR
I
ST
(
Dee
m
ed
to
b
e
Un
iv
er
s
it
y
)
,
B
en
g
al
u
r
u
,
Kar
n
a
tak
a,
5
6
0
0
2
9
-
I
n
d
ia
.
E
m
ail:
n
id
h
i
n
.
r
aj
u
@
r
es.c
h
r
is
t
u
n
i
v
er
s
i
t
y
.
i
n
1.
I
NT
RO
D
UCT
I
O
N
Dee
p
L
ea
r
n
i
n
g
i
s
an
e
x
ce
lle
n
t
to
o
l
f
o
r
f
ea
t
u
r
e
lear
n
i
n
g
,
class
i
f
y
in
g
,
id
en
t
if
y
i
n
g
,
a
n
d
q
u
an
ti
f
y
i
n
g
p
atter
n
s
in
m
ed
ical
i
m
a
g
es.
T
h
e
p
r
o
g
r
ess
io
n
s
o
f
ad
v
a
n
ce
d
g
r
ap
h
ics
p
r
o
ce
s
s
in
g
u
n
its
(
GP
Us)
an
d
ce
n
tr
al
p
r
o
ce
s
s
in
g
u
n
its
(
C
P
Us),
a
lar
g
e
v
o
lu
m
e
o
f
d
ataset
a
v
ai
lab
il
it
y
,
an
d
th
e
d
ev
elo
p
m
en
t
o
f
m
an
y
d
ee
p
lear
n
in
g
alg
o
r
ith
m
s
ar
e
s
o
m
e
f
ac
to
r
s
t
o
th
e
s
u
cc
es
s
o
f
d
ee
p
lear
n
i
n
g
in
e
v
er
y
f
ield
.
T
h
e
d
ee
p
lear
n
in
g
tech
n
iq
u
e
s
ca
n
ef
f
ec
tiv
e
l
y
u
s
e
w
h
e
n
th
er
e
is
a
m
as
s
i
v
e
a
m
o
u
n
t
o
f
s
a
m
p
l
e
d
ata
ar
e
av
ailab
le
in
th
e
tr
ain
i
n
g
s
ta
g
e
its
e
lf
.
T
h
e
s
u
f
f
icien
t
i
m
p
le
m
e
n
tatio
n
o
f
Dee
p
L
ea
r
n
i
n
g
is
i
n
t
h
e
e
x
ten
s
i
v
e
d
ata
s
et
o
n
l
y
.
T
h
is
is
t
h
e
m
ai
n
ch
alle
n
g
i
n
g
ta
s
k
f
o
r
ap
p
ly
in
g
d
ee
p
lear
n
in
g
i
n
m
ed
ical
i
m
a
g
es
to
b
u
i
ld
a
d
ee
p
lear
n
i
n
g
m
o
d
el
w
ith
a
m
i
n
i
m
al
a
m
o
u
n
t
o
f
tr
ai
n
in
g
d
ata
[
1
]
.
T
h
e
Me
d
i
ca
l
I
m
ag
in
g
m
et
h
o
d
o
lo
g
y
is
ef
f
o
r
tles
s
,
n
o
n
-
o
b
tr
u
s
i
v
e,
a
n
d
th
e
v
a
s
t
m
aj
o
r
ity
o
f
t
h
e
m
d
o
n
o
t
r
eq
u
ir
e
a
n
y
e
x
tr
ao
r
d
in
ar
y
ar
r
an
g
e
m
en
t.
Ge
n
er
all
y
,
to
p
r
o
v
id
e
in
p
u
t
to
th
e
d
ee
p
lear
n
in
g
m
o
d
els,
t
w
o
t
y
p
es
o
f
co
n
ce
p
ts
ar
e
u
s
in
g
;
in
th
e
f
ir
s
t
ca
s
e,
th
e
v
ec
to
r
t
y
p
e
v
al
u
e
s
w
il
l
b
e
tak
en
f
o
r
m
u
lti
-
la
y
er
n
eu
r
al
n
et
w
o
r
k
s
,
an
d
in
an
o
t
h
er
ca
s
e,
2
D
o
r
3
D
i
m
a
g
e
v
al
u
es
w
i
ll
b
e
tak
e
n
f
o
r
th
e
co
n
v
o
l
u
tio
n
a
l
n
et
w
o
r
k
s
.
I
n
th
is
p
r
o
p
o
s
ed
w
o
r
k
,
a
C
NN
ar
ch
itectu
r
e
ca
lled
s
m
al
ler
VGGN
et,
w
h
ich
is
a
s
u
b
lin
e
o
f
VGG
Net
u
s
ed
to
class
i
f
y
th
e
I
L
D
ca
te
g
o
r
y
f
r
o
m
1
7
d
if
f
er
e
n
t c
ateg
o
r
i
es.
Usu
al
l
y
,
I
L
D
h
as
i
n
cr
e
m
en
tal
b
r
ea
th
less
n
es
s
,
lu
n
g
cr
ac
k
li
n
g
,
an
d
an
ir
r
eg
u
lar
ch
e
s
t
x
-
r
a
y
.
A
v
ar
iet
y
o
f
o
th
er
illn
e
s
s
e
s
,
s
u
c
h
as
b
ac
ter
ial
p
n
eu
m
o
n
ia,
p
u
l
m
o
n
a
r
y
ed
e
m
a,
an
d
m
ali
g
n
an
c
y
(
e.
g
.
,
ca
r
cin
o
m
ato
s
i
s
l
y
m
p
h
a
n
g
i
tis
)
,
ar
e
id
en
tif
ied
a
s
d
if
f
er
e
n
tial
d
iag
n
o
s
es
[
2
]
.
T
y
p
ica
ll
y
,
tes
ts
f
o
r
p
u
l
m
o
n
ar
y
f
u
n
ct
io
n
s
h
o
w
lo
w
er
v
o
lu
m
es
o
f
l
u
n
g
s
,
p
o
o
r
g
as
tr
an
s
m
i
s
s
io
n
,
an
d
h
y
p
o
x
e
m
ia.
R
ed
u
cin
g
th
e
tr
an
s
f
er
f
a
cto
r
an
d
tr
an
s
p
o
r
t
co
ef
f
icie
n
t
f
o
r
ca
r
b
o
n
m
o
n
o
x
id
e
ar
e
t
y
p
ical
o
f
l
u
n
g
p
a
r
en
ch
y
m
a
d
is
ea
s
es
a
n
d
t
h
eir
b
lo
o
d
s
u
p
p
l
y
[
3
,
4
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
10
,
No
.
6
,
Decem
b
er
2020
:
6
2
8
3
-
6
2
9
1
6284
T
h
er
ef
o
r
e,
in
I
L
D,
b
u
t
al
s
o
em
p
h
y
s
e
m
a
a
n
d
p
u
l
m
o
n
ar
y
v
a
s
cu
lar
d
is
ea
s
e,
th
e
s
e
p
ar
a
m
et
er
s
ar
e
d
ec
r
ea
s
ed
.
I
n
e
m
p
h
y
s
e
m
a,
t
h
e
c
h
e
s
t
x
-
r
a
y
s
u
s
u
al
l
y
s
h
o
w
h
y
p
er
i
n
f
lat
ed
lu
n
g
s
,
w
h
ile,
w
it
h
r
etic
u
lo
n
o
d
u
lar
i
n
f
iltra
tio
n
,
th
e
I
L
D
s
h
o
w
s
g
e
n
er
all
y
d
e
cr
ea
s
ed
lu
n
g
v
o
lu
m
e
s
.
On
ce
I
L
D
h
as
s
u
s
p
ec
ted
th
e
f
ir
s
t
s
tep
is
a
t
h
o
r
o
u
g
h
ex
a
m
in
at
io
n
o
f
t
h
e
h
ea
lt
h
h
is
t
o
r
y
w
it
h
a
p
ar
ticu
lar
f
o
c
u
s
o
n
o
cc
u
p
atio
n
al
h
i
s
to
r
y
(
e.
g
.
,
a
s
b
esto
s
,
co
al
d
u
s
t)
,
en
v
ir
o
n
m
e
n
tal
e
x
p
o
s
u
r
e
(
e.
g
.
,
b
ir
d
t
o
u
ch
,
s
m
o
k
e
f
r
o
m
ci
g
ar
ettes),
a
lis
t
o
f
all
m
ed
icatio
n
s
(
e.
g
.
,
am
io
d
ar
o
n
e,
m
et
h
o
tr
ex
ate)
an
d
a
n
y
s
y
m
p
t
o
m
s
t
h
at
co
u
ld
i
n
cl
u
d
e
in
f
ec
t
io
u
s
d
is
ea
s
es
o
f
th
e
l
u
n
g
s
[
5
]
.
T
h
e
f
u
ll
m
ed
ical
h
is
to
r
y
o
f
a
p
atien
t
an
d
an
y
r
is
k
f
ac
to
r
s
f
o
r
an
i
m
m
u
n
o
co
m
p
r
o
m
is
ed
co
n
d
itio
n
is
e
s
s
e
n
tial
as
t
h
e
cli
n
ical
co
n
tex
t
d
ep
en
d
s
o
n
h
o
w
t
h
e
s
u
b
s
eq
u
e
n
t
e
x
a
m
i
n
atio
n
s
ar
e
i
n
ter
p
r
eted
.
I
n
s
o
m
e
i
n
s
ta
n
ce
s
,
it
is
li
k
el
y
to
h
a
v
e
eo
s
in
o
p
h
ilia,
a
u
to
an
t
ic,
o
r
av
ian
p
r
ec
ip
itin
.
I
t
is
c
u
m
b
er
s
o
m
e
to
d
iag
n
o
s
e
th
e
p
r
ese
n
ce
o
f
I
L
D
in
a
p
atie
n
t
b
y
cli
n
ical
d
at
a
a
n
d
to
g
o
t
h
r
o
u
g
h
all
s
i
m
ilar
t
y
p
e
s
o
f
HR
C
T
i
m
a
g
es
o
f
a
p
atien
t
s
i
n
ce
I
L
D
en
c
o
m
p
a
s
s
es
m
an
y
d
i
f
f
er
en
t
p
at
h
o
lo
g
ical
p
r
o
ce
s
s
e
s
.
T
h
e
ef
f
icie
n
c
y
o
f
t
h
e
d
iag
n
o
s
is
o
f
I
L
D
th
r
o
u
g
h
clin
ica
l
h
is
to
r
y
is
les
s
th
a
n
2
0
%.
C
u
r
r
en
tl
y
,
an
o
p
en
ch
es
t
b
io
p
s
y
is
th
e
b
est
w
a
y
o
f
co
n
f
ir
m
i
n
g
t
h
e
p
r
ese
n
ce
o
f
I
L
D
.
I
n
t
h
e
d
iag
n
o
s
tic
s
o
f
s
o
m
e
I
L
Ds,
f
o
r
e
x
a
m
p
le,
L
u
n
g
b
io
p
s
y
is
a
cr
u
cia
l
co
m
p
o
n
en
t
an
d
is
r
ar
el
y
r
eq
u
i
r
ed
f
o
r
th
e
d
iag
n
o
s
i
s
o
f
in
t
er
s
titi
al
id
io
p
ath
ic
p
n
eu
m
o
n
ia.
A
f
le
x
ib
le
b
r
o
n
ch
o
s
co
p
e
ca
n
b
e
p
er
f
o
r
m
ed
s
i
m
u
lta
n
eo
u
s
l
y
w
i
th
b
r
o
n
c
h
o
alv
e
o
lar
lav
ag
e
(
B
A
L
)
,
an
d
s
m
all
s
ec
tio
n
s
o
f
l
u
n
g
s
a
r
e
co
llected
ad
j
ac
en
t
to
th
e
b
r
o
n
ch
i
u
s
i
n
g
tr
an
s
b
r
o
n
ch
ial
b
io
p
s
ies.
A
s
u
r
g
ical
b
io
p
s
y
r
eq
u
ir
es
g
en
er
al
a
n
e
s
t
h
esia
w
it
h
a
co
m
p
lica
tio
n
r
ate
o
f
ap
p
r
o
x
i
m
atel
y
1
0
%
-
2
0
%
an
d
a
m
o
r
tali
t
y
r
ate
o
f
les
s
t
h
a
n
1
%
f
o
r
t
h
e
g
r
o
u
p
o
f
p
atien
t
s
cu
r
r
e
n
tl
y
u
n
d
er
s
elec
tio
n
.
Ma
n
y
p
atien
ts
ar
e
d
ee
m
ed
u
n
f
it
f
o
r
b
io
p
s
y
,
a
n
d
t
h
e
p
o
s
s
ib
le
b
en
e
f
its
o
f
a
s
s
e
s
s
i
n
g
th
e
h
i
s
to
p
ath
o
lo
g
ical
h
i
s
to
r
y
o
f
t
h
e
d
is
ea
s
e
s
h
o
u
ld
b
e
w
eig
h
ed
ag
ain
s
t t
h
e
p
r
o
ce
d
u
r
al
r
is
k
s
.
A
t
t
h
is
m
o
m
e
n
t,
t
h
e
p
r
o
p
o
s
ed
m
et
h
o
d
u
s
e
s
a
d
ee
p
lear
n
i
n
g
ar
ch
itect
u
r
e
to
ca
te
g
o
r
ize
t
h
e
I
L
D
f
r
o
m
HR
C
T
i
m
a
g
es.
T
h
e
I
L
D
co
m
es
w
i
th
v
ar
io
u
s
ca
te
g
o
r
ies,
an
d
alm
o
s
t
ev
er
y
ca
te
g
o
r
y
lo
o
k
s
lik
e
t
h
e
s
a
m
e
in
HR
C
T
i
m
a
g
es.
I
t
ca
u
s
e
s
to
cli
n
icia
n
s
to
id
en
ti
f
y
th
e
e
x
ac
t
p
ar
a
m
eter
s
f
r
o
m
th
o
s
e
i
m
a
g
es.
So
m
eti
m
es
it
lead
s
to
co
n
f
u
s
io
n
ev
e
n
f
o
r
th
e
d
o
ct
o
r
s
also
to
co
n
clu
d
e.
I
t
w
i
ll
h
elp
in
t
h
e
i
n
clu
s
i
v
e
n
e
s
s
o
f
cli
n
ical
e
v
al
u
atio
n
f
o
r
a
b
etter
u
n
d
er
s
tan
d
i
n
g
o
f
t
h
e
d
is
ea
s
e.
On
ce
it
d
etec
ted
,
th
e
tr
ea
t
m
e
n
t
to
th
e
p
atien
t
s
ca
n
g
et
a
s
tar
t
as
s
o
o
n
as
po
s
s
ib
le.
I
f
th
e
clin
ical
an
d
HR
C
T
f
ea
t
u
r
es
ar
e
t
y
p
ical
o
f
I
L
D,
a
b
io
p
s
y
m
a
y
n
o
t
b
e
r
eq
u
ir
ed
.
So
,
in
th
is
p
r
o
p
o
s
ed
w
o
r
k
,
w
e
ai
m
to
ca
teg
o
r
ize
1
7
ca
teg
o
r
ies
o
f
I
L
D
f
r
o
m
HR
C
T
im
a
g
es
b
y
u
s
in
g
a
d
ee
p
lear
n
in
g
n
et
w
o
r
k
n
a
m
ed
S
m
aller
VG
G
Net.
T
h
e
p
r
im
ar
y
o
b
j
ec
tiv
e
o
f
t
h
i
s
ex
p
er
i
m
e
n
t
i
s
to
f
i
n
d
th
e
p
r
esen
ce
o
f
I
L
D
b
y
a
n
al
y
zi
n
g
v
ar
io
u
s
HR
C
T
i
m
a
g
es.
HR
C
T
g
iv
e
s
g
r
ea
ter
ac
cu
r
ac
y
th
a
n
a
ch
e
s
t
r
ad
io
g
r
ap
h
f
o
r
th
e
d
iag
n
o
s
is
o
f
I
L
D
clas
s
if
icatio
n
.
C
ateg
o
r
ies
o
f
I
L
D
in
cl
u
d
e
n
o
d
u
les,
t
h
ick
e
n
ed
s
ep
ta,
r
etic
u
latio
n
,
r
ed
u
ce
d
atte
n
u
a
t
io
n
ar
ea
s
,
g
r
o
u
n
d
-
g
las
s
o
p
ac
ities
,
h
o
n
e
y
co
m
b
i
n
g
,
a
n
d
l
y
m
p
h
n
o
d
es
an
d
p
le
u
r
a
i
n
v
o
lv
e
m
en
t
i
n
ce
r
tai
n
d
is
ea
s
es.
T
h
e
p
r
ese
n
ce
o
f
I
L
D
in
H
R
C
T
i
m
a
g
e
s
ca
n
b
e
en
s
u
r
ed
b
y
a
n
al
y
zi
n
g
its
p
atter
n
b
e
ca
u
s
e
ea
c
h
ca
te
g
o
r
y
o
f
I
L
D
h
as
d
if
f
er
en
t
p
atter
n
s
in
th
e
H
R
C
T
i
m
ag
e.
I
d
en
ti
f
ica
tio
n
o
f
th
e
t
y
p
e
o
f
I
L
D
i
s
ess
e
n
tial
to
tr
ea
t
th
e
d
is
ea
s
e.
T
h
er
e
ar
e
m
an
y
t
y
p
e
s
o
f
I
L
D.
T
h
e
p
r
o
p
o
s
ed
s
ec
o
n
d
ar
y
o
b
j
ec
tiv
e
is
to
ca
teg
o
r
ize
th
e
I
L
D
f
r
o
m
1
7
d
if
f
er
en
t
t
y
p
es.
T
h
e
ter
m
I
L
D
ap
p
lies
to
a
w
id
e
v
ar
iet
y
o
f
m
o
r
e
th
an
2
0
0
lu
n
g
d
i
s
o
r
d
er
s
.
I
t
is
a
cr
u
cial
tas
k
f
o
r
t
h
e
cli
n
i
cian
s
to
d
eter
m
i
n
e
th
e
p
ar
a
m
eter
s
f
r
o
m
H
R
C
T
im
ag
e
s
b
ec
au
s
e
ev
e
n
t
h
o
u
g
h
I
L
D
h
as
a
v
ar
iet
y
o
f
p
atter
n
s
will
lo
o
k
lik
e
t
h
e
s
a
m
e
f
o
r
h
u
m
a
n
e
y
e
s
.
T
h
e
p
r
o
p
o
s
ed
d
ee
p
lear
n
in
g
tec
h
n
iq
u
e
w
ill
h
elp
to
ca
teg
o
r
ize
th
e
I
L
D
f
r
o
m
HR
C
T
i
m
ag
e
s.
2.
RE
L
AT
E
D
WO
RK
S
T
o
th
e
b
est
o
f
o
u
r
k
n
o
w
led
g
e,
d
ee
p
lear
n
in
g
h
as
n
o
t
b
ee
n
r
e
p
o
r
ted
to
liter
atu
r
e
f
o
r
th
e
cla
s
s
i
f
icatio
n
o
f
I
L
D.
I
n
s
o
m
e
ap
p
r
o
ac
h
es
,
th
er
e
ar
e
u
s
ed
v
ar
io
u
s
d
ee
p
lear
n
in
g
co
n
ce
p
t
s
r
elev
a
n
t
to
m
ed
ical
i
m
a
g
e
an
al
y
s
is
.
T
h
e
s
u
r
v
e
y
d
id
m
ai
n
l
y
b
ase
o
n
i
m
ag
e
cla
s
s
i
f
icat
io
n
,
r
eg
is
tr
atio
n
,
s
eg
m
e
n
tatio
n
,
o
b
j
ec
t d
etec
tio
n
,
an
d
o
th
er
tas
k
s
.
T
h
e
m
o
s
t
s
tu
d
y
w
as
r
e
s
p
ec
ted
in
t
h
e
ar
ea
o
f
t
h
e
b
r
ea
s
t,
n
e
u
r
o
,
r
etin
al,
d
ig
ital
p
ath
o
lo
g
y
,
p
u
l
m
o
n
ar
y
,
ca
r
d
iac,
ab
d
o
m
in
al,
m
u
s
c
u
lo
s
k
eleta
l
[
6
]
.
Fro
m
v
ar
io
u
s
s
t
u
d
ies,
t
h
e
a
u
th
o
r
s
f
o
u
n
d
m
an
y
t
h
i
n
g
s
,
s
u
c
h
as
t
h
e
i
m
p
ac
t
o
f
d
ee
p
lear
n
in
g
al
g
o
r
ith
m
s
i
n
t
h
e
a
n
al
y
s
i
s
o
f
m
ed
ical
i
m
ag
e
s
,
c
h
all
en
g
e
s
i
n
a
n
al
y
zin
g
,
an
d
b
en
ef
it
s
f
r
o
m
t
h
is
p
r
o
ce
s
s
.
T
h
e
b
est
k
in
d
o
f
m
o
d
els
f
o
r
th
e
an
al
y
s
i
s
o
f
i
m
a
g
es
to
d
ate
w
er
e
co
n
v
o
l
u
tio
n
al
n
eu
r
al
s
y
s
te
m
s
(
C
NN)
.
C
NN
's
co
n
tai
n
ed
n
u
m
er
o
u
s
la
y
er
s
th
at
c
h
an
g
e
t
h
eir
co
n
tr
ib
u
tio
n
w
it
h
co
n
v
o
l
u
tio
n
ch
an
n
el
s
o
f
a
litt
le
d
e
g
r
ee
.
I
n
co
m
p
u
ter
aid
,
d
ee
p
co
n
v
o
l
u
tio
n
al
s
y
s
te
m
s
h
ad
tu
r
n
ed
in
to
th
e
m
et
h
o
d
o
f
c
h
o
ice.
T
h
e
an
al
y
s
i
s
o
f
t
h
e
m
ed
ical
i
m
ag
e
co
m
m
u
n
it
y
n
e
t
w
o
r
k
h
ad
p
aid
h
ee
d
to
t
h
ese
cr
u
cial
ad
v
an
ce
m
e
n
t
s
.
T
h
e
Nif
ty
Net
p
l
atf
o
r
m
u
s
ed
to
ad
d
r
ess
th
e
i
d
i
o
s
y
n
cr
asi
es
o
f
m
ed
ic
al
im
ag
in
g
b
y
s
u
p
p
lem
en
tin
g
th
e
cu
r
r
en
t
d
e
ep
l
ea
r
n
in
g
in
f
r
a
s
tr
u
ctu
r
e
.
B
ase
d
o
n
th
e
T
en
s
o
r
Flo
w
lib
r
a
r
y
,
th
e
Nif
ty
Net
b
u
i
lt.
T
h
e
T
en
s
o
r
Fl
o
w
lib
r
a
r
y
p
r
o
v
i
d
e
d
th
e
t
o
o
ls
f
o
r
ex
ec
u
tin
g
th
em
ef
f
icien
tly
o
n
h
ar
d
w
ar
e
r
es
o
u
r
c
es
an
d
d
ef
in
in
g
co
m
p
u
tati
o
n
al
p
i
p
el
in
es
[
7
]
.
V
GGN
et
em
er
g
ed
f
r
o
m
th
e
n
ee
d
t
o
r
e
d
u
c
e
an
d
b
o
o
s
t
t
r
ain
in
g
tim
e
th
e
n
u
m
b
e
r
o
f
p
a
r
am
ete
r
s
in
th
e
C
ONV
l
ay
er
s
.
VGGN
et
(
VGG1
6
,
V
GG
1
9
,
et
c.
)
is
av
a
ila
b
le
in
m
u
ltip
l
e
v
er
s
i
o
n
s
,
w
h
ich
o
n
l
y
d
if
f
er
in
th
e
to
tal
n
u
m
b
er
o
f
n
et
w
o
r
k
c
o
m
p
o
n
en
t
s
.
T
h
e
m
a
x
i
m
u
m
p
a
r
a
m
eter
o
f
VG
G1
6
is
1
3
8
m
ill
io
n
.
I
t
is
i
m
p
o
r
tan
t
to
n
o
te
h
er
e
t
h
at
a
ll
k
er
n
el
s
i
n
C
o
n
v
ar
e
3
x
3
s
ize
an
d
2
x
2
m
ax
p
o
o
l
w
ith
t
w
o
s
ta
g
es.
T
h
e
p
u
r
p
o
s
e
o
f
a
f
ix
ed
k
er
n
el
is
t
o
r
ep
r
o
d
u
ce
all
v
ar
iab
le
-
s
ized
k
er
n
el
s
u
s
ed
i
n
Ale
x
n
et
(
1
1
x
1
1
,
5
x
5
,
3
x
3
)
u
s
in
g
s
ev
er
al
3
x
3
k
er
n
el
s
as
a
b
u
ild
i
n
g
b
lo
ck
.
I
t
i
m
p
r
o
v
es
o
n
A
le
x
Net
b
y
s
u
b
s
tit
u
ti
n
g
lar
g
e
k
er
n
el
f
i
lter
s
(
1
1
an
d
5
r
esp
ec
tiv
el
y
i
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
I
d
en
tifi
ca
tio
n
o
f i
n
ters
titi
a
l lu
n
g
d
is
ea
s
es u
s
in
g
d
ee
p
lea
r
n
in
g
(
N
id
h
in
R
a
ju
)
6285
th
e
f
ir
s
t
an
d
s
ec
o
n
d
la
y
er
o
f
co
n
v
o
lu
t
io
n
)
w
it
h
s
m
all
k
er
n
el
-
s
ized
3
X3
f
ilter
s
o
n
e
a
f
t
er
an
o
th
er
.
I
n
th
i
s
co
n
tex
t,
m
u
ltip
le
s
m
aller
k
er
n
els,
s
tac
k
ed
,
ar
e
b
etter
th
a
n
a
lar
g
er
o
n
e,
g
iv
e
n
t
h
e
f
ac
t
th
at
m
u
ltip
le
n
o
n
l
in
ea
r
la
y
er
s
in
cr
ea
s
e
t
h
e
n
et
w
o
r
k
d
ep
th
t
h
at
e
n
ab
les
it
to
lear
n
m
o
r
e
co
m
p
le
x
f
ea
t
u
r
es
a
t
a
l
o
w
er
co
s
t.
T
h
e
k
er
n
e
l
h
as a
d
if
f
er
en
t r
ec
ep
tiv
e
f
ield
[
6
,
7
]
.
Fo
r
d
iag
n
o
s
i
n
g
a
n
d
s
cr
ee
n
in
g
o
f
m
a
n
y
l
u
n
g
d
i
s
ea
s
e
s
,
th
e
ch
est
X
-
r
a
y
is
co
m
m
o
n
l
y
u
s
i
n
g
as
th
e
to
o
l
f
o
r
th
e
r
ad
io
lo
g
ical
ex
a
m
in
at
io
n
s
.
T
h
e
o
b
j
ec
t
s
eg
m
e
n
tatio
n
an
d
d
etec
tio
n
b
y
d
ee
p
lear
n
in
g
p
r
o
d
u
c
e
b
etter
p
er
f
o
r
m
a
n
ce
i
n
t
h
e
m
ed
ical
i
m
ag
e
an
al
y
s
is
d
o
m
ai
n
[
8
]
.
I
n
m
ed
ical
i
m
a
g
in
g
,
t
h
e
p
r
ec
is
e
an
al
y
s
is
,
a
s
w
ell
as
ev
alu
a
tio
n
o
f
d
i
s
ea
s
e,
r
elies
u
p
o
n
b
o
th
i
m
ag
e
i
n
ter
p
r
etati
o
n
an
d
i
m
a
g
e
ac
q
u
i
s
itio
n
.
I
m
ag
e
ac
q
u
is
itio
n
h
as
i
m
p
r
o
v
ed
co
n
s
id
er
ab
l
y
,
f
i
n
is
h
ed
la
te
y
ea
r
s
,
w
i
th
g
ad
g
ets
g
a
in
i
n
g
i
n
f
o
r
m
atio
n
at
q
u
ic
k
er
r
ates
a
n
d
e
x
p
an
d
ed
g
o
als.
T
h
e
i
m
ag
e
i
n
ter
p
r
etatio
n
p
r
o
ce
s
s
,
b
e
th
at
as
it
m
a
y
,
h
as
as
o
f
late
p
r
o
f
ited
b
y
co
m
p
u
ter
tech
n
o
lo
g
y
[
9
]
.
p
o
ly
m
y
o
s
iti
s
(
P
M)
an
d
d
e
r
m
ato
m
y
o
s
i
tis
(
DM
)
ar
e
f
o
u
n
d
atio
n
al
p
r
o
v
o
ca
tiv
e
d
is
ar
r
an
g
e
s
w
it
h
o
b
s
cu
r
e
etio
lo
g
y
,
f
u
r
th
er
m
o
r
e,
p
at
h
o
g
en
esi
s
.
T
h
e
y
p
r
in
c
ip
all
y
in
f
l
u
en
ce
s
tr
ia
ted
m
u
s
c
les,
b
r
in
g
i
n
g
ab
o
u
t
p
r
o
x
i
m
al
m
u
s
cle
s
h
o
r
tco
m
i
n
g
.
P
o
ly
m
y
o
s
itis
a
n
d
d
er
m
a
to
m
y
o
s
iti
s
ar
e
ex
tr
e
m
e
s
ick
n
e
s
s
ele
m
en
ts
i
n
f
lu
e
n
ci
n
g
s
k
e
letal
m
u
s
cles
an
d
d
i
f
f
er
e
n
t
o
r
g
a
n
s
,
in
cl
u
d
in
g
t
h
e
l
u
n
g
s
.
I
n
ter
s
t
itia
l
lu
n
g
d
is
ea
s
e
(
I
L
D)
i
n
P
M/D
M
is
p
r
o
g
r
ess
iv
el
y
p
er
ce
iv
ed
as
a
g
e
n
u
i
n
e
e
n
ta
n
g
le
m
en
t
o
f
t
h
e
i
n
f
ec
tio
n
.
I
L
D
is
a
t
y
p
ical
ad
d
itio
n
al
ar
ti
cu
lar
ap
p
ea
r
an
ce
o
f
r
h
eu
m
ato
id
ar
th
r
iti
s
(
R
A
)
,
an
d
a
cr
itical
r
ea
s
o
n
f
o
r
b
leak
n
e
s
s
an
d
m
o
r
talit
y
in
t
h
is
p
atie
n
t p
o
p
u
lace
[
1
0
,
1
1
]
.
HR
C
T
is
b
r
o
ad
l
y
ac
ce
s
s
ib
le,
r
eliab
le
in
th
e
h
a
n
d
s
o
f
e
x
p
er
ien
ce
d
r
ad
io
lo
g
is
ts
,
ea
s
e,
an
d
o
k
a
y
co
n
tr
asted
w
it
h
a
ca
r
ef
u
l
l
u
n
g
b
io
p
s
y
.
E
v
al
u
at
io
n
o
f
t
h
e
d
eg
r
ee
o
f
r
ad
io
lo
g
ical
f
ib
r
o
s
i
s
lo
an
s
e
x
tr
a
p
r
o
g
n
o
s
t
ic
estee
m
.
T
h
e
an
n
o
u
n
ce
d
p
r
e
d
o
m
i
n
a
n
ce
o
f
I
L
D
i
n
P
M/DM
in
p
r
io
r
in
v
esti
g
atio
n
s
g
en
er
all
y
f
lu
ct
u
ate
s
attr
ib
u
tab
le
to
th
e
ab
s
en
ce
o
f
u
n
if
o
r
m
s
y
m
p
to
m
atic
cr
iter
ia
f
o
r
I
L
D,
t
h
e
d
i
f
f
er
en
t
p
h
ases
o
f
t
h
e
s
ic
k
n
ess
i
n
w
h
ic
h
p
atien
ts
w
er
e
e
x
a
m
in
e
d
,
an
d
th
e
w
ell
s
p
r
in
g
o
f
p
atie
n
t
r
ef
er
r
al
[
1
2
,
1
3
]
.
Fo
r
a
s
tu
d
y
,
th
e
au
t
h
o
r
s
co
u
l
d
u
s
e
a
n
al
y
s
is
i
n
5
0
p
atien
ts
w
i
th
b
io
p
s
y
-
d
e
m
o
n
s
tr
ated
NSI
P
,
an
d
a
C
T
ch
ec
k
w
a
s
s
u
r
v
e
y
e
d
b
y
t
w
o
th
o
r
ac
ic
r
ad
io
lo
g
is
ts
in
ac
co
r
d
.
A
f
t
e
r
th
e
o
b
s
er
v
atio
n
s
w
er
e
p
o
r
tr
ay
ed
,
th
e
e
y
e
w
it
n
es
s
es
d
ec
id
ed
w
h
et
h
er
th
e
o
b
s
er
v
atio
n
s
w
er
e
g
o
o
d
w
ith
r
ec
en
tl
y
d
i
s
tr
ib
u
ted
p
o
r
tr
a
y
als
o
f
n
o
n
s
p
ec
i
f
ic
in
ter
s
titi
al
p
n
eu
m
o
n
ia
(
NSI
P
)
o
r
w
h
e
th
er
t
h
e
d
is
co
v
er
ies
w
o
u
ld
b
o
ls
ter
th
e
co
n
clu
s
io
n
o
f
an
o
t
h
er
u
n
e
n
d
in
g
i
n
f
iltra
tiv
e
l
u
n
g
s
ic
k
n
e
s
s
.
T
h
e
C
T
o
b
s
er
v
atio
n
s
in
p
atie
n
ts
w
ith
NSI
P
an
d
to
co
n
tr
ast
th
ese
a
n
d
t
h
e
C
T
d
is
co
v
er
ies
o
f
o
th
er
p
er
p
etu
a
l
in
f
iltra
ti
v
e
lu
n
g
s
ic
k
n
ess
e
s
w
e
r
e
d
escr
ib
ed
[
1
4
]
.
Dee
p
n
eu
r
al
s
y
s
te
m
s
h
av
e,
as
o
f
late,
i
n
cr
ea
s
ed
s
ig
n
i
f
ica
n
t
b
u
s
i
n
ess
en
th
u
s
ia
s
m
b
ec
au
s
e
o
f
th
e
i
m
p
r
o
v
e
m
en
t
o
f
n
e
w
v
ar
iatio
n
s
o
f
C
NNs
an
d
t
h
e
co
m
i
n
g
o
f
p
r
o
f
icie
n
t
p
ar
allel
s
o
lv
er
s
u
p
g
r
ad
ed
f
o
r
p
r
esen
t
-
d
a
y
GP
Us
[
1
5
]
.
No
tw
ith
s
tan
d
i
n
g
,
co
n
tr
asted
w
ith
2
D
i
m
a
g
es
f
o
r
th
e
m
o
s
t
p
ar
t
u
t
ilized
in
co
m
p
u
ter
v
is
io
n
,
s
y
m
p
to
m
atic
a
n
d
in
t
er
v
e
n
tio
n
al
i
m
ag
e
s
d
ata
in
th
e
m
ed
icin
al
f
ield
ar
e
f
r
eq
u
en
t
l
y
v
o
lu
m
etr
ic.
T
h
is
m
a
k
es
a
n
ee
d
f
o
r
ca
lc
u
latio
n
s
p
er
f
o
r
m
i
n
g
d
i
v
is
io
n
s
in
3
D
b
y
ta
k
i
n
g
th
e
e
n
tire
v
o
lu
m
e
co
n
ten
t
i
n
to
th
e
r
ec
o
r
d
w
ith
o
u
t
a
m
o
m
e
n
t
'
s
d
ela
y
[
1
6
]
.
P
r
ep
ar
in
g
a
d
e
ep
C
NN
f
r
o
m
s
cr
atch
is
tr
o
u
b
leso
m
e
b
ec
au
s
e
it
r
eq
u
ir
es
a
lo
t
o
f
lab
eled
tr
ain
i
n
g
d
ata
an
d
a
lo
t
o
f
ap
titu
d
es
to
g
u
ar
an
tee
ap
p
r
o
p
r
iate
in
ter
m
in
g
li
n
g
.
A
p
r
o
m
is
i
n
g
o
p
tio
n
is
to
ad
j
u
s
t
a
C
NN
th
at
h
a
s
b
ee
n
p
r
e
-
tr
ain
ed
u
ti
lizi
n
g
,
f
o
r
ex
a
m
p
le,
an
ex
p
an
s
i
v
e
ar
r
an
g
e
m
en
t
o
f
n
a
m
ed
s
ta
n
d
ar
d
im
a
g
e
s
.
No
n
et
h
eles
s
,
t
h
e
s
i
g
n
i
f
ica
n
t
co
n
tr
ast
s
a
m
o
n
g
r
eg
u
lar
an
d
m
ed
ici
n
al
i
m
a
g
es
m
a
y
e
x
h
o
r
t a
g
ain
s
t s
u
c
h
lear
n
i
n
g
ex
c
h
an
g
e
[
1
7
]
.
T
h
e
p
er
f
ec
t
b
io
p
s
y
p
r
o
ce
d
u
r
e
to
d
is
ti
n
g
u
is
h
m
e
n
w
i
th
p
r
o
s
tate
ca
n
ce
r
w
o
u
ld
b
e
f
i
n
d
in
g
s
o
f
j
u
s
t
s
ig
n
i
f
ica
n
t
p
r
o
s
tate
ca
n
ce
r
g
r
o
w
th
a
n
d
m
i
n
i
m
izatio
n
o
f
in
co
n
s
eq
u
e
n
tial
p
r
o
s
tate
ca
n
ce
r
d
etec
tio
n
an
d
s
u
b
s
eq
u
en
t
o
v
er
tr
ea
t
m
e
n
t.
M
R
I
-
T
B
x
is
a
p
r
o
m
is
i
n
g
s
y
s
te
m
t
h
at
m
a
y
o
f
f
er
a
p
o
r
tio
n
o
f
th
e
s
e
f
o
ca
l
p
o
in
ts
co
n
tr
asted
w
it
h
s
ta
n
d
ar
d
ef
f
ici
en
t
T
R
US
-
B
x
,
as
ap
p
ea
r
ed
d
ir
ec
t
ex
a
m
in
a
tio
n
o
f
th
e
t
w
o
b
io
p
s
y
ap
p
r
o
ac
h
es
i
n
th
e
an
al
y
s
i
s
[
1
8
,
1
9
]
.
T
h
e
an
al
y
s
i
s
o
f
p
r
o
s
tate
ca
n
ce
r
d
if
f
e
r
s
f
r
o
m
th
at
i
n
o
th
er
s
tr
o
n
g
o
r
g
an
ca
n
ce
r
w
h
er
e
i
m
a
g
in
g
is
u
tili
ze
d
to
d
is
tin
g
u
i
s
h
th
o
s
e
p
atien
t
s
w
h
o
r
eq
u
ir
e
a
b
io
p
s
y
.
T
h
e
p
r
o
s
tate
m
al
ig
n
a
n
t
g
r
o
w
th
d
em
o
n
s
tr
ati
v
e
p
at
h
w
a
y
o
f
f
er
s
tr
an
s
r
ec
tal
u
l
tr
aso
u
n
d
-
g
u
id
ed
b
io
p
s
y
(
T
R
US
-
b
io
p
s
y
)
i
n
m
e
n
w
h
o
p
r
esen
t
w
it
h
a
r
aised
s
er
u
m
p
r
o
s
tate
-
s
p
ec
if
ic
an
ti
g
e
n
(
P
S
A
)
[
2
0
]
.
I
m
a
g
i
n
g
b
io
m
ar
k
er
s
(
I
B
s
)
ar
e
es
s
en
t
ial
to
t
h
e
s
tan
d
ar
d
ad
m
in
i
s
tr
atio
n
o
f
p
atie
n
ts
w
it
h
t
h
e
d
is
ea
s
e.
I
B
S
u
ti
lized
d
ay
b
y
d
a
y
i
n
o
n
co
lo
g
y
to
i
n
co
r
p
o
r
ate
clin
ical
T
NM
ar
r
an
g
e,
tar
g
et
r
ea
ctio
n
,
an
d
l
ef
t
v
e
n
tr
ic
u
lar
d
is
ch
ar
g
e
d
i
v
is
io
n
.
Oth
er
C
T
,
MRI,
PE
T
,
an
d
u
ltra
s
o
n
o
g
r
ap
h
y
b
io
m
ar
k
er
s
ar
e
u
tili
ze
d
w
id
el
y
in
m
ali
g
n
an
t
g
r
o
w
t
h
r
esear
c
h
an
d
m
ed
icatio
n
i
m
p
r
o
v
e
m
en
t
[
2
1
,
2
2
]
.
Ne
w
I
B
s
s
h
o
u
ld
b
e
b
u
ilt
u
p
eit
h
er
as
h
e
lp
f
u
l
ap
p
ar
atu
s
es
f
o
r
test
in
g
r
esear
ch
s
p
ec
u
latio
n
s
i
n
cli
n
ica
l
p
r
eli
m
i
n
ar
ies
a
n
d
r
esear
ch
e
x
a
m
in
e
s
o
r
as
cli
n
ic
al
n
ec
e
s
s
ar
y
lead
er
s
h
ip
in
s
tr
u
m
en
ts
f
o
r
u
s
e
in
h
u
m
an
s
er
v
i
ce
s
,
b
y
i
n
ter
s
ec
tio
n
'
tr
a
n
s
latio
n
al
g
ap
s
'
th
r
o
u
g
h
ap
p
r
o
v
al
an
d
ca
p
ab
ilit
y
[
2
3
,
2
4
]
.
E
n
d
ea
v
o
r
s
to
s
et
u
p
a
q
u
a
n
ti
tativ
e
w
a
y
to
d
ea
l
w
it
h
t
h
e
C
T
-
b
ased
p
o
r
tr
ay
al
o
f
th
e
lu
n
g
p
ar
en
c
h
y
m
a
i
n
i
n
ter
s
titi
al
l
u
n
g
m
a
lad
y
(
co
u
n
ti
n
g
e
m
p
h
y
s
e
m
a)
h
a
s
b
ee
n
lo
o
k
ed
at
co
n
t
in
u
o
u
s
l
y
.
T
h
e
ex
ac
tn
e
s
s
o
f
t
h
ese
ap
p
ar
atu
s
e
s
m
u
s
t
b
e
s
ite
f
r
ee
.
Mu
l
ti
-
in
d
icato
r
p
u
s
h
C
T
h
as
r
e
m
ai
n
ed
th
e
h
i
g
h
est
q
u
a
lit
y
lev
e
l
f
o
r
i
m
ag
in
g
th
e
l
u
n
g
,
an
d
it
g
iv
e
s
th
e
ca
p
ac
it
y
t
o
p
ictu
r
e
b
o
th
lu
n
g
s
tr
u
ct
u
r
e
j
u
s
t a
s
l
u
n
g
w
o
r
k
[
2
5
]
.
3.
I
M
P
L
E
M
E
NT
AT
I
O
N
3
.
1
.
Da
t
a
c
o
llect
io
n
T
h
e
d
ata
s
et
co
m
p
r
is
e
s
3
0
4
5
i
m
a
g
es
o
f
H
R
C
T
w
it
h
t
h
r
ee
-
d
i
m
en
s
io
n
al
an
n
o
tated
p
ath
o
l
o
g
ical
l
u
n
g
tis
s
u
e
r
eg
io
n
s
to
g
et
h
er
w
it
h
m
ed
ical
cr
iter
ia
o
f
I
L
D
co
n
d
itio
n
s
th
at
h
a
v
e
b
ee
n
co
n
f
ir
m
ed
p
ath
o
lo
g
ica
ll
y
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
10
,
No
.
6
,
Decem
b
er
2020
:
6
2
8
3
-
6
2
9
1
6286
I
t
co
n
tain
s
1
2
8
p
atien
ts
w
it
h
o
n
e
o
f
th
ir
tee
n
I
L
D
h
i
s
to
l
o
g
ical
co
n
d
itio
n
s
,
1
0
8
im
ag
e
s
er
ie
s
o
f
o
v
er
4
1
liter
s
o
f
an
n
o
tated
p
u
l
m
o
n
ar
y
ti
s
s
u
e
p
atter
n
s
,
an
d
a
co
m
p
r
e
h
en
s
iv
e
co
llectio
n
o
f
9
9
m
ed
ical
p
ar
am
eter
s
.
I
t
p
r
o
v
id
es
DI
C
OM
i
m
ag
e
s
an
d
a
.
tx
t
f
ile
w
it
h
a
n
n
o
tatio
n
s
.
T
h
en
t
h
er
e
is
u
s
ed
Fil
e
v
ie
w
er
P
lu
s
s
o
f
t
w
ar
e
to
co
n
v
er
t
.
r
o
i
to
.
j
p
g
f
o
r
m
at
a
s
s
h
o
w
n
i
n
T
ab
le
1
.
L
ater
,
w
e
b
u
ilt
an
al
g
o
r
it
h
m
in
Ma
t
lab
to
ch
an
g
e
t
h
e
n
a
m
e
o
f
i
m
a
g
e
f
iles
(
b
ec
au
s
e
t
h
e
s
a
m
e
n
a
m
e
s
w
er
e
ex
i
s
ti
n
g
f
o
r
m
a
n
y
f
iles
)
.
E
a
ch
i
m
a
g
e
co
n
tain
s
9
6
d
p
i
an
d
5
1
2
p
ix
els
f
o
r
b
o
th
w
id
t
h
a
n
d
h
e
ig
h
t.
T
h
e
d
ata
o
r
g
an
ized
i
n
d
i
f
f
er
en
t
c
lass
e
s
ac
co
r
d
in
g
to
t
h
e
v
ar
iet
y
o
f
I
L
D
s
u
c
h
a
s
f
ib
r
o
s
is
,
h
ea
lt
h
y
,
e
m
p
h
y
s
e
m
a,
g
r
o
u
n
d
g
las
s
,
co
n
s
o
lid
atio
n
,
m
icr
o
n
o
d
u
les,
b
r
o
n
ch
ial
w
al
l
th
ick
en
in
g
,
r
etic
u
latio
n
,
c
y
s
t
s
,
m
ac
r
o
n
o
d
u
le
s
,
p
er
ip
h
er
al
m
icr
o
n
o
d
u
les,
b
r
o
n
ch
iec
tasi
s
,
air
tr
a
p
p
in
g
,
i
n
cr
ea
s
ed
atten
u
a
tio
n
,
ea
r
l
y
f
ib
r
o
s
is
,
tu
b
er
c
u
lo
s
is
,
a
n
d
P
C
P
.
I
t
c
o
n
tain
ed
th
e
co
m
m
o
n
I
L
D
ca
te
g
o
r
ies
w
er
e
h
ea
lt
h
y
,
e
m
p
h
y
s
e
m
a,
g
r
o
u
n
d
-
g
la
s
s
,
f
ib
r
o
s
is
,
a
n
d
m
icr
o
n
o
d
u
les as s
h
o
w
n
i
n
T
ab
le
2
.
T
ab
le
1
.
Data
s
et
lab
els an
d
th
e
to
tal
n
u
m
b
er
o
f
i
m
a
g
e
s
f
o
r
ea
ch
lab
el
w
ith
a
n
ex
a
m
p
le
C
a
t
e
g
o
r
y
I
mag
e
s
Ex
a
mp
l
e
C
a
t
e
g
o
r
y
I
mag
e
s
Ex
a
mp
l
e
B
r
o
n
c
h
i
e
c
t
a
si
s
88
h
e
a
l
t
h
y
2
1
9
c
o
n
so
l
i
d
a
t
i
o
n
1
6
5
i
n
c
r
e
a
se
d
a
t
t
e
n
u
a
t
i
o
n
20
c
y
st
s
24
mac
r
o
n
o
d
u
l
e
s
1
1
5
e
a
r
l
y
_
f
i
b
r
o
si
s
26
mi
c
r
o
n
o
d
u
l
e
s
5
2
9
T
ab
le
2
.
C
o
m
m
o
n
I
L
D
ca
teg
o
r
y
a
n
d
its
d
escr
ip
tio
n
C
o
mm
o
n
I
L
D
C
a
t
e
g
o
r
y
D
e
scri
p
t
i
o
n
h
e
a
l
t
h
y
N
o
r
mal
c
o
n
d
i
t
i
o
n
e
mp
h
y
se
ma
Emp
h
y
se
ma
i
s
c
o
mm
o
n
l
y
u
se
d
f
o
r
r
h
e
u
ma
t
o
i
d
a
r
t
h
r
i
t
i
s
-
a
sso
c
i
a
t
e
d
d
i
so
r
d
e
r
s
r
e
l
a
t
e
d
t
o
t
o
b
a
c
c
o
smo
k
i
n
g
i
n
p
a
t
i
e
n
t
s w
i
t
h
i
d
i
o
p
a
t
h
i
c
p
u
l
mo
n
a
r
y
f
i
b
r
o
si
s (I
P
F
)
a
n
d
I
L
D
.
g
r
o
u
n
d
-
g
l
a
ss
I
t
i
s
d
e
scri
p
t
i
v
e
o
f
a
n
a
r
e
a
i
n
w
h
i
c
h
C
T
w
i
t
h
b
r
o
n
c
h
i
a
l
a
n
d
v
a
sc
u
l
a
r
mar
k
i
n
g
s
a
r
e
mo
r
e
a
t
t
e
n
u
a
t
e
d
i
n
t
h
e
l
u
n
g
.
f
i
b
r
o
si
s
W
h
e
n
d
a
m
a
g
e
d
o
r
sca
r
r
e
d
l
u
n
g
t
i
ss
u
e
mi
c
r
o
n
o
d
u
l
e
s
I
n
c
l
u
d
e
s mi
l
i
a
r
y
t
u
b
e
r
c
u
l
o
si
s
a
n
d
e
n
d
o
b
r
o
n
c
h
i
a
l
d
i
se
a
se
d
i
sse
mi
n
a
t
i
o
n
3
.
2
.
Me
t
ho
do
lo
g
y
S
m
aller
V
GGNe
t
is
th
e
C
N
N
ar
ch
itect
u
r
e
u
s
ed
f
o
r
th
e
p
r
o
p
o
s
ed
w
o
r
k
,
a
s
tr
ea
m
li
n
ed
v
er
s
io
n
o
f
it
s
en
o
r
m
o
u
s
s
ib
li
n
g
,
VG
GNe
t.
T
h
e
s
tag
es
o
f
t
h
e
p
r
o
p
o
s
ed
w
o
r
k
p
ictu
r
ed
in
th
e
g
iv
e
n
b
lo
ck
d
iag
r
a
m
a
s
s
h
o
wn
in
Fi
g
u
r
e
1
.
T
h
e
en
tire
i
m
p
le
m
en
t
atio
n
p
r
o
ce
s
s
h
as t
h
r
ee
s
ta
g
es.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
I
d
en
tifi
ca
tio
n
o
f i
n
ters
titi
a
l lu
n
g
d
is
ea
s
es u
s
in
g
d
ee
p
lea
r
n
in
g
(
N
id
h
in
R
a
ju
)
6287
3
.
2
.
1
.
B
uil
d
t
he
da
t
a
s
et
T
h
e
d
ataset
cr
ea
tio
n
alg
o
r
it
h
m
b
u
i
lt
b
y
i
n
cl
u
d
in
g
t
h
e
B
in
g
I
m
a
g
e
Sear
c
h
A
P
I
o
f
Mic
r
o
s
o
f
t,
w
h
ich
w
a
s
p
ar
t
o
f
Mic
r
o
s
o
f
t
's
co
g
n
i
tiv
e
s
er
v
ice
s
,
to
b
r
in
g
A
I
to
th
e
te
x
t,
s
p
ee
ch
,
v
i
s
io
n
,
a
n
d
m
o
r
e
f
o
r
ap
p
s
a
n
d
s
o
f
t
w
ar
e
to
b
u
ild
th
e
th
o
r
o
u
g
h
l
y
lear
n
ed
i
m
a
g
e
d
atasets
.
A
f
ter
th
is
p
er
io
d
,
if
it
n
ee
d
s
to
u
s
e
f
o
r
f
u
r
t
h
er
i
m
p
le
m
en
ta
tio
n
,
t
h
en
it r
eq
u
ir
ed
to
p
ay
s
o
m
e
a
m
o
u
n
t to
g
o
o
g
le
f
o
r
th
e
u
s
a
g
e.
3
.
2
.
2
.
T
ra
i
n t
he
m
o
del
T
h
e
d
ata
au
g
m
en
tatio
n
is
d
o
n
e
b
y
u
s
in
g
a
class
ca
lled
I
m
ag
e
Data
Ge
n
er
ato
r
,
a
m
et
h
o
d
u
s
ed
to
ca
p
tu
r
e
cu
r
r
en
t
i
m
a
g
es
in
o
u
r
d
ataset
a
n
d
to
p
r
o
d
u
ce
ex
tr
a
tr
ain
in
g
d
ata
u
s
i
n
g
r
a
n
d
o
m
tr
an
s
f
o
r
m
atio
n
s
(
r
o
tatio
n
s
,
s
h
ea
r
,
etc.
)
.
T
h
e
au
g
m
e
n
tatio
n
o
f
d
ata
u
s
ed
t
o
av
o
id
o
v
er
f
itti
n
g
.
I
n
cr
e
m
e
n
tal
i
m
p
r
o
v
e
m
en
t
s
th
r
o
u
g
h
b
ac
k
p
r
o
p
ag
atio
n
w
o
u
ld
co
n
tr
ib
u
te
to
t
h
e
tr
ain
i
n
g
o
f
o
u
r
Net
w
o
r
k
f
o
r
7
5
E
P
OC
HS.
W
e
s
et
an
i
n
it
ial
lear
n
in
g
r
ate
as
3
(
A
d
a
m
Op
t
i
m
izer
d
e
f
au
l
t
v
alu
e)
.
W
e
u
s
e
d
a
b
atch
s
ize
o
f
3
2
f
o
r
t
h
is
p
r
o
j
ec
t.
T
h
e
i
m
ag
e
s
w
il
l s
u
b
s
eq
u
e
n
tl
y
b
e
r
esized
a
s
9
6
x
9
6
w
i
th
th
r
ee
c
h
an
n
el
s
.
F
ir
s
t,
th
e
a
lg
o
r
it
h
m
lo
ad
s
ea
ch
i
m
ag
e
in
to
m
e
m
o
r
y
f
r
o
m
d
is
k
.
T
h
en
,
it p
er
f
o
r
m
s
p
r
ep
r
o
ce
s
s
in
g
b
y
r
es
izin
g
i
m
a
g
es
w
ith
w
id
t
h
a
n
d
h
ei
g
h
t a
s
9
6
an
d
th
e
n
co
n
v
er
t it
in
to
an
ar
r
a
y
f
o
r
m
.
A
m
u
lt
i
-
l
ab
el
clas
s
i
f
icatio
n
ta
s
k
p
er
f
o
r
m
ed
to
s
p
lit
th
e
p
at
h
i
n
to
s
ev
er
al
lab
el
s
.
A
2
-
ele
m
e
n
t
li
s
t
i
s
g
en
er
ated
af
ter
t
h
is
ex
ec
u
tio
n
a
n
d
th
e
n
ad
d
ed
to
th
e
lab
els.
T
h
en
,
co
n
v
er
t
th
e
lis
t
to
a
n
ar
r
ay
o
f
Nu
m
P
y
a
n
d
s
ca
le
t
h
e
p
ix
el
in
ten
s
itie
s
to
th
e
r
an
g
e
o
f
[
0
,
1
]
.
R
eg
ar
d
in
g
t
h
e
b
in
ar
iza
tio
n
o
f
m
u
lti
-
clas
s
class
i
f
icatio
n
lab
els,
t
h
e
Mu
ltiL
ab
elB
i
n
ar
izer
class
o
f
th
e
s
cien
ce
-
lear
n
l
ib
r
ar
y
u
s
ed
.
T
h
en
th
ese
lab
el
s
m
atc
h
ed
an
d
tr
an
s
f
o
r
m
ed
in
t
o
a
v
ec
to
r
,
w
h
ic
h
e
n
co
d
es
t
h
e
clas
s
(
s
)
o
f
th
e
i
m
a
g
e.
I
n
m
ac
h
in
e
lear
n
i
n
g
p
r
ac
tices
,
th
e
n
e
x
t
s
tep
i
s
to
d
iv
id
e
t
h
e
d
ata
f
o
r
tr
ai
n
in
g
a
n
d
tes
tin
g
.
8
0
%
o
f
th
e
i
m
ag
e
s
ar
e
allo
ca
ted
f
o
r
tr
ain
i
n
g
,
an
d
2
0
%
f
o
r
test
i
n
g
.
T
h
is
th
en
i
n
itial
ized
th
e
d
ata
au
g
m
e
n
tatio
n
o
b
j
ec
t b
ec
au
s
e
th
e
d
ataset
co
n
tain
ed
less
t
h
a
n
1
,
0
0
0
i
m
ag
e
s
f
o
r
s
o
m
e
ca
te
g
o
r
ies.
An
d
n
e
x
t,
t
h
e
S
m
aller
V
GGNe
t
m
o
d
el
co
m
e
s
in
to
th
e
p
ict
u
r
e
f
o
r
th
e
M
u
lt
i
-
lab
el
clas
s
i
f
icatio
n
.
T
h
er
e
u
s
ed
b
i
n
ar
y
cr
o
s
s
-
e
n
tr
o
p
y
r
at
h
er
t
h
an
ca
te
g
o
r
ical
cr
o
s
s
-
en
tr
o
p
y
to
co
m
p
ile
t
h
e
m
o
d
el.
A
f
ter
t
h
e
tr
ain
i
n
g
co
m
p
leted
,
w
e
s
av
ed
th
e
m
o
d
el
an
d
lab
el
b
in
ar
izer
to
d
is
k
b
y
tr
ai
n
i
n
g
w
it
h
o
u
r
d
ata
au
g
m
en
tatio
n
g
en
er
ato
r
s
.
T
h
e
ac
cu
r
ac
y
an
d
lo
s
s
o
f
th
e
p
lo
t
w
er
e
s
to
r
ed
as
a
n
i
m
a
g
e
f
ile
f
r
o
m
t
h
er
e.
Fig
u
r
e
1
.
B
lo
ck
d
iag
r
a
m
o
f
t
h
e
p
r
o
p
o
s
ed
p
r
o
j
ec
t
3
.
2
.
3
.
T
est
t
he
m
o
del
On
ce
C
NN
h
as
b
ee
n
tr
ain
ed
,
it
co
n
s
tr
u
cted
a
s
c
r
ip
t
t
h
at
co
u
ld
id
en
ti
f
y
o
b
j
ec
ts
th
at
wer
e
n
o
t
i
n
tr
ain
i
n
g
o
r
v
alid
atio
n
/
test
co
ll
ec
tio
n
.
T
h
e
p
r
e
-
p
r
o
ce
s
s
o
f
a
n
i
n
p
u
t i
m
a
g
e
i
s
d
o
n
e
i
n
t
h
e
s
a
m
e
w
a
y
a
s
m
en
tio
n
ed
in
t
h
e
tr
ain
i
n
g
p
h
a
s
e.
T
h
e
m
o
d
el
an
d
m
u
lti
-
lab
el
b
in
ar
izer
a
r
e
lo
ad
ed
f
r
o
m
d
is
k
to
m
e
m
o
r
y
b
y
an
al
g
o
r
ith
m
.
I
t
class
i
f
ied
th
e
i
m
ag
e
a
n
d
ex
tr
ac
ted
th
e
to
p
m
o
s
t
lab
els
b
ased
o
n
in
d
ices.
T
h
e
ar
r
ay
in
d
e
x
is
t
h
en
s
o
r
ted
in
d
escen
d
in
g
o
r
d
er
b
y
its
r
elate
d
p
r
o
b
ab
ilit
y
,
r
ec
o
r
d
in
g
t
h
e
t
o
p
m
o
s
t
lab
el
i
n
d
ices,
w
h
ich
t
h
er
ef
o
r
e
is
th
e
b
est
o
n
e
in
t
h
e
n
et
w
o
r
k
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
10
,
No
.
6
,
Decem
b
er
2020
:
6
2
8
3
-
6
2
9
1
6288
Algorithm
-
1
Input: HRCT images in .jpg format.
Output: Training loss and accuracy
1.
Set the paths of dataset, model, MLB object and image plotting path
2.
Load the images from dataset
3.
Pre
-
process each images
a. resize the image into 96*96
b. Change the raw pixel intensities of each image into the range [0, 1]
4.
Updating data list by extracting class labels (ILD category) and append it
5.
Binarize labels with the advanced multi
-
label of scikit learn
6.
Di
vi
de
da
ta
in
to
se
ts
of
tr
ai
ni
ng
an
d
te
st
in
g
us
in
g
80
%
of
tr
ai
ni
ng
da
ta
an
d
20
%
of
test data.
7.
Binary cross
-
entropy compile the model and then store the model and MLB to the disk
8.
Store an image in disk with training loss and accuracy
Algorithm
-
2
Input: HRCT image in .jpg format.
Output: Predict ILD
category with index
1.
Set the paths of dataset, model, MLB object and image plotting path
2.
Load the images from dataset
3.
Pre
-
process each images
a. resize into 96*96
b. Change the raw pixel intensities of each image into the range [0, 1]
4.
Load the CNN and MLB
5.
Predict the category of input image with index of the category
6.
Sh
ow
th
e
ou
tp
ut
by
di
sp
la
yi
ng
in
pu
t
im
ag
e
wi
th
pr
e
di
ct
ed
ca
te
go
ry
an
d
co
rr
e
sp
on
di
ng
index
3
.
3
.
E
x
peri
m
ent
a
l r
esu
lt
I
n
th
i
s
ex
p
er
i
m
e
n
t,
t
h
e
to
tal
n
u
m
b
er
o
f
3
0
4
5
HR
C
T
im
ag
es
w
er
e
p
r
o
ce
s
s
ed
.
E
a
ch
im
ag
e
w
a
s
co
n
v
er
ted
f
r
o
m
.
r
o
i
to
.
j
p
g
f
o
r
m
at
f
o
r
th
e
s
m
o
o
th
e
x
ec
u
ti
o
n
o
f
th
e
s
y
s
te
m
.
T
h
en
,
th
e
s
e
i
m
a
g
es
m
an
u
all
y
ar
r
an
g
ed
in
1
7
d
if
f
er
en
t
ca
te
g
o
r
ies
b
ased
o
n
I
L
D
ca
te
g
o
r
i
es
f
o
r
tr
ai
n
i
n
g
th
e
n
et
w
o
r
k
.
On
ce
it
h
a
s
w
h
o
l
l
y
tr
ain
ed
,
g
o
t
ac
cu
r
ac
y
,
an
d
v
alid
ate
d
th
e
ac
cu
r
a
c
y
,
9
5
%,
an
d
9
4
%,
r
esp
ec
tiv
el
y
as
s
h
o
w
n
i
n
Fig
u
r
e
2
.
So
m
e
s
a
m
p
le
i
m
ag
e
s
w
er
e
ap
p
lied
to
th
e
tr
ain
ed
n
et
w
o
r
k
,
an
d
it
co
u
ld
p
r
ed
ic
t
th
e
ca
te
g
o
r
ies
s
u
cc
e
s
s
i
v
el
y
,
a
n
d
its
ca
te
g
o
r
y
p
r
ed
ictio
n
r
esu
lts
ar
e
g
iv
e
n
as s
h
o
w
n
in
Fig
u
r
es 3
an
d
4
.
Fig
u
r
e
2
.
Gr
ap
h
ical
r
ep
r
esen
ta
tio
n
o
f
tr
ain
in
g
lo
s
s
a
n
d
ac
cu
r
ac
y
Fig
u
r
e
3
.
P
r
ed
icte
d
as b
r
o
n
ch
iecta
s
is
f
o
r
a
n
e
w
in
p
u
t
Fig
u
r
e
4
.
P
r
ed
icte
d
as f
r
o
u
n
d
f
o
r
a
n
e
w
i
n
p
u
t
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
I
d
en
tifi
ca
tio
n
o
f i
n
ters
titi
a
l lu
n
g
d
is
ea
s
es u
s
in
g
d
ee
p
lea
r
n
in
g
(
N
id
h
in
R
a
ju
)
6289
T
h
e
av
ailab
ilit
y
o
f
t
h
e
d
ata
f
o
r
th
is
e
x
p
er
i
m
e
n
t
w
as
m
i
n
i
m
al.
T
h
e
to
tal
n
u
m
b
er
o
f
3
0
4
5
HR
C
T
i
m
a
g
es
o
n
l
y
co
u
ld
co
llect
f
o
r
th
is
e
x
p
er
i
m
e
n
t.
As
it
w
as
a
d
ee
p
lear
n
in
g
m
o
d
el,
it
r
eq
u
ir
e
s
a
m
a
s
s
i
v
e
a
m
o
u
n
t
o
f
d
ata
to
tr
ain
t
h
e
m
o
d
el.
T
h
e
s
u
f
f
icie
n
t
n
u
m
b
er
o
f
i
m
ag
e
s
p
er
ea
ch
class
ca
n
p
r
o
d
u
ce
a
g
o
o
d
r
esu
lt
w
h
en
it
test
ed
.
T
h
e
an
a
l
y
s
is
o
f
th
is
wo
r
k
w
as
d
o
n
e
b
y
cr
ea
ti
n
g
1
2
d
if
f
er
en
t
m
o
d
el
s
b
y
co
n
s
id
er
in
g
t
h
e
n
u
m
b
er
o
f
class
es
h
a
s
tak
e
n
a
n
d
test
ed
i
t
w
ith
d
i
f
f
er
en
t
i
m
a
g
es.
I
n
t
h
e
f
ir
s
t
m
o
d
el,
th
er
e
h
a
v
e
tak
e
n
1
7
cla
s
s
e
s
an
d
it
s
g
r
ap
h
icall
y
p
icto
r
is
ed
in
Fi
g
u
r
e
5
.
Fig
u
r
e
5
.
Gr
ap
h
ical
r
ep
r
esen
ta
tio
n
f
o
r
class
if
icatio
n
o
f
1
7
cla
s
s
es
T
h
e
n
ex
t
m
o
d
el
b
u
il
t
w
it
h
f
i
v
e
clas
s
es,
w
h
ic
h
ar
e
t
h
e
to
p
co
m
m
o
n
ca
te
g
o
r
ies
a
m
o
n
g
1
7
class
es.
I
n
th
e
r
e
m
ai
n
i
n
g
m
o
d
els,
o
n
l
y
t
w
o
cl
ass
e
s
w
er
e
co
n
s
id
er
ed
,
an
d
f
o
r
ea
ch
m
o
d
el,
p
ick
ed
u
p
w
it
h
t
w
o
v
ar
io
u
s
class
es
f
r
o
m
5
p
o
p
u
lar
class
e
s
as
s
h
o
w
n
in
Fi
g
u
r
e
6
.
W
h
e
n
s
a
m
p
le
i
m
a
g
es
test
ed
w
i
th
ea
c
h
m
o
d
el,
t
h
er
e
w
a
s
s
h
o
w
i
n
g
ac
c
u
r
ac
y
v
ar
iatio
n
s
i
n
ea
ch
m
o
d
el.
T
h
e
aim
o
f
cr
ea
tin
g
m
o
d
els
w
i
th
t
w
o
c
la
s
s
es
w
a
s
t
h
at
o
n
ce
i
f
co
u
ld
ab
le
to
f
in
d
th
e
to
p
2
c
lass
es
f
r
o
m
t
h
e
1
7
class
es
m
o
d
el,
it
ca
n
ap
p
ly
in
2
class
e
s
m
o
d
el,
w
h
ich
w
a
s
cr
ea
ted
b
y
th
a
t to
p
p
r
ed
icted
c
lass
es.
T
h
is
ac
ti
v
it
y
ca
n
b
e
u
s
e
d
to
clar
if
y
th
e
d
is
ea
s
e
o
n
ce
m
o
r
e.
Fig
u
r
e
6
.
Gr
ap
h
ical
r
ep
r
esen
ta
tio
n
f
o
r
class
if
icatio
n
o
f
5
clas
s
es
4.
L
I
M
I
T
AT
I
O
NS
T
h
e
li
m
ited
n
u
m
b
er
o
f
tr
ain
i
n
g
d
ata
w
as
o
n
e
o
f
th
e
cr
itic
al
d
r
a
w
b
ac
k
s
o
f
th
i
s
al
g
o
r
ith
m
.
I
d
ea
ll
y
,
w
h
e
n
tr
ain
i
n
g
a
co
n
v
o
l
u
tio
n
al
n
eu
r
al
n
et
w
o
r
k
,
it
r
eq
u
ir
ed
at
least
5
0
0
-
1
0
0
0
im
ag
e
s
p
er
ea
ch
class
.
Ot
h
er
w
i
s
e,
it
af
f
ec
ted
th
e
ac
c
u
r
ac
y
o
f
th
e
r
esu
lt.
O
n
e
o
th
er
d
is
ad
v
a
n
t
ag
e
is
t
h
at
t
h
e
s
y
s
te
m
ta
k
es
2
-
3
h
o
u
r
s
to
tr
ai
n
th
e
n
et
w
o
r
k
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
10
,
No
.
6
,
Decem
b
er
2020
:
6
2
8
3
-
6
2
9
1
6290
5.
CO
NCLU
SI
O
N
I
n
th
i
s
p
r
o
p
o
s
ed
w
o
r
k
,
a
d
ee
p
lear
n
in
g
C
NN
ar
c
h
itect
u
r
e
n
a
m
ed
S
m
aller
V
GGNe
t
u
s
ed
to
class
if
y
th
e
I
L
D
ca
te
g
o
r
y
f
r
o
m
1
7
d
if
f
er
en
t
ca
teg
o
r
ies
b
y
p
r
o
ce
s
s
in
g
HR
C
T
i
m
ag
e
s
.
Fo
r
th
is
ex
p
e
r
i
m
en
t,
1
2
v
ar
io
u
s
d
ee
p
lear
n
in
g
m
o
d
els
co
n
s
tr
u
c
ted
ac
co
r
d
in
g
to
th
e
n
u
m
b
er
o
f
clas
s
es
u
s
ed
f
o
r
ea
ch
m
o
d
el.
T
h
e
n
et
w
o
r
k
w
it
h
1
7
class
e
s
co
u
ld
ab
le
to
tr
ai
n
t
h
e
s
y
s
te
m
w
it
h
a
9
5
%
ac
c
u
r
ac
y
r
ate.
Af
ter
t
h
at,
a
f
e
w
n
u
m
b
er
s
o
f
i
n
p
u
t
i
m
ag
e
s
w
er
e
s
u
b
m
itted
to
th
e
tr
ai
n
ed
th
e
s
y
s
te
m
,
a
n
d
it
co
u
ld
class
if
y
t
h
e
I
L
D
ca
teg
o
r
ies
s
u
cc
es
s
i
v
el
y
.
T
h
e
r
em
ai
n
in
g
m
o
d
els
also
cr
ea
ted
b
ased
o
n
th
e
d
is
ea
s
es
w
h
ic
h
h
a
v
e
b
ee
n
o
cc
u
r
r
in
g
m
o
s
t
co
m
m
o
n
l
y
.
On
ce
t
h
e
to
p
m
o
s
t
d
is
ea
s
e
ca
te
g
o
r
y
ca
n
b
e
ab
le
to
id
en
ti
f
y
,
a
g
ain
,
i
t
ca
n
c
h
ec
k
w
it
h
th
e
s
u
b
-
m
o
d
els
f
o
r
b
etter
clar
it
y
.
Af
ter
ap
p
l
y
i
n
g
s
o
m
e
s
a
m
p
le
s
o
n
t
h
e
p
r
o
ce
d
u
r
e
m
en
tio
n
ed
ab
o
v
e,
it
s
h
o
w
ed
s
o
m
e
v
ar
iatio
n
i
n
p
r
ed
ic
tin
g
t
h
e
r
es
u
lt.
B
ased
o
n
t
h
e
a
v
ailab
ilit
y
o
f
d
ata
f
o
r
ea
ch
clas
s
,
t
h
e
m
o
d
els
g
en
er
at
ed
d
if
f
er
e
n
t
r
es
u
lt
s
.
T
h
e
m
o
d
els
i
n
w
h
ic
h
clas
s
es
co
n
tain
ed
a
r
ig
h
t
a
m
o
u
n
t
o
f
d
ata
g
av
e
a
g
o
o
d
r
esu
lt
in
p
r
ed
ictin
g
th
e
I
L
D
ca
teg
o
r
y
t
h
a
n
m
o
d
els
in
w
h
ic
h
cla
s
s
e
s
i
n
cl
u
d
ed
less
a
m
o
u
n
t
o
f
d
ata.
Hen
ce
t
h
is
s
t
u
d
y
ca
n
b
e
p
r
o
ce
s
s
ed
f
o
r
ea
r
l
y
-
s
ta
g
e
d
etec
tio
n
o
f
I
L
D
f
o
r
b
etter
tr
ea
tm
en
t
to
t
h
e
p
atien
ts
.
I
n
t
h
e
f
u
t
u
r
e,
th
i
s
s
y
s
te
m
ca
n
b
e
u
t
ilized
to
ex
ec
u
te
t
h
e
s
y
s
te
m
w
it
h
o
u
t
r
esizin
g
t
h
e
i
m
a
g
es
an
d
to
ap
p
ly
m
o
r
e
f
ilter
s
to
i
n
cr
ea
s
e
th
e
ac
c
u
r
ac
y
o
f
th
e
r
esu
lt.
RE
F
E
R
E
NC
E
S
[1
]
Işın
,
A
li
,
Ce
m
Dire
k
o
ğ
lu
,
a
n
d
M
e
li
k
e
Ş
a
h
,
"
R
e
v
ie
w
o
f
M
RI
-
b
a
se
d
b
ra
in
tu
m
o
r
i
m
a
g
e
se
g
m
e
n
ta
t
io
n
u
si
n
g
d
e
e
p
lea
rn
in
g
m
e
th
o
d
s,
"
Pro
c
e
d
ia
Co
m
p
u
ter
S
c
ien
c
e
,
v
o
l.
1
0
2
,
p
p
.
3
1
7
-
3
2
4
,
2
0
1
6
.
[2
]
L
iu
,
S
iq
i
,
e
t
a
l.
"
Early
d
iag
n
o
sis
o
f
A
lzh
e
i
m
e
r
'
s
d
ise
a
se
w
it
h
d
e
e
p
l
e
a
rn
in
g
,
"
IEE
E
1
1
th
in
ter
n
a
ti
o
n
a
l
sy
mp
o
siu
m o
n
b
io
me
d
ic
a
l
ima
g
i
n
g
(
IS
BI)
,
p
p
.
1
0
1
5
-
1
0
1
8
,
2
0
1
4
.
[3
]
L
isk
o
w
s
k
i,
P
a
w
e
ł,
a
n
d
K.
K
ra
w
i
e
c
,
"
S
e
g
m
e
n
ti
n
g
re
ti
n
a
l
b
lo
o
d
v
e
ss
e
ls
w
it
h
d
e
e
p
n
e
u
r
a
l
n
e
tw
o
rk
s,
"
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
M
e
d
ica
l
I
ma
g
in
g
,
v
o
l.
3
5
,
n
o
.
1
1
,
p
p
.
2
3
6
9
-
2
3
8
0
,
2
0
1
6
.
[4
]
S
c
h
o
ll
,
In
g
rid
,
A
a
c
h
,
T
.
,
De
se
rn
o
,
T
.
M
.
,
a
n
d
K
u
h
le
n
,
T
.
,
"
Ch
a
ll
e
n
g
e
s
o
f
m
e
d
ica
l
i
m
a
g
e
p
ro
c
e
ss
in
g
,
"
Co
mp
u
te
r
s
c
ien
c
e
-
Res
e
a
rc
h
a
n
d
d
e
v
e
lo
p
me
n
t,
v
o
l
.
2
6
,
n
o
.
1
-
2
,
p
p
.
5
-
1
3
,
2
0
1
1
.
[5
]
W
e
e
se
,
Jü
rg
e
n
,
a
n
d
Cristi
a
n
L
o
re
n
z
,
"
F
o
u
r
c
h
a
ll
e
n
g
e
s
in
m
e
d
ica
l
i
m
a
g
e
a
n
a
l
y
sis
f
ro
m
a
n
in
d
u
stri
a
l
p
e
rsp
e
c
ti
v
e
,
"
El
se
v
ier
,
n
o
.
4
4
-
4
9
,
2
0
1
6
.
[6
]
L
it
jen
s,
G
e
e
rt,
e
t
a
l.
,
"
A
su
rv
e
y
o
n
d
e
e
p
lea
rn
in
g
i
n
m
e
d
ica
l
ima
g
e
a
n
a
l
y
sis,
"
M
e
d
ica
l
ima
g
e
a
n
a
lys
is
,
v
o
l.
4
2
,
p
p
.
6
0
-
8
8
,
2
0
1
7
.
[7
]
G
ib
so
n
,
El
i,
e
t
a
l.
,
"
Nif
t
y
Ne
t:
a
d
e
e
p
-
lea
rn
in
g
p
latf
o
rm
f
o
r
m
e
d
ica
l
im
a
g
i
n
g
,
"
Co
mp
u
ter
me
th
o
d
s
a
n
d
p
ro
g
r
a
ms
i
n
b
io
me
d
ici
n
e
,
v
o
l.
1
5
8
,
p
p
.
1
1
3
-
1
2
2
,
2
0
1
6
.
[8
]
W
a
n
g
,
X
iao
so
n
g
,
e
t
a
l.
"
Ch
e
stx
-
ra
y
8
:
Ho
sp
it
a
l
-
sc
a
le
c
h
e
st
x
-
ra
y
d
a
tab
a
se
a
n
d
b
e
n
c
h
m
a
rk
s
o
n
we
a
k
l
y
-
su
p
e
rv
ise
d
c
las
si
f
ica
ti
o
n
a
n
d
lo
c
a
li
z
a
ti
o
n
o
f
c
o
m
m
o
n
th
o
ra
x
d
ise
a
se
s,
"
Pro
c
e
e
d
in
g
s
o
f
t
h
e
IEE
E
Co
n
fer
e
n
c
e
o
n
Co
mp
u
ter
Vi
sio
n
a
n
d
P
a
tt
e
rn
Rec
o
g
n
it
i
o
n
,
p
p
.
2
0
9
7
-
2
1
0
6
,
2
0
1
7
.
[9
]
G
re
e
n
sp
a
n
,
Ha
y
it
,
Bra
m
V
a
n
G
in
n
e
k
e
n
,
a
n
d
R
o
n
a
ld
M
.
S
u
m
m
e
rs,
"
G
u
e
st
e
d
it
o
rial
d
e
e
p
lea
rn
in
g
in
m
e
d
ica
l
im
a
g
in
g
:
Ov
e
r
v
ie
w
a
n
d
f
u
tu
re
p
ro
m
i
se
o
f
a
n
e
x
c
it
in
g
n
e
w
tec
h
n
iq
u
e
,
"
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
M
e
d
ica
l
Ima
g
i
n
g
,
v
o
l.
3
5
,
n
o
.
5
,
p
p
.
1
1
5
3
-
1
1
5
9
,
2
0
1
6
.
[1
0
]
S
p
a
m
p
in
a
to
,
Co
n
c
e
tt
o
,
e
t
a
l.
,
"
De
e
p
lea
rn
in
g
f
o
r
a
u
to
m
a
ted
sk
e
let
a
l
b
o
n
e
a
g
e
a
ss
e
ss
m
e
n
t
in
X
-
ra
y
i
m
a
g
e
s,
"
M
e
d
ica
l
ima
g
e
a
n
a
lys
is,
v
o
l.
3
6
,
p
p
.
4
1
-
5
1
,
2
0
1
7
.
[1
1
]
Kim
,
Eu
n
ice
J.,
e
t
a
l.
"
Us
u
a
l
in
ters
ti
ti
a
l
p
n
e
u
m
o
n
ia
in
r
h
e
u
m
a
to
id
a
rth
rit
is
-
a
ss
o
c
iate
d
i
n
ters
ti
ti
a
l
lu
n
g
d
ise
a
se
,
"
Eu
ro
p
e
a
n
Res
p
ira
to
ry
J
o
u
rn
a
l
,
v
o
l.
3
5
,
n
o
.
6
,
p
p
.
1
3
2
2
-
1
3
2
8
,
2
0
1
6
.
[1
2
]
F
a
th
i,
M
a
ry
a
m
,
e
t
a
l.
,
"
In
ters
ti
ti
a
l
lu
n
g
d
ise
a
se
,
a
c
o
m
m
o
n
m
a
n
if
e
sta
ti
o
n
o
f
n
e
w
l
y
d
iag
n
o
se
d
p
o
ly
m
y
o
siti
s
a
n
d
d
e
rm
a
to
m
y
o
siti
s,
"
An
n
a
ls o
f
th
e
r
h
e
u
ma
ti
c
d
ise
a
se
s
,
v
o
l.
6
3
,
n
o
.
3
,
p
p
.
2
9
7
-
3
0
1
,
2
0
0
4
.
[1
3
]
L
e
e
,
Yo
o
n
Jin
,
e
t
a
l.
,
"
He
p
a
to
c
e
ll
u
lar
c
a
rc
in
o
m
a
:
d
iag
n
o
stic
p
e
rf
o
rm
a
n
c
e
o
f
m
u
lt
id
e
tec
to
r
CT
a
n
d
M
R
im
a
g
in
g
-
a
s
y
st
e
m
a
ti
c
re
v
ie
w
a
n
d
m
e
ta
-
a
n
a
ly
sis,
"
Ra
d
io
l
o
g
y
,
v
o
l.
2
7
5
,
n
o
.
1
,
p
p
.
9
7
-
1
0
9
,
2
0
1
6
.
[1
4
]
Ha
rt
m
a
n
,
T
h
o
m
a
s
E.
,
e
t
a
l.
,
"
N
o
n
sp
e
c
if
ic
in
ters
ti
ti
a
l
p
n
e
u
m
o
n
ia
:
v
a
riab
le
a
p
p
e
a
ra
nc
e
a
t
h
ig
h
-
re
so
lu
ti
o
n
c
h
e
st
CT
,
"
Ra
d
io
l
o
g
y
,
v
o
l
.
2
1
7
,
n
o
.
3
,
p
p
.
7
0
1
-
7
0
5
,
2
0
0
0
.
[1
5
]
Ka
n
g
,
Eu
n
h
e
e
,
Ju
n
h
o
n
g
M
i
n
,
a
n
d
Jo
n
g
Ch
u
l
Ye
,
"
A
d
e
e
p
c
o
n
v
o
lu
t
io
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
u
sin
g
d
irec
ti
o
n
a
l
w
a
v
e
lets
f
o
r
l
o
w
‐d
o
se
X
‐ra
y
CT
re
c
o
n
stru
c
ti
o
n
,
"
M
e
d
ica
l
p
h
y
sic
s
,
v
o
l.
4
4
,
n
o
.
1
0
,
2
0
1
7
.
[1
6
]
M
il
leta
ri,
F
a
u
sto
,
Na
ss
i
r
Na
v
a
b
,
a
n
d
S
e
y
e
d
-
A
h
m
a
d
A
h
m
a
d
i,
"V
-
n
e
t:
F
u
ll
y
c
o
n
v
o
lu
ti
o
n
a
l
n
e
u
r
a
l
n
e
tw
o
rk
s
f
o
r
v
o
lu
m
e
tri
c
m
e
d
ica
l
i
m
a
g
e
se
g
m
e
n
tatio
n
,
"
Fo
u
rth
I
n
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
3
D V
isi
o
n
(
3
DV),
p
p
.
5
6
5
-
5
7
1
,
2
0
1
6
.
[1
7
]
T
a
jb
a
k
h
sh
,
Nim
a
,
e
t
a
l.
"
Co
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
s
f
o
r
m
e
d
ica
l
ima
g
e
a
n
a
l
y
sis
:
F
u
ll
train
in
g
o
r
f
in
e
tu
n
i
n
g
?
,
"
IEE
E
tra
n
sa
c
ti
o
n
s
o
n
m
e
d
ica
l
ima
g
i
n
g
,
v
o
l.
3
5
,
n
o
.
5
,
p
p
.
1
2
9
9
-
1
3
1
2
,
2
0
1
9
.
[1
8
]
Av
e
n
d
i,
M
.
R.
,
A
ra
sh
Kh
e
ra
d
v
a
r,
a
n
d
Ha
m
id
Ja
f
a
rk
h
a
n
i,
"
A
c
o
m
b
in
e
d
d
e
e
p
-
lea
rn
in
g
a
n
d
d
e
f
o
r
m
a
b
le
-
m
o
d
e
l
a
p
p
ro
a
c
h
t
o
f
u
ll
y
a
u
to
m
a
ti
c
se
g
m
e
n
tatio
n
o
f
th
e
lef
t
v
e
n
tri
c
le
in
c
a
rd
iac
M
RI,
"
M
e
d
ica
l
ima
g
e
a
n
a
lys
is
,
v
o
l
.
3
0
,
p
p
.
1
0
8
-
1
1
9
,
2
0
1
6
.
[1
9
]
S
c
h
o
o
ts,
Iv
o
G
.
,
e
t
a
l.
,
"
M
a
g
n
e
ti
c
re
so
n
a
n
c
e
im
a
g
in
g
-
targ
e
ted
b
i
o
p
sy
m
a
y
e
n
h
a
n
c
e
th
e
d
iag
n
o
sti
c
a
c
c
u
ra
c
y
o
f
sig
n
if
ica
n
t
p
ro
sta
te
c
a
n
c
e
r
d
e
tec
ti
o
n
c
o
m
p
a
re
d
to
sta
n
d
a
r
d
tran
s
re
c
tal
u
lt
ra
so
u
n
d
-
g
u
i
d
e
d
b
i
o
p
sy
:
A
s
y
st
e
m
a
ti
c
re
v
ie
w
a
n
d
m
e
ta
-
a
n
a
l
y
sis,
"
Eu
ro
p
e
a
n
u
r
o
lo
g
y
,
v
o
l
.
6
8
,
n
o
.
3
,
p
p
.
4
3
8
-
4
5
0
,
2
0
1
5
.
[2
0
]
A
h
m
e
d
,
Ha
sh
i
m
U.,
e
t
a
l.
,
"
Dia
g
n
o
s
ti
c
a
c
c
u
ra
c
y
o
f
m
u
lt
i
-
p
a
ra
m
e
tri
c
M
RI
a
n
d
T
RUS
b
io
p
sy
in
p
ro
sta
te
c
a
n
c
e
r
(P
ROMI
S
):
A
p
a
ire
d
v
a
li
d
a
ti
n
g
c
o
n
f
irm
a
to
r
y
stu
d
y
,
"
T
h
e
L
a
n
c
e
t
,
v
o
l.
3
8
9
,
p
p
.
8
1
5
-
8
2
2
,
2
0
1
7
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
I
d
en
tifi
ca
tio
n
o
f i
n
ters
titi
a
l lu
n
g
d
is
ea
s
es u
s
in
g
d
ee
p
lea
r
n
in
g
(
N
id
h
in
R
a
ju
)
6291
[2
1
]
S
m
it
h
-
Bin
d
m
a
n
,
Re
b
e
c
c
a
,
e
t
a
l.
,
"
Us
e
o
f
d
iag
n
o
stic
im
a
g
in
g
stu
d
i
e
s
a
n
d
a
ss
o
c
iate
d
ra
d
iatio
n
e
x
p
o
s
u
re
f
o
r
p
a
ti
e
n
ts
e
n
ro
ll
e
d
in
larg
e
in
teg
ra
ted
h
e
a
lt
h
c
a
re
s
y
ste
m
s,
1
9
9
6
-
2
0
1
0
,
"
J
a
ma
,
v
o
l.
3
0
7
,
n
o
.
2
2
,
p
p
.
2
4
0
0
-
2
4
0
9
,
2
0
1
2
.
[2
2
]
Yin
,
X
iao
-
X
ia,
S
i
ll
a
s
Ha
d
ji
lo
u
c
a
s,
a
n
d
Ya
n
c
h
u
n
Zh
a
n
g
,
"
In
tro
d
u
c
ti
o
n
a
n
d
M
o
ti
v
a
ti
o
n
f
o
r
Co
n
d
u
c
ti
n
g
M
e
d
ica
l
Im
a
g
e
A
n
a
l
y
sis,
"
Pa
tt
e
rn
Cla
ss
if
i
c
a
ti
o
n
o
f
M
e
d
ica
l
Im
a
g
e
s: Co
m
p
u
ter
Ai
d
e
d
Dia
g
n
o
sis,
p
p
.
1
-
2
6
,
2
0
1
7
.
[2
3
]
P
h
a
m
,
Dz
u
n
g
L
.
,
C
h
e
n
y
a
n
g
X
u
,
a
n
d
Je
rry
L
.
P
rin
c
e
,
"
Cu
rre
n
t
m
e
th
o
d
s
in
m
e
d
ica
l
i
m
a
g
e
se
g
m
e
n
tatio
n
,
"
An
n
u
a
l
r
e
v
iew o
f
b
i
o
me
d
ica
l
e
n
g
i
n
e
e
rin
g
,
v
o
l.
2
,
n
o
.
1
,
p
p
.
3
1
5
-
3
3
7
,
2
0
0
0
.
[2
4
]
O'
c
o
n
n
o
r,
Ja
m
e
s
P
B,
e
t
a
l.
,
"
I
m
a
g
in
g
b
io
m
a
r
k
e
r
ro
a
d
m
a
p
f
o
r
c
a
n
c
e
r
stu
d
ies
,
"
N
a
tu
re
re
v
iews
Cli
n
ica
l
o
n
c
o
lo
g
y
,
v
o
l.
1
4
,
n
o
.
3
,
p
p
.
1
6
9
-
1
8
6
,
2
0
1
7
.
[2
5
]
Ho
ffm
a
n
,
Eri
c
A
.
,
e
t
a
l.
,
"
Ch
a
ra
c
teriz
a
ti
o
n
o
f
th
e
in
ters
ti
ti
a
l
lu
n
g
d
ise
a
se
s
v
i
a
d
e
n
sit
y
-
b
a
se
d
a
n
d
tex
tu
re
-
b
a
s
e
d
a
n
a
ly
sis
o
f
c
o
m
p
u
ted
to
m
o
g
ra
p
h
y
i
m
a
g
e
s
o
f
lu
n
g
stru
c
tu
re
a
n
d
fu
n
c
ti
o
n
1
,
"
Aca
d
e
mic
ra
d
i
o
l
o
g
y
,
v
o
l.
1
0
,
n
o
.
1
0
,
p
p
.
1
1
0
4
-
1
1
1
8
,
2
0
0
3
.
B
I
O
G
RAP
H
I
E
S
O
F
AUTH
O
RS
Nid
h
i
n
Ra
j
u
h
a
s
b
e
e
n
w
o
rk
in
g
a
s
a
n
A
s
so
c
iate
S
o
f
t
w
a
r
e
E
n
g
in
e
e
r
in
Ce
n
tre
f
o
r
Di
g
it
a
l
In
n
o
v
a
ti
o
n
,
CHRIST
(De
e
m
e
d
to
b
e
Un
iv
e
rsit
y
),
In
d
ia.
He
p
e
ru
sin
g
M
.
P
h
il
i
n
Co
m
p
u
ter
S
c
ien
c
e
f
ro
m
CHRIST
(D
e
e
m
e
d
to
b
e
U
n
iv
e
rsity
),
In
d
ia.
His
re
se
a
rc
h
in
tere
st
is
I
m
a
g
e
p
ro
c
e
ss
in
g
a
n
d
De
e
p
L
e
a
rn
in
g
.
He
h
a
s p
u
b
l
ish
e
d
a
re
se
a
rc
h
p
a
p
e
r
in
a
n
I
n
tern
a
ti
o
n
a
l
jo
u
r
n
a
l.
Em
a
il
:
n
id
h
i
n
.
ra
ju
@re
s.c
h
rist
u
n
iv
e
rsit
y
.
in
Dr
.
An
ita
H
.
B
.
is
w
o
rk
in
g
a
s
A
s
so
c
iate
P
r
o
f
e
ss
o
r
in
t
h
e
De
p
a
rtm
e
n
t
o
f
Co
m
p
u
ter
S
c
ien
c
e
a
t
CHRIST
(De
e
m
e
d
to
b
e
Un
iv
e
rsity
).
S
h
e
h
a
s
re
c
e
iv
e
d
a
P
h
D
f
r
o
m
G
u
lb
a
rg
a
Un
iv
e
rsit
y
,
In
d
ia,
in
2
0
1
4
.
M
CA
d
e
g
re
e
f
ro
m
In
d
ir
a
G
a
n
d
h
i
Na
ti
o
n
a
l
O
p
e
n
U
n
iv
e
rsity
in
2
0
0
3
.
He
r
re
se
a
rc
h
in
te
re
st
is
Im
a
g
e
p
ro
c
e
ss
in
g
,
S
o
u
n
d
p
ro
c
e
ss
in
g
a
n
d
M
a
c
h
i
n
e
L
e
a
rn
in
g
.
S
h
e
h
a
s
p
u
b
li
s
h
e
d
m
a
n
y
p
a
p
e
rs
in
i
n
tern
a
ti
o
n
a
l
a
n
d
n
a
ti
o
n
a
l
c
o
n
f
e
re
n
c
e
a
n
d
p
u
b
li
sh
e
d
tw
o
Bo
o
k
c
h
a
p
ters
in
S
p
ri
n
g
e
r
L
in
k
,
In
d
ia.
A
lso
,
sh
e
h
a
s p
u
b
li
sh
e
d
m
a
n
y
re
s
e
a
rc
h
p
a
p
e
rs i
n
I
n
tern
a
ti
o
n
a
l
j
o
u
r
n
a
ls.
Em
a
il
:
a
n
it
a
.
h
b
@c
h
rist
u
n
iv
e
rsity
.
in
Pete
r
A
u
g
u
sti
n
e
h
a
s
b
e
e
n
w
o
rk
in
g
a
s
a
n
A
ss
o
c
iat
e
P
ro
f
e
ss
o
r
in
th
e
De
p
a
rtm
e
n
t
o
f
Co
m
p
u
ter
S
c
ien
c
e
,
CHRIST
(De
e
m
e
d
to
b
e
Un
iv
e
rsit
y
),
Ba
n
g
a
lo
re
.
P
e
ter
A
u
g
u
stin
e
h
a
s
a
P
h
D
i
n
M
e
d
ica
l
Im
a
g
e
P
ro
c
e
ss
in
g
in
Clo
u
d
En
v
ir
o
n
m
e
n
t,
w
it
h
o
v
e
r
8
y
e
a
r
s
in
c
lo
u
d
c
o
m
p
u
ti
n
g
a
n
d
5
y
e
a
rs
in
Big
Da
ta
A
n
a
l
y
ti
c
s.
He
h
a
s
a
u
th
o
re
d
v
a
rio
u
s
re
se
a
rc
h
p
a
p
e
rs
p
u
b
li
sh
e
d
in
p
e
e
r
-
re
v
ie
w
e
d
jo
u
rn
a
ls
.
He
h
a
s
b
e
e
n
in
v
o
lv
e
d
i
n
a
M
a
jo
r
Re
se
a
rc
h
P
ro
jec
t
u
si
n
g
Clo
u
d
C
o
m
p
u
ti
n
g
w
h
ich
c
o
sts
m
o
re
th
a
n
1
8
lak
h
s.
He
h
a
s
a
lso
c
o
ll
a
b
o
ra
t
e
d
w
it
h
S
t.
J
o
h
n
’s
M
e
d
ica
l
Re
s
e
a
rc
h
In
stit
u
te
f
o
r
t
h
e
re
se
a
rc
h
p
ro
jec
t
to
d
iag
n
o
se
L
u
n
g
Dise
a
se
s
u
sin
g
c
u
tt
in
g
-
e
d
g
e
A
I
a
n
d
M
a
c
h
in
e
L
e
a
rn
in
g
.
His
re
se
a
r
c
h
in
tere
sts
in
c
lu
d
e
a
rti
f
icia
l
in
telli
g
e
n
c
e
,
Io
T
a
n
d
b
ig
d
a
ta
a
n
a
ly
sis
in
th
e
a
re
a
o
f
He
a
lt
h
c
a
re
a
n
d
d
a
ta
m
in
in
g
,
a
n
d
h
u
m
a
n
c
o
m
p
u
ter
i
n
t
e
r
a
c
ti
o
n
.
He
h
a
s
w
rit
ten
c
h
a
p
ters
f
o
r
th
e
b
o
o
k
s
f
o
c
u
sin
g
o
n
so
m
e
o
f
th
e
e
m
e
rg
in
g
tec
h
n
o
lo
g
ies
su
c
h
a
s
Io
T
,
Da
ta
A
n
a
l
y
t
ic
s
a
n
d
S
c
ien
c
e
,
Bl
o
c
k
c
h
a
in
a
n
d
Dig
it
a
l
Tw
in
.
Em
a
il
:
p
e
ter.au
g
u
stin
e
@c
h
ristu
n
i
v
e
rsit
y
.
in
Evaluation Warning : The document was created with Spire.PDF for Python.