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.
1
5
,
No
.
3
,
J
u
n
e
2
0
2
6
,
p
p
.
2595
~
2
6
0
5
I
SS
N:
2
2
5
2
-
8
9
3
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijai.v
15
.i
3
.
p
p
2
5
9
5
-
2
6
0
5
2595
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
Utilizatio
n of
dep
th
-
wise
and sp
a
ti
a
lly
sepa
ra
ble conv
o
lutiona
l
network
fusio
n f
o
r clas
sifica
tion o
f
white
blo
o
d cells
F
ira
s
M
un
ea
m
B
a
cha
y
1
,
Ali A
bb
a
s
Alza
heiree
2
,
H
a
s
a
na
i
n Ali
H
us
s
ein
3
,
Ahm
ed
No
o
ruldeen A
l
s
a
f
i
4,
5
,
M
o
ha
m
m
ed
H
a
s
a
n Abdu
la
m
ee
r
3
1
D
e
p
a
r
t
me
n
t
o
f
S
c
i
e
n
t
i
f
i
c
A
f
f
a
i
r
,
U
n
i
v
e
r
si
t
y
o
f
K
u
f
a
,
N
a
j
a
f
,
I
r
a
q
2
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
En
g
i
n
e
e
r
i
n
g
T
e
c
h
n
o
l
o
g
y
,
I
mam
A
l
k
a
d
h
i
m
U
n
i
v
e
r
si
t
y
C
o
l
l
e
g
e
,
B
a
g
h
d
a
d
,
I
r
a
q
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
,
F
a
c
u
l
t
y
o
f
E
d
u
c
a
t
i
o
n
f
o
r
G
i
r
l
s,
U
n
i
v
e
r
si
t
y
o
f
K
u
f
a
,
N
a
j
a
f
,
I
r
a
q
4
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
a
n
d
I
n
f
o
r
m
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
,
F
a
c
u
l
t
y
o
f
En
g
i
n
e
e
r
i
n
g
,
U
n
i
v
e
r
s
i
t
y
o
f
Q
o
m,
Q
o
m,
I
r
a
n
5
A
l
ma
k
e
e
n
I
n
st
i
t
u
t
i
o
n
f
o
r
S
c
i
e
n
c
e
a
n
d
K
n
o
w
l
e
d
g
e
,
N
a
j
a
f
,
I
r
a
q
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Dec
1
9
,
2
0
2
4
R
ev
is
ed
Feb
1
6
,
2
0
2
6
Acc
ep
ted
Ap
r
2
0
,
2
0
2
6
Wh
it
e
b
lo
o
d
c
e
ll
s
(W
BCs
)
a
re
a
n
e
ss
e
n
ti
a
l
p
a
rt
o
f
th
e
h
u
m
a
n
imm
u
n
e
sy
ste
m
,
p
lay
in
g
a
sig
n
if
ica
n
t
ro
l
e
in
fig
h
t
in
g
d
ise
a
se
s
a
n
d
in
fe
c
ti
o
n
s.
T
h
e
ir
d
e
tec
ti
o
n
a
n
d
c
las
sifica
ti
o
n
fro
m
m
icro
sc
o
p
ic
b
lo
o
d
ima
g
e
s
is
a
c
ru
c
ial
ste
p
in
d
ia
g
n
o
sin
g
v
a
rio
u
s
d
ise
a
se
s.
Lo
o
k
in
g
a
t
c
e
ll
s
b
y
h
a
n
d
is
stil
l
k
e
y
,
b
u
t
i
t
tak
e
s
a
lo
t
o
f
w
o
rk
a
n
d
m
istak
e
s
c
a
n
h
a
p
p
e
n
.
S
o
,
th
is
stu
d
y
tri
e
s
t
o
imp
r
o
v
e
h
o
w
to
fin
d
a
n
d
s
o
rt
WBCs
u
si
n
g
so
m
e
c
o
o
l
c
o
m
p
u
ter
t
rick
s.
Th
e
stu
d
y
tac
k
li
n
g
issu
e
s
li
k
e
c
e
ll
s
b
e
in
g
o
n
to
p
o
f
e
a
c
h
o
t
h
e
r,
l
o
o
k
in
g
d
if
f
e
re
n
t,
a
n
d
n
o
t
h
a
v
in
g
a
to
n
o
f
d
a
ta.
T
o
a
c
h
iev
e
th
is,
ima
g
e
e
n
h
a
n
c
e
m
e
n
t
tec
h
n
iq
u
e
s
we
re
a
p
p
li
e
d
u
sin
g
c
o
n
tras
t
e
n
h
a
n
c
e
m
e
n
t
a
lg
o
rit
h
m
,
c
o
n
tras
t
-
li
m
it
e
d
a
d
a
p
ti
v
e
h
ist
o
g
ra
m
e
q
u
a
li
z
a
t
io
n
(CLAHE)
,
a
n
d
ima
g
e
se
g
m
e
n
tatio
n
tec
h
n
iq
u
e
s
u
sin
g
c
o
lo
r
is
o
latio
n
a
re
e
m
p
lo
y
e
d
,
w
h
ich
c
o
n
tri
b
u
tes
to
m
o
re
a
c
c
u
ra
te
se
p
a
ra
ti
o
n
o
f
o
v
e
rlap
p
in
g
c
e
ll
s,
a
n
d
e
n
a
b
les
fa
ste
r
a
n
d
m
o
re
e
fficie
n
t
d
iag
n
o
sis.
T
o
e
fficie
n
tl
y
c
o
m
p
lete
th
e
c
las
sifica
ti
o
n
p
r
o
c
e
ss
a
fter
th
e
se
g
m
e
n
tatio
n
p
r
o
c
e
ss
,
a
n
e
u
ra
l
n
e
tw
o
rk
str
u
c
tu
re
c
o
n
s
isti
n
g
o
f
c
o
m
b
in
i
n
g
t
h
re
e
ty
p
e
s
o
f
c
o
n
v
o
lu
ti
o
n
a
l
la
y
e
rs
(d
e
p
th
wise
,
sp
a
ti
a
ll
y
,
a
n
d
c
o
n
v
o
lu
ti
o
n
)
wa
s
u
se
d
.
T
o
e
v
a
l
u
a
te
th
e
p
ro
p
o
se
d
tec
h
n
iq
u
e
,
e
x
p
e
rime
n
ts
we
re
c
o
n
d
u
c
ted
u
si
n
g
a
n
o
p
e
n
-
so
u
rc
e
b
l
o
o
d
c
e
ll
c
o
u
n
t
a
n
d
d
e
tec
ti
o
n
(BCCD)
d
a
tas
e
t
fro
m
th
e
Ka
g
g
le
p
latfo
rm
,
a
n
d
re
su
l
ted
in
a
c
h
iev
i
n
g
a
c
las
sifica
ti
o
n
a
c
c
u
ra
c
y
o
f
9
9
.
0
6
%
a
n
d
a
n
F
1
-
sc
o
re
o
f
9
9
.
0
5
%
.
Th
is
h
i
g
h
li
g
h
t
o
f
th
e
m
o
d
e
l’s
a
b
il
i
ty
t
o
e
fficie
n
t
ly
d
e
a
l
with
t
h
e
c
h
a
ll
e
n
g
e
s
a
ss
o
c
iate
d
with
WBC
c
las
sifica
ti
o
n
.
K
ey
w
o
r
d
s
:
C
L
AHE
Dee
p
lear
n
in
g
D
e
p
t
h
w
is
e
s
e
p
a
r
a
b
l
e
c
o
n
v
o
l
u
t
io
n
Sp
atially
s
ep
ar
ab
le
co
n
v
o
lu
tio
n
W
h
ite
b
lo
o
d
ce
lls
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
:
Fira
s
Mu
n
ea
m
B
ac
h
ay
Dep
ar
tm
en
t o
f
Scien
tific
Af
f
air
,
Un
iv
er
s
ity
o
f
Ku
f
a
Al
-
Z
u
h
u
r
R
o
ad
,
Naja
f
,
I
r
aq
E
m
ail:
f
ir
asm
.
b
ac
h
ay
@
u
o
k
u
f
a
.
ed
u
.
iq
1.
I
NT
RO
D
UCT
I
O
N
B
lo
o
d
is
an
ess
en
tial
elem
en
t
f
o
r
ass
ess
in
g
h
u
m
an
h
ea
lth
,
as
ea
ch
d
r
o
p
o
f
it
c
o
n
tain
s
m
u
ltip
le
co
m
p
o
n
en
ts
th
at
r
ev
ea
l
th
e
h
ea
lth
s
tatu
s
.
An
o
th
er
p
er
tain
in
g
ass
es
s
m
en
t
is
th
e
p
er
ip
h
er
al
b
lo
o
d
s
m
ea
r
,
wh
er
e
a
b
lo
o
d
s
am
p
le
s
p
r
ea
d
is
p
lace
d
o
n
a
g
lass
s
lid
e,
s
tain
ed
in
a
p
ar
ticu
lar
m
an
n
er
,
a
n
d
s
u
b
s
eq
u
en
tly
o
b
s
er
v
ed
in
a
m
icr
o
s
co
p
e.
T
h
is
test
f
ig
u
r
es
o
u
t
th
e
d
if
f
er
en
t
k
in
d
s
o
f
wh
ite
b
lo
o
d
ce
lls
(
W
B
C
s
)
an
d
te
lls
th
em
ap
ar
t
f
r
o
m
r
ed
b
lo
o
d
ce
lls
an
d
o
t
h
er
s
tu
f
f
,
g
iv
in
g
in
f
o
a
b
o
u
t
h
o
w
h
e
alth
y
s
o
m
eo
n
e
is
[
1
]
.
Fo
r
th
e
y
ar
e
ab
le
to
f
ig
h
t
b
ac
ter
ia,
v
ir
u
s
es,
an
d
o
th
er
g
er
m
s
,
W
B
C
s
ar
e
a
cr
u
cial
co
m
p
o
n
en
t o
f
th
e
im
m
u
n
e
s
y
s
tem
[
2
]
,
[
3
]
.
B
lo
o
d
d
is
o
r
d
er
ca
s
es,
lik
e
leu
k
em
ia
an
d
ly
m
p
h
o
m
a,
ar
e
in
c
r
ea
s
in
g
.
T
h
ese
ce
lls
ar
e
s
u
p
er
im
p
o
r
tan
t
b
ec
au
s
e
th
ey
k
ee
p
y
o
u
r
im
m
u
n
e
s
y
s
tem
wo
r
k
in
g
r
ig
h
t
[
4
]
–
[
6
]
.
T
h
ese
ce
lls
as
a
g
r
o
u
p
co
o
p
er
ate
to
en
h
a
n
ce
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
5
,
No
.
3
,
J
u
n
e
2
0
2
6
:
2
5
9
5
-
2
6
0
5
2596
th
e
im
m
u
n
ity
o
f
th
e
b
o
d
y
b
y
h
elp
in
g
in
r
em
o
v
al
o
f
f
o
r
eig
n
a
n
d
p
ath
o
g
en
ic
elem
en
ts
[
7
]
.
I
t
is
al
s
o
k
n
o
wn
th
at
th
e
b
lo
o
d
h
as
a
s
et
co
m
p
o
s
itio
n
th
at
is
co
n
s
tan
t
in
ea
ch
in
d
i
v
id
u
al,
wh
ic
h
co
n
s
is
ts
o
f
f
i
v
e
m
ain
ty
p
es
o
f
ce
lls
.
T
h
ese
ar
e
n
eu
tr
o
p
h
ils
(
4
0
-
6
0
%),
ly
m
p
h
o
cy
tes
(
2
0
-
4
0
%)
,
m
o
n
o
cy
tes
(
2
-
8
%),
eo
s
in
o
p
h
ils
(
1
-
4
%),
an
d
b
aso
p
h
ils
(
0
.
5
-
1
%).
Star
tin
g
f
r
o
m
th
o
s
e
with
th
e
lo
west
p
er
ce
n
tag
es,
b
aso
p
h
ils
ar
e
m
o
s
t
r
ar
ely
s
ee
n
wh
er
ea
s
ea
ch
ty
p
e
p
er
f
o
r
m
s
v
a
r
io
u
s
f
u
n
ctio
n
s
to
p
r
o
tect
t
h
e
b
o
d
y
[
8
]
,
[
9
]
.
C
h
an
g
es
in
W
B
C
s
n
u
m
b
er
,
o
r
in
th
eir
m
o
r
p
h
o
lo
g
y
s
u
ch
as
s
ize
an
d
s
h
ap
e
o
r
co
l
o
r
in
b
lo
o
d
s
m
ea
r
,
ar
e
s
ig
n
s
o
f
d
iv
e
r
s
e
b
o
d
y
d
is
o
r
d
er
s
.
I
n
cy
to
p
ath
o
l
o
g
y
,
th
e
W
B
C
m
o
r
p
h
o
lo
g
y
an
d
th
eir
co
u
n
t
in
th
e
s
lid
es
is
also
h
elp
f
u
l
in
th
e
d
iag
n
o
s
is
o
f
m
a
n
y
d
is
o
r
d
er
s
.
Hen
ce
,
th
e
i
d
en
tific
atio
n
an
d
ch
ar
ac
te
r
izatio
n
o
f
t
h
e
b
lo
o
d
ce
lls
in
th
is
co
n
tex
t,
co
r
r
esp
o
n
d
s
to
th
e
m
ec
h
an
is
m
s
th
at
p
r
o
m
o
te
e
ar
ly
d
iag
n
o
s
is
an
d
tr
ea
tm
en
t
o
f
m
an
y
d
is
ea
s
e
p
r
o
ce
s
s
es
[
6
]
,
[
1
0
]
,
[
1
1
]
.
C
o
n
v
en
tio
n
al
m
eth
o
d
s
u
s
e
b
lo
o
d
s
m
ea
r
as
s
u
ch
r
em
ain
s
co
s
tly
an
d
tim
e
co
n
s
u
m
in
g
as
it
r
elies
o
n
a
v
is
u
al
in
s
p
ec
tio
n
b
y
th
e
clin
ician
s
tr
ain
ed
o
n
th
e
h
em
ato
lo
g
y
.
Su
c
h
m
eth
o
d
s
ar
e
ted
i
o
u
s
an
d
ar
e
s
u
b
ject
to
h
u
m
a
n
f
allib
ilit
y
.
W
h
er
ea
s
with
co
m
p
u
ter
ass
is
t
v
is
io
n
s
y
s
tem
s
,
it
p
r
o
v
id
es
m
o
r
e
ef
f
ec
tiv
e
an
d
d
ep
en
d
ab
le
ap
p
r
o
ac
h
with
d
iag
n
o
s
tic
p
r
ec
is
io
n
a
n
d
less
h
u
m
an
in
ter
f
e
r
en
ce
as
well
as
v
ar
iab
ilit
y
[
1
2
]
–
[
1
4
]
.
B
u
t,
to
d
is
cr
im
in
ate
W
B
C
s
p
h
o
to
m
icr
o
g
r
a
p
h
ic
p
o
in
ter
s
ar
e
u
s
ed
al
o
n
g
with
t
h
e
im
ag
e
p
r
ep
r
o
ce
s
s
in
g
m
et
h
o
d
s
an
d
s
tatis
tical
m
ea
s
u
r
es
o
b
tain
ed
th
r
o
u
g
h
m
an
u
al
f
ea
tu
r
e
en
g
in
ee
r
in
g
.
All
th
ese
p
r
o
ce
d
u
r
es
ar
e
lab
o
r
i
o
u
s
an
d
ca
n
co
n
s
u
m
e
a
lo
t
o
f
tim
e,
an
d
th
e
ac
cu
r
ac
y
o
f
th
e
class
if
icatio
n
g
r
ea
tly
d
ep
en
d
s
o
n
th
e
m
an
u
ally
s
elec
ted
f
ea
tu
r
es,
wh
ich
lead
s
to
lo
w
ac
cu
r
ac
y
an
d
a
lo
n
g
tim
e
o
f
class
if
icatio
n
[
1
5
]
.
W
ith
th
e
d
ev
elo
p
m
en
t
o
f
c
o
m
p
u
tin
g
ca
p
ab
ilit
ies,
th
e
u
s
e
o
f
ar
tific
ial
in
tellig
en
ce
a
n
d
d
ee
p
lear
n
in
g
alg
o
r
ith
m
s
is
in
cr
ea
s
in
g
in
m
ed
ical
an
d
b
i
o
lo
g
ical
ap
p
licatio
n
s
.
T
h
ese
tech
n
iq
u
es
ar
e
u
s
e
d
in
v
ar
i
o
u
s
f
ield
s
,
f
r
o
m
m
e
d
ical
im
ag
e
an
aly
s
is
to
p
r
o
ce
s
s
in
g
b
ig
d
ata
o
f
t
h
e
h
u
m
an
g
en
o
m
e
.
T
h
er
e
f
o
r
e,
t
h
e
d
ev
elo
p
m
e
n
t
o
f
lo
w
-
co
s
t
an
d
r
eliab
le
s
y
s
tem
s
f
o
r
class
if
y
in
g
an
d
ev
alu
ati
n
g
WBCs
h
as
b
ec
o
m
e
im
p
er
ativ
e
[
1
6
]
.
Desp
ite
ad
v
an
ce
s
in
im
a
g
e
-
b
ased
class
if
icatio
n
o
f
W
B
C
s
,
th
e
f
ield
s
till
f
ac
es
s
ig
n
if
ican
t
ch
all
en
g
es.
Var
iatio
n
i
n
ce
ll
m
o
r
p
h
o
lo
g
y
is
af
f
ec
ted
b
y
p
atien
t
co
n
d
itio
n
a
n
d
s
tain
in
g
tech
n
i
q
u
es,
m
ak
i
n
g
it
d
if
f
icu
lt
to
d
ev
elo
p
ac
cu
r
ate
class
if
icatio
n
s
y
s
te
m
s
.
Oth
er
ce
llu
lar
co
m
p
o
n
en
ts
,
s
u
ch
as
r
ed
b
lo
o
d
c
ells
an
d
p
latelets,
co
m
p
licate
im
ag
e
an
aly
s
is
a
n
d
in
c
r
ea
s
e
lik
elih
o
o
d
o
f
class
if
icatio
n
er
r
o
r
s
.
I
n
ad
d
itio
n
,
lack
o
f
lar
g
e
an
d
d
iv
er
s
e
d
atasets
,
co
u
p
led
with
v
ar
iatio
n
s
in
illu
m
i
n
atio
n
an
d
s
tain
in
g
i
n
ten
s
ity
,
lim
it
th
e
ac
cu
r
ac
y
an
d
o
p
tim
izatio
n
o
f
s
y
s
tem
s
[
1
7
]
,
[
1
8
]
.
A
n
u
m
b
er
o
f
m
eth
o
d
o
lo
g
ies
b
ased
o
n
d
ee
p
lear
n
in
g
h
av
e
b
ee
n
p
r
o
p
o
s
ed
in
th
is
r
e
g
a
r
d
f
o
r
th
e
d
etec
tio
n
an
d
class
if
icatio
n
o
f
W
B
C
s
.
On
e
o
f
th
e
m
eth
o
d
s
,
n
am
ed
W
B
C
s
Net
u
tili
ze
d
th
e
d
ataset
-
m
aster
an
d
r
ep
o
r
ted
9
8
.
9
2
%
o
f
ac
c
u
r
ac
y
.
Ho
wev
er
,
th
is
m
eth
o
d
d
i
d
n
o
t
tak
e
in
to
c
o
n
s
id
er
atio
n
t
h
e
s
eg
m
en
tatio
n
o
f
W
B
C
im
ag
es
f
r
o
m
th
e
m
ain
i
m
ag
e
an
d
u
s
ed
p
r
e
-
t
r
ain
ed
d
e
ep
lear
n
in
g
m
o
d
els
an
d
th
er
e
f
o
r
e
d
etailed
an
aly
s
is
was
n
o
t
p
o
s
s
ib
le
[
1
5
]
.
Similar
ly
,
an
o
t
h
er
s
tu
d
y
u
s
ed
Alex
N
et,
Go
o
g
L
eNe
t,
a
n
d
R
esNet
-
5
0
with
o
p
e
n
-
ac
ce
s
s
d
ataset
f
r
o
m
Kag
g
le
an
d
s
h
o
wed
a
9
7
.
9
5
%
ac
cu
r
ac
y
b
u
t
a
g
ain
d
id
n
o
t
ex
tr
ac
t
W
B
C
im
ag
es
f
r
o
m
th
e
m
ain
im
ag
e
h
en
ce
n
o
t
af
f
ec
tin
g
th
e
g
r
an
u
lar
ity
o
f
th
eir
a
n
aly
s
is
[
2
]
.
T
h
e
co
m
b
in
atio
n
o
f
R
esNet
-
5
0
,
VGG
-
1
9
,
an
d
Mo
b
ileNet
-
V3
-
s
m
all
with
R
a
ab
in
-
W
B
C
d
ata
s
et
r
esu
lted
in
9
8
.
8
6
%
ac
cu
r
ac
y
.
W
h
ile
ef
f
icien
t,
th
is
m
eth
o
d
was
also
with
o
u
t
s
eg
m
en
tatio
n
o
f
W
B
C
s
f
r
o
m
th
e
p
r
im
a
r
y
p
ictu
r
es
an
d
p
r
im
ar
ily
d
ep
e
n
d
ed
o
n
p
r
e
-
tr
ain
ed
m
o
d
els
[
3
]
.
An
o
th
er
wo
r
k
u
s
in
g
c
o
n
v
o
lu
tio
n
al
n
eu
r
al
n
etw
o
r
k
(
C
NN
)
o
n
th
e
p
u
b
licly
a
v
ailab
le
d
ataset
f
r
o
m
Kag
g
le
ac
h
iev
e
d
h
ig
h
ac
cu
r
ac
y
9
9
.
5
%,
b
u
t
it
d
id
n
o
t
p
er
f
o
r
m
th
e
e
x
tr
ac
tio
n
o
f
W
B
C
im
ag
es
f
r
o
m
th
e
m
ain
im
ag
e,
wh
ich
im
p
ac
ted
th
e
r
o
b
u
s
tn
ess
o
f
th
e
class
if
icatio
n
[
1
6
]
.
Den
s
eNe
t
-
1
2
1
,
wh
en
ap
p
lied
to
th
e
o
p
en
-
ac
ce
s
s
d
ataset
f
r
o
m
Kag
g
le,
attain
ed
a
n
ac
cu
r
ac
y
o
f
9
8
.
8
4
%.
T
h
is
s
tu
d
y
h
ad
s
im
ilar
lim
itatio
n
s
b
y
n
o
t
e
x
tr
ac
ti
n
g
W
B
C
im
ag
es
an
d
u
s
in
g
p
r
e
-
tr
ain
ed
m
o
d
els,
p
o
ten
tially
af
f
ec
tin
g
f
ea
tu
r
e
ex
tr
ac
tio
n
s
p
ec
if
icity
[
1
9
]
.
M
o
b
ileNet,
ap
p
lie
d
to
th
e
s
am
e
d
ataset,
ac
h
ie
v
ed
9
8
.
4
%
ac
cu
r
ac
y
a
n
d
in
cl
u
d
ed
s
eg
m
en
tatio
n
.
Ho
wev
e
r
,
it
wa
s
b
ased
o
n
p
r
e
-
tr
ain
e
d
d
ee
p
le
ar
n
in
g
m
o
d
els
an
d
m
ay
h
av
e
m
is
s
ed
im
p
o
r
tan
t/u
n
iq
u
e
f
ea
tu
r
es
f
r
o
m
th
is
d
ataset
[
2
0
]
.
Den
s
eNe
-
t2
0
1
an
d
Dar
k
N
et
-
5
3
,
wh
ic
h
wo
r
k
ed
o
n
an
o
p
en
-
ac
ce
s
s
d
ataset
s
y
n
th
esized
f
r
o
m
a
r
ea
l
-
wo
r
ld
d
ataset,
r
ep
o
r
ted
a
cc
u
r
ac
y
o
f
9
9
.
9
%.
Ho
wev
er
,
th
e
u
s
e
o
f
s
y
n
th
etic
d
ata
an
d
p
r
e
-
tr
ain
e
d
m
o
d
els
with
o
u
t
th
e
ex
tr
ac
tio
n
o
f
W
B
C
im
ag
es
m
ak
es
th
e
f
in
d
in
g
s
o
f
th
is
s
tu
d
y
p
o
s
s
ib
ly
in
ap
p
licab
le
[
1
3
]
.
Sti
ll,
an
o
th
er
C
NN
-
b
ased
s
tu
d
y
u
s
ed
o
p
e
n
-
ac
ce
s
s
d
ata
f
r
o
m
Kag
g
le,
d
id
n
'
t e
x
tr
ac
t WB
C
i
m
ag
es f
r
o
m
t
h
e
b
ac
k
g
r
o
u
n
d
p
i
ctu
r
e,
an
d
r
ep
o
r
ted
9
8
% a
cc
u
r
ac
y
[
1
7
]
.
Fu
r
th
er
r
esear
ch
u
s
in
g
C
NN
with
th
e
s
am
e
d
ataset
ac
h
iev
ed
9
8
.
3
3
%
ac
cu
r
ac
y
b
u
t
also
d
id
n
o
t
s
ep
ar
ate
W
B
C
im
ag
es
an
d
th
u
s
is
co
n
s
is
ten
t
with
th
e
lim
itatio
n
s
o
f
p
r
ev
io
u
s
wo
r
k
s
[
2
1
]
.
C
o
n
v
o
lu
tio
n
a
l
atten
tio
n
B
L
STM
n
etwo
r
k
(
C
AB
-
Net)
,
wh
en
test
ed
u
s
in
g
th
e
p
er
ip
h
er
al
b
lo
o
d
ce
ll
(
PB
C
)
an
d
b
lo
o
d
ce
ll
co
u
n
t
an
d
d
etec
tio
n
(
B
C
C
D)
d
atasets
,
ac
h
iev
ed
9
8
.
2
9
%
ac
cu
r
ac
y
b
u
t
was
lim
ited
b
y
n
o
t
b
ein
g
ab
le
to
ex
tr
ac
t
W
B
C
im
ag
es
f
r
o
m
th
e
p
r
im
ar
y
im
ag
e,
h
en
ce
lim
itin
g
th
e
d
etailed
m
o
r
p
h
o
l
o
g
ical
an
al
y
s
is
[
1
8
]
.
L
astl
y
,
a
co
m
b
in
ed
C
NN
an
d
r
ec
u
r
r
e
n
t
n
eu
r
al
n
etwo
r
k
(
R
NN
)
ap
p
r
o
ac
h
u
s
in
g
th
e
p
u
b
licly
av
aila
b
le
Kag
g
le
d
ataset
ac
h
iev
ed
9
5
.
8
9
%
[
2
2
]
.
T
h
e
cu
r
r
en
t
r
esear
ch
d
i
d
n
'
t
s
eg
m
en
t
th
e
W
B
C
im
ag
es
f
r
o
m
th
e
p
ar
en
t
im
ag
e
an
d
also
u
s
ed
p
r
e
-
tr
ai
n
ed
n
etwo
r
k
s
,
th
a
t m
ig
h
t n
o
t c
a
p
tu
r
e
all
th
e
s
u
b
t
leties in
th
e
d
ataset.
Ab
o
z
eid
et
a
l
.
[
2
3
]
h
av
e
u
s
e
d
p
r
e
tr
a
in
ed
m
o
d
e
l
(
Op
-
YO
L
Ov
8
)
o
v
er
B
C
C
D
d
at
as
et
an
d
o
b
t
ain
ed
p
er
f
o
r
m
an
c
e
(
9
9
.
2
%
ac
cu
r
a
cy
)
w
i
th
li
tt
le
c
u
s
to
m
iz
at
io
n
s
.
C
h
en
et
a
l
.
[
2
4
]
in
tr
o
d
u
c
ed
a
d
u
al
a
tt
en
t
io
n
m
o
d
e
l
(
DA
F
F
Net
)
ev
alu
at
ed
o
n
s
ix
d
if
f
e
r
en
t
d
a
ta
s
et
s
,
w
ith
ac
cu
r
ac
i
e
s
b
e
tw
ee
n
9
1
.
3
0
%
to
9
9
.
7
1
%,
b
u
t
th
e
w
in
n
er
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
Utiliz
a
tio
n
o
f d
ep
th
-
w
is
e
a
n
d
s
p
a
tia
lly
s
ep
a
r
a
b
le
co
n
vo
lu
tio
n
a
l n
etw
o
r
k
fu
s
io
n
…
(
F
ir
a
s
Mu
n
ea
m
B
a
c
h
a
y
)
2597
m
o
d
el
p
r
e
s
en
ted
a
r
el
at
iv
e
ly
h
ig
h
co
m
p
u
ta
tio
n
al
co
m
p
le
x
ity
,
wh
ich
m
ay
m
i
tig
at
e
th
e
r
ea
l
d
ep
lo
y
m
en
t.
Fay
y
az
e
t
a
l
.
[
2
5
]
p
r
o
p
o
s
e
d
a
3
-
s
tag
e
C
N
N
m
o
d
e
l
to
t
h
e
B
C
C
D
d
at
a
s
e
t,
w
it
h
a
r
ep
o
r
ted
o
v
er
a
ll
a
cc
u
r
ac
y
o
f
9
7
.
5
9
%
,
b
u
t
i
t
d
o
e
s
n
o
t
h
av
e
a
s
eg
m
en
ta
tio
n
s
tep
an
d
h
as
a
f
a
ir
ly
co
m
p
l
ic
at
ed
p
i
p
e
lin
e.
T
h
e
ex
i
s
t
in
g
li
ter
atu
r
e
h
a
s
g
en
er
al
ly
s
h
a
r
e
d
s
o
m
e
l
im
it
at
io
n
s
,
s
u
ch
a
s
u
s
i
n
g
a
lr
e
ad
y
p
r
e
tr
a
in
ed
m
o
d
el
s
o
f
d
ee
p
lea
r
n
in
g
an
d
f
ai
lin
g
to
ex
tr
ac
t
W
B
C
im
ag
e
s
f
r
o
m
o
r
ig
in
a
l
im
a
g
e
s
.
E
v
en
if
it
a
ch
i
ev
ed
h
ig
h
a
cc
u
r
a
cy
,
th
e
s
e
t
wo
f
ac
to
r
s
a
ctu
al
ly
li
m
i
t
th
e
m
o
d
el
s
'
ab
i
li
ty
in
e
x
tr
ac
tin
g
u
n
iq
u
e
f
ea
tu
r
e
s
n
ec
es
s
ar
y
f
o
r
th
e
ac
cu
r
a
te
cla
s
s
if
i
ca
tio
n
o
f
W
B
C
s
.
I
n
o
r
d
e
r
to
r
eso
lv
e
th
ese
is
s
u
es,
th
is
p
ap
er
p
r
esen
ts
an
ap
p
r
o
ac
h
th
at
em
p
l
o
y
s
th
e
co
n
tr
a
s
t
-
lim
ited
ad
ap
tiv
e
h
is
to
g
r
am
e
q
u
aliza
tio
n
(
C
L
AHE
)
m
eth
o
d
f
o
r
im
ag
e
ac
cu
r
ac
y
im
p
r
o
v
em
en
t
an
d
ch
r
o
m
atic
s
p
in
n
in
g
-
b
ased
s
eg
m
en
tatio
n
o
f
b
lo
o
d
s
m
ea
r
im
ag
es.
T
h
is
is
an
im
ag
e
c
o
n
tr
ast
en
h
a
n
ce
m
en
t
tech
n
iq
u
e
th
at
h
elp
s
v
is
u
al
in
te
r
p
r
etatio
n
an
d
p
e
r
f
o
r
m
an
ce
f
o
r
s
u
b
s
eq
u
e
n
ce
task
s
s
u
ch
as
im
ag
e
s
eg
m
en
tatio
n
.
Stan
d
ar
d
m
eth
o
d
s
h
a
v
e
s
ev
er
al
lim
itati
o
n
s
,
an
d
in
o
r
d
er
to
o
v
er
c
o
m
e
th
ese
lim
itatio
n
s
C
L
AHE
was
in
tr
o
d
u
ce
d
wh
ich
en
h
an
ce
s
lo
ca
l
co
n
tr
ast
[
2
6
]
,
[
2
7
]
.
A
d
ee
p
lea
r
n
in
g
m
o
d
el
with
th
r
ee
s
o
p
h
is
ticated
ty
p
e
s
o
f
co
n
v
o
lu
tio
n
a
l
lay
er
s
f
o
r
f
ea
tu
r
e
ex
tr
ac
tio
n
a
n
d
W
B
C
cla
s
s
if
icatio
n
was
d
ev
elo
p
ed
in
th
is
s
tu
d
y
.
T
h
e
k
ey
co
n
tr
ib
u
tio
n
s
o
f
th
is
r
esear
ch
ar
e
as f
o
llo
ws:
i)
Usi
n
g
th
e
C
L
AHE
m
eth
o
d
to
en
h
an
ce
im
a
g
e
ac
cu
r
ac
y
.
ii)
I
m
p
lem
en
tin
g
a
n
ew
s
eg
m
e
n
tatio
n
ap
p
r
o
ac
h
u
s
in
g
ch
r
o
m
atic
s
p
in
n
in
g
.
iii)
Dev
elo
p
in
g
a
r
o
b
u
s
t
d
ee
p
lear
n
in
g
m
o
d
el
with
th
r
ee
ad
v
an
ce
d
c
o
n
v
o
lu
tio
n
al
lay
e
r
s
f
o
r
W
B
C
class
if
icatio
n
.
iv
)
Pro
v
id
in
g
a
n
ef
f
icien
t a
n
d
r
eliab
le
s
o
lu
tio
n
f
o
r
ea
r
ly
d
iag
n
o
s
is
b
y
ex
am
in
in
g
b
lo
o
d
s
m
ea
r
s
.
2.
M
E
T
H
O
D
I
n
th
is
s
ec
tio
n
,
t
h
e
m
eth
o
d
o
l
o
g
y
u
s
ed
in
th
e
p
r
o
p
o
s
ed
a
p
p
r
o
ac
h
f
o
r
class
if
y
in
g
W
B
C
s
will
b
e
r
ev
iewe
d
in
d
etail
.
I
t
co
m
p
r
is
es
o
f
two
p
r
im
ar
y
s
tag
es:
p
r
ep
r
o
ce
s
s
in
g
im
a
g
e
an
d
f
ea
t
u
r
e
ex
tr
ac
tio
n
a
n
d
class
if
icatio
n
u
s
in
g
th
e
p
r
o
p
o
s
ed
m
o
d
el.
Fig
u
r
e
1
illu
s
tr
ates th
e
b
asic step
s
o
f
th
is
ap
p
r
o
ac
h
.
Fig
u
r
e
1
.
T
h
e
g
e
n
er
al
s
tr
u
ctu
r
e
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
2
.
1
.
P
re
pro
ce
s
s
ing
im
a
g
e
s
t
a
g
e
T
o
im
p
r
o
v
e
th
e
e
f
f
icien
cy
an
d
s
p
ee
d
o
f
d
iag
n
o
s
is
in
ca
s
es
o
f
ce
ll
o
v
er
lap
in
b
lo
o
d
s
m
e
ar
im
ag
es,
ad
v
an
ce
d
im
ag
e
p
r
o
ce
s
s
in
g
te
ch
n
iq
u
es
s
u
ch
as
th
e
C
L
AHE
alg
o
r
ith
m
a
r
e
u
s
ed
to
en
h
a
n
ce
th
e
c
o
n
tr
ast
o
f
th
e
im
ag
es.
T
h
e
im
ag
e
is
tr
an
s
f
o
r
m
ed
in
to
th
e
L
ab
co
lo
r
m
o
d
el,
an
d
th
en
th
e
C
L
AHE
alg
o
r
ith
m
is
u
s
ed
o
n
th
e
lig
h
t
ch
an
n
el
(
th
e
L
ch
an
n
el)
to
b
o
o
s
t
co
n
tr
ast
wh
ile
k
ee
p
in
g
th
e
c
o
lo
r
c
o
m
p
o
n
en
ts
in
t
ac
t.
T
h
is
p
r
o
ce
d
u
r
e
en
h
an
ce
s
th
e
d
etails
in
ar
ea
s
o
f
lo
w
co
n
tr
ast,
wh
ile
r
ed
u
cin
g
n
o
is
e.
Af
ter
co
n
tr
ast
e
n
h
an
c
em
en
t,
p
r
ep
ar
atio
n
f
o
r
t
h
e
c
o
lo
r
is
o
latio
n
p
r
o
ce
s
s
is
ca
r
r
ied
o
u
t
b
y
c
o
n
v
e
r
tin
g
th
e
im
ag
e
to
th
e
HSV
co
lo
r
m
o
d
el.
I
n
t
h
is
m
o
d
el,
th
e
co
lo
r
v
io
let
⎯
wh
ich
u
s
u
al
ly
r
ep
r
esen
ts
W
B
C
s
⎯
is
d
eter
m
in
ed
b
y
s
ettin
g
lo
wer
a
n
d
u
p
p
er
lim
its
f
o
r
t
h
e
co
lo
r
v
alu
es.
T
o
en
s
u
r
e
th
at
an
im
ag
e
co
n
tain
in
g
o
n
ly
W
B
C
s
is
ex
tr
ac
ted
,
th
e
lar
g
est
ar
ea
is
s
elec
ted
an
d
cr
o
p
p
e
d
af
ter
co
lo
r
is
o
latio
n
p
r
o
ce
s
s
.
Fig
u
r
e
2
illu
s
tr
ates
t
h
e
d
if
f
er
en
t
s
tag
es
o
f
im
ag
e
p
r
o
ce
s
s
in
g
,
s
tar
tin
g
f
r
o
m
c
o
n
tr
ast en
h
a
n
ce
m
en
t u
s
in
g
C
L
AHE
an
d
en
d
in
g
with
t
h
e
ex
tr
ac
tio
n
o
f
WBCs
.
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
5
,
No
.
3
,
J
u
n
e
2
0
2
6
:
2
5
9
5
-
2
6
0
5
2598
Fig
u
r
e
2
.
Stag
es o
f
th
e
p
r
o
p
o
s
ed
WBC
s
eg
m
en
tatio
n
p
r
o
ce
s
s
2
.
2
.
F
e
a
t
ure
ex
t
r
a
ct
io
n a
nd
cla
s
s
if
ica
t
io
n sta
g
e
I
n
th
is
s
tag
e,
a
d
ee
p
lear
n
in
g
m
o
d
el
s
p
ec
if
ically
d
esig
n
ed
f
o
r
p
r
o
ce
s
s
in
g
W
B
C
i
m
ag
es
was
d
ev
elo
p
e
d
.
T
h
e
m
o
d
el
co
n
s
is
ts
o
f
th
r
ee
ty
p
es
o
f
co
n
v
o
lu
ti
o
n
al
lay
er
s
.
T
h
e
p
r
o
ce
s
s
in
g
p
r
o
ce
s
s
b
eg
in
s
with
ad
ju
s
tin
g
th
e
in
p
u
t
lay
er
to
ac
co
m
m
o
d
ate
th
e
c
o
lo
r
WBC
i
m
ag
es
with
a
r
eso
lu
tio
n
o
f
(
6
4
×6
4
)
p
ix
els.
First,
two
co
n
v
o
l
u
tio
n
al
lay
er
s
ar
e
u
s
ed
,
ea
ch
co
n
tain
in
g
1
2
8
k
e
r
n
els,
f
o
llo
wed
b
y
a
n
o
r
m
aliza
ti
o
n
lay
er
a
n
d
th
e
n
a
p
o
o
lin
g
lay
er
,
wh
er
e
p
o
o
lin
g
o
p
er
atio
n
s
a
r
e
ap
p
lied
u
s
in
g
m
ax
p
o
o
lin
g
win
d
o
ws.
T
h
is
is
f
o
llo
wed
b
y
tw
o
co
n
v
o
l
u
tio
n
al
lay
er
s
ac
r
o
s
s
s
p
atial
d
im
en
s
io
n
s
,
wh
er
e
th
e
f
ir
s
t
co
n
tain
s
6
4
k
er
n
els
to
o
p
e
r
ate
o
n
th
e
v
er
tical
d
im
en
s
io
n
,
wh
ile
t
h
e
s
ec
o
n
d
o
p
er
ates
o
n
t
h
e
h
o
r
izo
n
tal
d
im
e
n
s
io
n
with
th
e
s
am
e
n
u
m
b
er
o
f
k
er
n
els,
f
o
llo
wed
b
y
a
p
o
o
lin
g
lay
e
r
.
Af
ter
th
at,
th
e
th
ir
d
ty
p
e
o
f
co
n
v
o
l
u
tio
n
al
lay
er
s
,
th
e
d
ep
th
wis
e
c
o
n
v
o
lu
tio
n
lay
er
,
is
u
s
ed
.
I
n
th
is
lay
er
,
a
s
ep
ar
ate
f
ilter
i
s
ap
p
lied
to
ea
ch
in
p
u
t c
h
an
n
e
l,
an
d
th
en
th
e
in
p
u
t c
h
a
n
n
els ar
e
co
m
b
in
ed
u
s
in
g
a
co
n
v
o
lu
tio
n
al
lay
e
r
with
a
k
er
n
el
s
ize
o
f
(
1
×
1
)
,
f
o
llo
we
d
b
y
a
p
o
o
lin
g
la
y
er
.
T
o
alig
n
th
e
o
u
tp
u
ts
o
f
th
e
co
n
v
o
l
u
tio
n
al
lay
er
s
with
th
e
in
p
u
ts
o
f
th
e
f
u
lly
c
o
n
n
ec
ted
lay
er
s
,
a
g
lo
b
al
av
er
a
g
e
p
o
o
li
n
g
lay
er
is
ap
p
lied
.
T
h
en
,
two
f
u
lly
c
o
n
n
ec
ted
lay
er
s
ar
e
u
s
ed
,
wh
ic
h
a
d
ap
t
to
t
h
e
n
u
m
b
er
o
f
n
o
d
es
in
th
e
p
r
e
v
io
u
s
lay
er
an
d
th
e
n
u
m
b
er
o
f
d
esire
d
class
es.
T
o
im
p
r
o
v
e
g
en
er
aliza
tio
n
a
n
d
r
ed
u
ce
th
e
r
is
k
o
f
o
v
e
r
f
itti
n
g
to
th
e
tr
ain
in
g
d
ata,
a
d
r
o
p
o
u
t
lay
e
r
is
ad
d
e
d
af
ter
ea
ch
f
u
lly
co
n
n
ec
ted
lay
e
r
.
Fin
ally
,
a
f
in
al
lay
e
r
co
n
s
is
ti
n
g
o
f
th
e
s
o
f
tm
ax
f
u
n
ctio
n
,
wh
ich
is
wid
ely
u
s
e
d
in
m
u
lti
-
class
class
if
icatio
n
p
r
o
b
lem
s
,
is
in
clu
d
ed
.
T
h
is
lay
er
is
ad
ju
s
ted
to
m
atch
th
e
n
u
m
b
er
o
f
tar
g
et
cla
s
s
es.
T
h
e
o
p
tim
al
d
esig
n
o
f
t
h
e
p
r
o
p
o
s
ed
m
o
d
el
was
d
eter
m
in
e
d
ex
p
e
r
im
en
tally
b
y
g
r
ad
u
ally
ad
ju
s
tin
g
th
e
n
u
m
b
er
o
f
co
n
v
o
lu
tio
n
al
l
ay
er
s
,
th
e
m
ax
im
u
m
p
o
o
lin
g
s
ize,
an
d
th
e
n
u
m
b
er
o
f
f
ilter
s
.
T
h
is
p
r
o
ce
s
s
was
co
n
tin
u
ed
u
n
til
th
e
ar
ch
itectu
r
e
with
th
e
b
est
p
er
f
o
r
m
an
ce
w
as
s
elec
ted
.
T
ab
le
1
s
h
o
ws
th
e
d
etailed
d
esig
n
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el,
wh
ile
Fig
u
r
e
3
p
r
o
v
i
d
es a
v
is
u
al
d
escr
ip
tio
n
o
f
th
e
m
o
d
el
a
r
ch
itectu
r
e.
T
ab
le
1
.
T
h
e
p
r
o
p
o
s
ed
m
o
d
el
ar
ch
itectu
r
e
La
y
e
r
K
e
r
n
el
si
z
e
K
e
r
n
el
N
o
.
I
n
p
ut
s
h
a
p
e
O
u
t
p
ut
s
h
a
p
e
P
a
r
a
m
e
t
e
r
n
o.
C
o
n
v
o
l
u
t
i
o
n
a
l
1
3
×
3
1
2
8
6
4
,
6
4
,
3
6
2
,
6
2
,
1
2
8
3
,
5
6
8
C
o
n
v
o
l
u
t
i
o
n
a
l
2
3
×
3
1
2
8
6
2
,
6
2
,
1
2
8
6
0
,
6
0
,
1
2
8
1
4
7
,
5
8
4
B
a
t
c
h
n
o
r
ma
l
i
z
a
t
i
o
n
-
-
6
0
,
6
0
,
1
2
8
6
0
,
6
0
,
1
2
8
5
1
2
M
a
x
p
o
o
l
.
1
2
×
2
-
6
0
,
6
0
,
1
2
8
3
0
,
3
0
,
1
2
8
0
H
o
r
i
z
o
n
t
a
l
c
o
n
v
o
l
u
t
i
o
n
1
×
3
64
3
0
,
3
0
,
1
2
8
3
0
,
2
8
,
64
2
4
,
6
4
0
V
e
r
t
i
c
a
l
c
o
n
v
o
l
u
t
i
o
n
3
×
1
64
3
0
,
2
8
,
64
2
8
,
2
8
,
64
1
2
,
3
5
2
M
a
x
p
o
o
l
.
2
2
×
2
-
2
8
,
2
8
,
64
1
4
,
1
4
,
64
0
D
e
p
t
h
w
i
s
e
c
o
n
v
o
l
u
t
i
o
n
3
×
3
1
1
4
,
1
4
,
64
1
2
,
1
2
,
64
6
4
0
P
o
i
n
t
w
i
se
c
o
n
v
o
l
u
t
i
o
n
1
×
1
32
1
2
,
1
2
,
64
1
2
,
1
2
,
32
2
,
0
8
0
M
a
x
p
o
o
l
.
3
2
×
2
-
1
2
,
1
2
,
32
6,
6,
32
0
G
l
o
b
a
l
a
v
e
r
a
g
e
p
o
o
l
i
n
g
-
-
6,
6,
32
32
0
F
u
l
l
y
c
o
n
n
e
c
t
e
d
-
-
32
5
1
2
1
6
,
8
9
6
D
r
o
p
o
u
t
(
0
.
5
)
-
-
5
1
2
5
1
2
0
F
u
l
l
y
c
o
n
n
e
c
t
e
d
-
-
5
1
2
2
5
6
1
3
1
,
3
2
8
D
r
o
p
o
u
t
(
0
.
5
)
-
-
2
5
6
2
5
6
0
S
o
f
t
ma
x
-
-
2
5
6
4
1
,
0
2
8
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
Utiliz
a
tio
n
o
f d
ep
th
-
w
is
e
a
n
d
s
p
a
tia
lly
s
ep
a
r
a
b
le
co
n
vo
lu
tio
n
a
l n
etw
o
r
k
fu
s
io
n
…
(
F
ir
a
s
Mu
n
ea
m
B
a
c
h
a
y
)
2599
Fig
u
r
e
3
.
Ar
c
h
itectu
r
e
o
f
th
e
s
u
g
g
ested
d
ee
p
lear
n
in
g
m
o
d
el
f
o
r
W
B
C
clas
s
if
icatio
n
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
A
n
ew
m
eth
o
d
f
o
r
th
e
class
if
icatio
n
o
f
leu
co
cy
tes
in
th
e
W
B
C
s
was
d
ev
elo
p
ed
u
tili
zin
g
th
e
f
r
ee
en
v
ir
o
n
m
en
t
f
o
r
p
r
o
g
r
am
m
in
g
Go
o
g
le
C
o
lab
b
ased
o
n
Py
th
o
n
3
.
E
ac
h
o
f
th
e
ex
p
e
r
im
en
ts
was
p
er
f
o
r
m
ed
o
n
a
1
2
GB
Nv
id
ia
T
esla
K8
0
GPU
with
1
3
GB
o
f
R
AM
.
Fo
r
th
e
ev
alu
atio
n
,
th
e
o
p
e
n
-
ac
ce
s
s
d
ataset
p
r
esen
ted
o
n
Kag
g
le
was
u
s
ed
.
T
h
e
ef
f
ec
tiv
en
ess
o
f
th
e
s
y
s
tem
u
n
d
er
co
n
s
id
er
atio
n
was
m
ea
s
u
r
ed
with
th
e
u
s
e
o
f
ev
alu
atio
n
in
d
icato
r
s
s
u
ch
as
a
cc
u
r
ac
y
an
d
F1
-
s
co
r
e
.
3
.
1
.
Da
t
a
s
et
T
h
e
B
C
C
D
d
ata
s
et
co
n
s
is
ts
o
f
1
2
,
5
0
0
m
ag
n
i
f
ied
im
ag
es
o
f
b
lo
o
d
ce
lls
in
J
PEG
f
o
r
m
at
[
2
8
]
.
I
t
in
clu
d
es
ab
o
u
t
3
,
0
0
0
im
ag
es
f
o
r
ea
ch
o
f
t
h
e
f
o
u
r
ty
p
es
o
f
W
B
C
s
:
eo
s
in
o
p
h
ils
,
ly
m
p
h
o
c
y
tes,
m
o
n
o
cy
tes,
an
d
n
eu
tr
o
p
h
ils
.
E
ac
h
im
ag
e
h
as
a
r
eso
lu
tio
n
o
f
3
2
0
×
2
4
0
p
ix
el
s
,
u
s
es
3
co
lo
r
ch
an
n
els
(
R
G
B
)
,
an
d
h
as
a
p
ix
el
d
en
s
ity
o
f
9
6
d
o
ts
p
er
in
ch
(
D
PI)
.
Fig
u
r
e
4
d
is
p
lay
s
s
am
p
le
im
ag
es o
f
WBCs
f
r
o
m
th
e
d
ata
s
et.
Fig
u
r
e
4
.
Sam
p
les o
f
W
B
C
s
i
m
ag
es
3
.
2
.
Seg
m
ent
a
t
i
o
n r
esu
lt
s
T
h
e
p
r
o
p
o
s
ed
m
eth
o
d
d
escr
ib
ed
in
s
ec
tio
n
2
.
1
was
u
s
ed
to
ex
tr
ac
t
WBC
im
ag
es
f
r
o
m
b
lo
o
d
s
m
ea
r
im
ag
es
an
d
en
h
an
ce
th
ei
r
r
eso
lu
tio
n
.
E
ac
h
cr
o
p
p
ed
WBC
im
ag
e
was
6
4
×6
4
p
ix
els
in
s
ize.
Fig
u
r
e
5
s
h
o
ws
th
e
r
esu
ltin
g
im
ag
es f
r
o
m
th
e
en
h
an
ce
m
en
t a
n
d
s
eg
m
en
tatio
n
s
tag
es.
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
5
,
No
.
3
,
J
u
n
e
2
0
2
6
:
2
5
9
5
-
2
6
0
5
2600
Fig
u
r
e
5
.
Seg
m
e
n
tatio
n
s
tag
es
3
.
3
.
E
v
a
lua
t
ing
t
he
pro
po
s
ed
t
ec
hn
iqu
e'
s
perf
o
rm
a
nce
T
h
e
p
r
o
p
o
s
ed
tech
n
iq
u
e
in
th
is
s
tu
d
y
was
ev
alu
ated
u
s
i
n
g
two
m
ain
in
d
icato
r
s
:
ac
c
u
r
ac
y
an
d
F1
-
s
co
r
e.
T
h
e
ac
c
u
r
ac
y
m
ea
s
u
r
e
is
d
ef
in
ed
ac
c
o
r
d
in
g
to
(
1
)
.
=
+
+
+
+
(
1
)
T
r
u
e
p
o
s
itiv
e
r
ep
r
esen
ts
th
e
n
u
m
b
er
o
f
ca
s
es
wh
er
e
W
B
C
s
wer
e
co
r
r
ec
tly
class
if
ie
d
an
d
s
u
cc
ess
f
u
lly
id
en
tifie
d
.
W
h
ile
tr
u
e
n
e
g
ativ
e
r
ef
er
s
to
th
e
n
u
m
b
er
o
f
ca
s
es
wh
er
e
W
B
C
s
wer
e
co
r
r
ec
tly
ex
clu
d
ed
.
On
th
e
o
th
er
h
a
n
d
,
f
alse
p
o
s
itiv
e
r
ef
e
r
s
to
th
e
n
u
m
b
er
o
f
ca
s
es
wh
er
e
W
B
C
s
wer
e
in
co
r
r
ec
tly
cl
ass
if
ied
as
co
r
r
ec
t,
wh
ile
f
alse n
eg
ativ
e
r
ef
er
s
to
th
e
n
u
m
b
er
o
f
ca
s
es wh
er
e
W
B
C
s
wer
e
in
co
r
r
ec
tly
ex
clu
d
e
d
as in
co
r
r
ec
t.
T
h
e
F1
-
s
co
r
e
r
ep
r
esen
ts
th
e
h
ar
m
o
n
ic
m
ea
n
b
etwe
en
p
r
ec
is
io
n
an
d
r
ec
all
.
I
t
is
ca
lcu
lated
u
s
in
g
(
2
)
.
B
o
th
p
r
ec
is
io
n
an
d
r
ec
all
ca
n
b
e
ca
lcu
lated
m
ath
em
atica
lly
ac
co
r
d
in
g
to
(
3
)
an
d
(
4
)
.
1
−
=
2
×
×
+
(
2
)
=
+
(
3
)
=
+
(
4
)
T
h
e
tr
ain
in
g
p
r
o
ce
s
s
u
tili
ze
d
Ker
as,
wh
ich
is
a
p
o
p
u
lar
d
ee
p
lear
n
in
g
f
r
am
ew
o
r
k
.
I
n
th
e
f
in
al
lay
er
,
th
e
s
o
f
tm
ax
f
u
n
ctio
n
was
ap
p
lied
to
tr
an
s
f
o
r
m
th
e
o
u
t
p
u
ts
i
n
to
class
p
r
o
b
a
b
ilit
ies.
T
h
e
c
o
n
v
o
lu
ti
o
n
al
lay
e
r
s
s
tar
ted
with
weig
h
ts
s
et
b
y
th
e
l
ea
k
y
r
ec
tifie
d
lin
ea
r
u
n
it
(
R
eL
U
)
f
u
n
ctio
n
,
wh
ich
is
a
v
ar
iatio
n
o
f
th
e
s
tan
d
ar
d
R
eL
U
f
u
n
ctio
n
.
T
h
e
Ad
am
m
eth
o
d
was
u
s
ed
t
o
ad
ju
s
t
th
e
weig
h
ts
d
u
r
i
n
g
tr
ai
n
in
g
,
with
a
c
o
n
s
tan
t
lear
n
in
g
r
ate
o
f
0
.
0
0
1
.
T
h
e
tr
ain
in
g
last
ed
f
o
r
1
0
0
r
o
u
n
d
s
,
g
iv
in
g
th
e
m
o
d
el
tim
e
to
lear
n
f
r
o
m
th
e
d
ata
an
d
im
p
r
o
v
e
its
p
er
f
o
r
m
a
n
ce
b
y
u
p
d
atin
g
th
e
weig
h
ts
.
Sev
er
al
test
s
wer
e
d
o
n
e
t
o
f
i
n
d
t
h
e
b
est
s
ettin
g
s
f
o
r
th
e
m
o
d
el.
T
h
ese
test
s
wer
e
d
esig
n
ed
to
en
h
an
ce
th
e
m
o
d
el’
s
ac
cu
r
ac
y
an
d
b
o
o
s
t
its
o
v
er
all
p
er
f
o
r
m
a
n
ce
.
Af
te
r
a
s
er
ies
o
f
tr
ain
in
g
iter
atio
n
s
,
a
class
if
icatio
n
ac
cu
r
ac
y
o
f
9
9
.
0
6
%
an
d
a
n
F1
-
s
co
r
e
o
f
9
9
.
0
5
%
wer
e
ac
h
iev
ed
,
wh
ich
r
ef
lects
th
e
h
ig
h
ef
f
ici
en
cy
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el.
T
ab
le
2
s
h
o
ws
th
e
ev
al
u
atio
n
r
esu
lts
o
f
d
if
f
er
e
n
t
p
ar
am
eter
s
u
s
ed
in
th
e
m
o
d
e
l.
T
h
e
tab
le
s
h
o
ws
th
e
ef
f
e
ct
o
f
ch
a
n
g
in
g
ea
ch
p
a
r
am
eter
in
d
iv
id
u
ally
wh
ile
k
ee
p
in
g
t
h
e
r
est o
f
th
e
p
ar
am
e
ter
’
s
co
n
s
tan
t.
T
ab
le
2
.
T
h
e
p
er
f
o
r
m
a
n
ce
r
es
u
lts
o
f
ass
ess
in
g
th
e
p
r
o
p
o
s
ed
tech
n
iq
u
e
N
o
.
o
f
f
i
l
t
e
r
s i
n
C
o
n
v
N
o
.
o
f
f
i
l
t
e
r
s
S
p
a
t
i
a
l
l
y
C
o
n
v
N
o
.
o
f
f
i
l
t
e
r
s
D
e
p
t
h
w
i
s
e
C
o
n
v
N
o
.
o
f
f
i
l
t
e
r
s
P
o
i
n
t
w
i
se
C
o
n
v
A
c
t
i
v
a
t
i
o
n
f
u
n
c
t
i
o
n
A
c
c
u
r
a
c
y
(
%)
F1
-
sc
o
r
e
(
%)
N
o
t
e
16
16
16
1
x
1
Le
a
k
y
R
e
LU
9
0
.
1
4
9
0
.
1
9
-
32
32
32
1
x
1
Le
a
k
y
R
e
LU
9
2
.
5
1
9
2
.
6
2
-
64
64
64
1
x
1
Le
a
k
y
R
e
LU
9
4
.
3
6
9
4
.
5
6
-
1
2
8
1
2
8
1
2
8
1
x
1
Le
a
k
y
R
e
LU
9
1
.
7
5
9
1
.
7
7
-
1
2
8
64
32
1
x
1
Le
a
k
y
R
e
LU
9
5
.
4
4
9
5
.
5
1
-
1
2
8
64
32
1
x
1
Le
a
k
y
R
e
LU
9
6
.
2
1
9
6
.
3
1
-
1
2
8
64
32
1
x
1
R
e
LU
9
0
.
1
4
9
0
.
4
4
-
1
2
8
64
32
1
x
1
sw
i
sh
9
3
.
5
9
9
3
.
6
5
-
1
2
8
64
32
1
x
1
Le
a
k
y
R
e
LU
9
7
.
9
0
9
8
.
0
3
W
i
t
h
b
a
t
c
h
n
o
r
m
a
l
i
z
a
t
i
o
n
1
2
8
64
32
1
x
1
Le
a
k
y
R
e
LU
9
9
.
0
6
9
9
.
0
5
B
a
t
c
h
n
o
r
ma
l
i
z
a
t
i
o
n
+
G
l
o
b
a
l
A
v
e
r
a
g
e
P
o
o
l
i
n
g
2
D
Fig
u
r
e
6
s
h
o
ws
th
e
r
elatio
n
s
h
ip
b
etwe
en
th
e
n
u
m
b
er
o
f
iter
atio
n
s
an
d
th
e
ac
cu
r
ac
y
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
d
u
r
in
g
th
e
tr
ain
in
g
p
r
o
ce
s
s
.
I
n
th
e
ea
r
ly
s
tag
es
o
f
tr
ain
in
g
,
a
s
h
ar
p
in
cr
ea
s
e
in
ac
cu
r
ac
y
ca
n
b
e
o
b
s
er
v
ed
,
r
ef
lectin
g
t
h
e
m
o
d
e
l's
ab
ilit
y
to
q
u
ick
l
y
lear
n
f
r
o
m
th
e
b
asic
p
atter
n
s
in
th
e
d
a
ta.
Af
ter
e
p
o
ch
3
0
,
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
Utiliz
a
tio
n
o
f d
ep
th
-
w
is
e
a
n
d
s
p
a
tia
lly
s
ep
a
r
a
b
le
co
n
vo
lu
tio
n
a
l n
etw
o
r
k
fu
s
io
n
…
(
F
ir
a
s
Mu
n
ea
m
B
a
c
h
a
y
)
2601
th
e
in
cr
ea
s
e
in
ac
cu
r
ac
y
b
ec
o
m
es
m
o
r
e
g
r
ad
u
al.
At
th
is
s
tag
e,
th
e
m
o
d
el
h
as
lear
n
ed
m
o
s
t
o
f
th
e
b
asic
p
atter
n
s
,
an
d
it
b
eg
in
s
to
im
p
r
o
v
e
its
ab
ilit
y
to
r
ec
o
g
n
ize
f
in
e
d
etails
in
th
e
d
ata.
As
ep
o
ch
8
5
ap
p
r
o
ac
h
es,
th
e
ac
cu
r
ac
y
r
ea
ch
es
a
r
elativ
ely
s
tab
le
lev
el,
in
d
icatin
g
th
at
th
e
m
o
d
el
h
as
r
ea
ch
e
d
a
s
tag
e
o
f
s
atu
r
atio
n
.
T
h
is
s
tab
ilit
y
in
d
icate
s
a
b
alan
ce
b
e
twee
n
lear
n
in
g
f
r
o
m
t
h
e
d
ata
a
n
d
av
o
i
d
in
g
o
v
er
f
itti
n
g
.
Fig
u
r
e
7
s
h
o
ws
h
o
w
th
e
lo
s
s
f
u
n
ctio
n
,
ca
lled
ca
teg
o
r
ical
cr
o
s
s
en
tr
o
p
y
,
ch
an
g
es
as
th
e
m
o
d
el
tr
ain
s
o
v
er
tim
e.
At
th
e
b
eg
i
n
n
in
g
,
t
h
e
lo
s
s
f
u
n
ctio
n
ju
m
p
s
ar
o
u
n
d
a
lo
t,
s
h
o
win
g
th
at
th
e
m
o
d
el
is
s
till
ad
ju
s
tin
g
it
s
s
ettin
g
s
to
u
n
d
er
s
tan
d
th
e
d
at
a
b
etter
.
As
th
e
tr
ain
in
g
co
n
ti
n
u
es,
esp
ec
ially
af
ter
2
0
ep
o
c
h
s
,
th
e
lo
s
s
f
u
n
ctio
n
s
tar
ts
to
s
ettle
d
o
wn
,
m
ea
n
in
g
th
e
m
o
d
el
is
g
ettin
g
b
etter
at
h
an
d
lin
g
th
e
d
ata
a
n
d
m
ak
in
g
f
ewe
r
m
is
tak
es.
B
y
th
e
tim
e
6
0
ep
o
ch
s
is
r
ea
ch
ed
,
th
e
lo
s
s
f
u
n
ctio
n
k
ee
p
s
d
r
o
p
p
in
g
s
tead
ily
.
Af
ter
9
0
ep
o
ch
s
,
th
e
lo
s
s
f
u
n
ctio
n
s
tay
s
b
etwe
en
0
.
0
3
a
n
d
0
.
0
5
,
wh
ich
m
ea
n
s
th
e
m
o
d
el
h
as le
ar
n
ed
well
an
d
is
m
ak
in
g
v
er
y
f
ew
er
r
o
r
s
.
Fig
u
r
e
6
.
Acc
u
r
ac
y
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
Fig
u
r
e
7
.
L
o
s
s
f
u
n
ctio
n
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
Fig
u
r
e
8
s
h
o
ws
th
e
r
ec
eiv
e
r
o
p
er
atin
g
ch
ar
ac
ter
is
tic
(
R
OC
)
cu
r
v
es,
wh
ich
ar
e
u
s
ed
t
o
e
v
alu
ate
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
m
o
d
el
i
n
class
if
y
in
g
th
e
f
o
u
r
t
y
p
es
o
f
W
B
C
s
.
T
h
e
h
o
r
izo
n
tal
ax
is
r
ep
r
esen
ts
th
e
f
alse
p
o
s
itiv
e
r
ate,
wh
ile
th
e
v
er
tical
ax
is
r
ep
r
esen
ts
th
e
tr
u
e
p
o
s
itiv
e
r
ate.
T
h
e
R
OC
cu
r
v
es
s
h
o
w
a
co
n
v
er
g
en
ce
to
war
d
s
th
e
u
p
p
er
le
f
t
co
r
n
e
r
o
f
th
e
g
r
ap
h
,
r
ef
lectin
g
t
h
e
h
ig
h
ef
f
icien
cy
o
f
th
e
m
o
d
e
l
in
d
is
tin
g
u
is
h
in
g
b
etwe
en
th
e
d
if
f
er
en
t ty
p
es o
f
W
B
C
s
.
Fig
u
r
e
8
.
R
OC
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
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
5
,
No
.
3
,
J
u
n
e
2
0
2
6
:
2
5
9
5
-
2
6
0
5
2602
Fig
u
r
e
9
s
h
o
ws
th
e
co
n
f
u
s
io
n
m
atr
ix
o
f
th
e
p
r
o
p
o
s
ed
tech
n
iq
u
e.
T
h
e
r
esu
lts
s
h
o
w
th
at
th
e
m
o
d
e
l
ac
h
iev
ed
1
0
0
%
ac
cu
r
ac
y
in
class
if
y
in
g
ly
m
p
h
o
cy
te,
wh
ile
th
e
class
if
icatio
n
ac
cu
r
ac
y
r
ea
ch
ed
9
9
.
0
6
%
f
o
r
n
eu
tr
o
p
h
ils
,
9
9
.
3
4
%
f
o
r
eo
s
in
o
p
h
ils
an
d
9
7
.
7
7
%
f
o
r
m
o
n
o
cy
te.
T
h
ese
r
esu
lts
r
ef
lect
th
e
o
u
ts
tan
d
in
g
p
er
f
o
r
m
an
ce
o
f
th
e
m
o
d
el
an
d
its
h
ig
h
ef
f
icien
cy
in
d
is
tin
g
u
i
s
h
in
g
b
etwe
en
d
i
f
f
er
en
t t
y
p
es
o
f
W
B
C
s
.
Fig
u
r
e
9
.
C
o
n
f
u
s
io
n
m
atr
i
x
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
3
.
4
.
Co
m
pa
riso
n a
nd
re
s
ults
a
na
ly
s
is
Desp
ite
th
e
co
m
p
lex
ch
allen
g
es
p
o
s
ed
b
y
im
ag
e
r
eso
lu
tio
n
,
o
v
e
r
lap
p
in
g
ce
lls
,
h
eter
o
g
e
n
e
ity
o
f
ce
ll
ty
p
es
an
d
s
izes,
as
well
as
d
at
a
s
ca
r
city
an
d
d
if
f
er
e
n
ce
s
in
l
ig
h
tin
g
co
n
d
itio
n
s
,
th
e
r
esu
lts
o
b
tain
e
d
u
s
in
g
t
h
e
p
r
o
p
o
s
ed
tech
n
iq
u
e
h
a
v
e
p
r
o
v
en
to
b
e
v
er
y
p
r
o
m
is
in
g
.
T
h
e
r
esu
lts
s
h
o
w
th
at
th
e
p
r
o
p
o
s
ed
m
eth
o
d
ad
o
p
ted
a
s
y
s
tem
atic
an
d
o
r
g
an
ized
ap
p
r
o
ac
h
to
b
u
ild
a
m
o
d
el
b
as
ed
o
n
d
if
f
e
r
en
t
c
o
n
v
o
lu
tio
n
al
lay
er
s
.
T
h
e
m
o
d
el
b
u
ild
in
g
p
r
o
ce
s
s
was
p
r
ec
ed
ed
b
y
im
ag
e
q
u
ality
im
p
r
o
v
em
en
t
an
d
s
eg
m
en
tatio
n
s
tep
s
,
wh
ich
co
n
tr
ib
u
ted
to
ac
h
iev
in
g
h
ig
h
ly
r
eliab
le
p
er
f
o
r
m
an
ce
.
I
n
co
n
tr
ast
to
p
r
e
v
io
u
s
s
tu
d
ies
[
1
6
]
,
[
1
7
]
,
[
2
1
]
,
[
2
5
]
,
th
at
d
id
n
o
t
r
el
y
o
n
s
eg
m
en
tatio
n
o
f
b
lo
o
d
s
m
e
ar
im
ag
es,
wh
ich
o
f
f
er
e
d
th
e
ad
v
an
tag
e
o
f
r
e
d
u
cin
g
co
m
p
u
t
atio
n
al
co
m
p
lex
ity
an
d
ac
ce
ler
atin
g
th
e
tr
ain
in
g
p
r
o
ce
s
s
wh
ile
im
p
r
o
v
in
g
r
elia
b
ilit
y
.
T
h
e
p
r
o
p
o
s
ed
m
o
d
el
is
d
is
tin
g
u
is
h
ed
f
r
o
m
tr
ad
itio
n
al
C
NN
m
o
d
els
u
s
ed
in
m
o
s
t
p
r
ev
io
u
s
s
tu
d
ies
[
2
]
,
[
1
9
]
,
[
2
0
]
,
[
2
2
]
–
[
2
4
]
,
as
it
was
s
p
ec
if
ically
d
esig
n
ed
to
a
d
d
r
ess
th
e
u
n
iq
u
e
ch
allen
g
es
ass
o
ciate
d
with
W
B
C
clas
s
if
icatio
n
.
T
ab
le
3
p
r
o
v
id
es
a
co
m
p
ar
ativ
e
an
aly
s
is
o
f
th
e
d
if
f
er
en
t
m
eth
o
d
o
lo
g
ies
an
d
th
eir
r
esu
lts
,
with
th
e
p
r
o
p
o
s
ed
m
o
d
el.
W
h
en
co
m
p
ar
in
g
th
e
r
esu
lts
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
with
r
ec
en
t
r
e
s
ea
r
ch
b
ased
o
n
th
e
s
am
e
d
at
aset,
it
is
clea
r
th
a
t
th
e
ac
h
iev
ed
r
esu
lts
ar
e
v
er
y
en
co
u
r
a
g
in
g
a
n
d
h
ig
h
lig
h
t
t
h
e
ef
f
icien
cy
o
f
th
e
ap
p
r
o
ac
h
in
im
p
r
o
v
in
g
th
e
ac
cu
r
ac
y
an
d
r
eliab
ilit
y
o
f
WBC
cla
s
s
if
icatio
n
.
T
ab
le
3
.
C
o
m
p
a
r
is
o
n
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
p
e
r
f
o
r
m
an
ce
an
d
th
e
r
elate
d
wo
r
k
s
R
e
f
e
r
e
n
c
e
Te
c
h
n
i
q
u
e
S
e
g
m
e
n
t
a
t
i
o
n
A
c
c
u
r
a
c
y
(
%
)
[
1
]
A
l
e
x
N
e
t
,
G
o
o
g
Le
N
e
t
,
a
n
d
R
e
sN
e
t
-
50
No
9
7
.
9
5
[
1
6
]
C
N
N
No
9
9
.
5
[
1
9
]
D
e
n
seN
e
t
-
1
2
1
No
9
8
.
8
4
[
2
0
]
M
o
b
i
l
e
N
e
t
Y
e
s
9
8
.
4
[
1
7
]
C
N
N
No
98
[
2
1
]
C
N
N
No
9
8
.
3
3
[
2
2
]
C
N
N
a
n
d
R
N
N
No
9
5
.
8
9
[
2
3
]
Op
-
Y
O
LO
v
8
Y
e
s
9
9
.
2
[
2
4
]
D
A
F
F
N
e
t
Y
e
s
9
9
.
7
1
[
2
5
]
C
N
N
No
9
7
.
5
9
Th
e
p
r
o
p
o
s
e
d
t
e
c
h
n
i
q
u
e
M
i
x
C
N
N
Y
e
s
9
9
.
0
6
4.
CO
NCLU
SI
O
N
T
h
is
p
ap
er
d
ev
elo
p
e
d
an
ad
v
an
ce
d
WBC
cla
s
s
if
icatio
n
tech
n
iq
u
e
b
ased
o
n
d
ee
p
lear
n
in
g
ap
p
r
o
ac
h
es,
f
o
c
u
s
in
g
o
n
th
e
u
s
e
o
f
d
if
f
e
r
en
t
t
y
p
es
o
f
c
o
n
v
o
lu
ti
o
n
al
lay
e
r
s
.
As
a
p
r
elu
d
e,
th
e
C
L
AHE
alg
o
r
ith
m
was
ap
p
lied
to
en
h
an
ce
im
ag
e
co
n
tr
ast,
alo
n
g
with
co
lo
r
is
o
latio
n
tec
h
n
o
lo
g
y
to
s
eg
m
en
t
im
ag
es.
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
Utiliz
a
tio
n
o
f d
ep
th
-
w
is
e
a
n
d
s
p
a
tia
lly
s
ep
a
r
a
b
le
co
n
vo
lu
tio
n
a
l n
etw
o
r
k
fu
s
io
n
…
(
F
ir
a
s
Mu
n
ea
m
B
a
c
h
a
y
)
2603
T
h
is
ap
p
r
o
ac
h
aim
s
to
ad
d
r
ess
th
e
ch
allen
g
es
r
esu
ltin
g
f
r
o
m
ce
ll
o
v
er
lap
an
d
d
if
f
er
e
n
t
lig
h
tin
g
co
n
d
itio
n
s
.
Usi
n
g
th
e
p
r
o
p
o
s
ed
tech
n
iq
u
e,
d
ee
p
f
ea
tu
r
es
wer
e
ex
tr
ac
ted
f
r
o
m
im
ag
es,
r
esu
ltin
g
in
h
ig
h
class
if
icatio
n
p
er
f
o
r
m
an
ce
with
h
ig
h
r
eliab
i
lity
,
ev
en
wh
en
d
ea
lin
g
with
a
lim
ited
d
ataset
s
ize,
wh
ich
i
s
a
m
ajo
r
ch
allen
g
e
in
d
ee
p
lea
r
n
in
g
ap
p
licatio
n
s
.
An
o
p
en
-
s
o
u
r
ce
d
ataset
f
r
o
m
Kag
g
le
was
u
s
ed
f
o
r
ex
p
er
im
en
ts
.
Sev
er
al
ex
p
er
im
en
ts
wer
e
d
esig
n
ed
an
d
im
p
lem
e
n
ted
to
u
n
d
er
s
tan
d
h
o
w
th
e
p
ar
am
eter
s
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
lay
e
r
s
ev
o
lv
e
o
v
er
tim
e.
T
h
e
r
esu
lts
s
h
o
wed
th
at
th
e
m
o
d
el
is
ab
le
to
ad
ap
t
to
a
v
a
r
iety
o
f
W
B
C
s
,
r
esu
ltin
g
in
h
ig
h
class
if
icatio
n
ac
cu
r
ac
y
.
W
h
en
co
m
p
ar
e
d
with
o
th
er
ap
p
r
o
ac
h
es,
th
e
p
r
o
p
o
s
ed
tech
n
iq
u
e
d
em
o
n
s
tr
ated
s
u
p
er
io
r
p
er
f
o
r
m
a
n
ce
.
T
h
e
b
e
s
t
r
esu
lts
ac
h
iev
ed
wer
e
a
clas
s
if
icatio
n
ac
cu
r
ac
y
o
f
9
9
.
0
6
%
an
d
an
F1
-
s
co
r
e
o
f
9
9
.
0
5
%,
in
d
icatin
g
th
e
ef
f
ici
en
cy
an
d
ef
f
ec
tiv
en
ess
o
f
th
e
m
o
d
el.
I
n
th
e
f
u
tu
r
e,
a
d
o
p
tin
g
d
ee
p
lear
n
in
g
m
eth
o
d
o
l
o
g
ies
in
all
s
tag
e
s
o
f
th
e
W
B
C
s
cla
s
s
if
icatio
n
p
r
o
ce
s
s
i
s
p
r
o
p
o
s
ed
,
f
r
o
m
co
n
tr
ast
en
h
an
ce
m
en
t
an
d
im
ag
e
s
eg
m
en
tatio
n
to
th
e
f
in
al
class
if
icatio
n
.
T
h
e
m
o
s
t
im
p
o
r
tan
t
co
n
t
r
ib
u
tio
n
s
o
f
th
is
wo
r
k
ar
e
an
in
teg
r
ated
s
eg
m
en
tatio
n
-
class
if
icatio
n
p
ip
elin
e,
d
e
v
elo
p
m
e
n
t
o
f
a
s
p
ec
ialized
d
ee
p
lea
r
n
in
g
m
o
d
el,
an
d
ap
p
ly
in
g
C
L
AHE
f
o
r
tr
ain
in
g
-
en
h
an
ce
d
co
n
tr
ast.
No
n
eth
el
ess
,
lim
itatio
n
s
s
u
ch
as
ev
alu
atin
g
o
n
ly
a
s
in
g
le
d
ataset
an
d
ab
s
en
ce
o
f
clin
ica
l
v
alid
atio
n
ar
e
p
r
esen
t.
I
n
f
u
t
u
r
e
wo
r
k
,
t
h
e
m
o
d
el
is
p
lan
n
e
d
to
b
e
r
ef
in
e
d
f
o
r
d
ep
lo
y
m
e
n
t
,
in
teg
r
ate
clin
ical
test
s
,
an
d
v
alid
ate
th
e
m
o
d
el
clin
ically
,
as
well
a
s
o
p
tim
ize
test
in
g
o
n
o
th
er
d
atasets
an
d
in
co
r
p
o
r
ate
atten
t
io
n
m
ec
h
a
n
is
m
s
f
o
cu
s
ed
o
n
f
e
atu
r
e
ex
tr
ac
tio
n
.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
Au
th
o
r
s
s
tate
n
o
f
u
n
d
in
g
in
v
o
lv
ed
.
AUTHO
R
CO
NT
RI
B
UT
I
O
NS ST
A
T
E
M
E
N
T
T
h
is
jo
u
r
n
al
u
s
es
th
e
C
o
n
tr
ib
u
to
r
R
o
les
T
ax
o
n
o
m
y
(
C
R
ed
iT)
to
r
ec
o
g
n
ize
in
d
iv
id
u
al
au
th
o
r
co
n
tr
ib
u
tio
n
s
,
r
ed
u
ce
au
th
o
r
s
h
ip
d
is
p
u
tes,
an
d
f
ac
ilit
ate
co
llab
o
r
atio
n
.
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
Fira
s
Mu
n
ea
m
B
ac
h
ay
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Ali A
b
b
as Alz
ah
eir
ee
✓
✓
✓
✓
✓
✓
Hasan
ain
Ali H
u
s
s
ein
✓
✓
✓
✓
✓
✓
✓
✓
✓
Ah
m
ed
No
o
r
u
ld
ee
n
Als
af
i
✓
✓
✓
✓
✓
✓
Mo
h
am
m
ed
Hasan
Ab
d
u
lam
ee
r
✓
✓
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
n
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
Au
th
o
r
s
s
tate
n
o
co
n
f
lict o
f
in
t
er
est.
DATA AV
AI
L
AB
I
L
I
T
Y
T
h
e
d
ata
th
at
s
u
p
p
o
r
t
th
e
f
in
d
in
g
s
o
f
th
is
s
tu
d
y
ar
e
o
p
en
ly
av
ailab
le
in
Kag
g
le
at
h
ttp
s
://www.
k
ag
g
le.
co
m
/d
atasets
/p
au
ltimo
th
y
m
o
o
n
ey
/b
lo
o
d
-
ce
lls
,
r
ef
er
en
ce
[
2
8
]
.
RE
F
E
R
E
NC
E
S
[
1
]
Y
.
Y
.
B
a
y
d
i
l
l
i
a
n
d
Ü
.
A
t
i
l
a
,
“
C
l
a
ss
i
f
i
c
a
t
i
o
n
o
f
w
h
i
t
e
b
l
o
o
d
c
e
l
l
s
u
s
i
n
g
c
a
p
su
l
e
n
e
t
w
o
r
k
s,
”
C
o
m
p
u
t
e
r
i
ze
d
M
e
d
i
c
a
l
I
m
a
g
i
n
g
a
n
d
G
ra
p
h
i
c
s
,
v
o
l
.
8
0
,
M
a
r
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
m
p
me
d
i
ma
g
.
2
0
2
0
.
1
0
1
6
9
9
.
[
2
]
M
.
To
ğ
a
ç
a
r
,
B
.
Er
g
e
n
,
a
n
d
Z.
C
ö
mert
,
“
C
l
a
ssi
f
i
c
a
t
i
o
n
o
f
w
h
i
t
e
b
l
o
o
d
c
e
l
l
s u
si
n
g
d
e
e
p
f
e
a
t
u
r
e
s
o
b
t
a
i
n
e
d
f
r
o
m
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
mo
d
e
l
s
b
a
s
e
d
o
n
t
h
e
c
o
m
b
i
n
a
t
i
o
n
o
f
f
e
a
t
u
r
e
s
e
l
e
c
t
i
o
n
m
e
t
h
o
d
s,”
Ap
p
l
i
e
d
S
o
f
t
C
o
m
p
u
t
i
n
g
,
v
o
l
.
9
7
,
D
e
c
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
a
so
c
.
2
0
2
0
.
1
0
6
8
1
0
.
[
3
]
O
.
K
a
t
a
r
a
n
d
İ
.
F
.
K
ı
l
ı
n
ç
e
r
,
“
A
u
t
o
m
a
t
i
c
c
l
a
ss
i
f
i
c
a
t
i
o
n
o
f
w
h
i
t
e
b
l
o
o
d
c
e
l
l
s
u
si
n
g
p
r
e
-
t
r
a
i
n
e
d
d
e
e
p
m
o
d
e
l
s
,
”
S
a
k
a
r
y
a
U
n
i
v
e
rs
i
t
y
J
o
u
rn
a
l
o
f
C
o
m
p
u
t
e
r
a
n
d
I
n
f
o
rm
a
t
i
o
n
S
c
i
e
n
c
e
s
,
v
o
l
.
5
,
n
o
.
3
,
p
p
.
4
6
2
–
4
7
6
,
D
e
c
.
2
0
2
2
,
d
o
i
:
1
0
.
3
5
3
7
7
/
sa
u
c
i
s.
0
5
.
0
3
.
1
1
9
6
9
3
4
.
[
4
]
X
.
Y
a
o
,
K
.
S
u
n
,
X
.
B
u
,
C
.
Zh
a
o
,
a
n
d
Y
.
J
i
n
,
“
C
l
a
ssi
f
i
c
a
t
i
o
n
o
f
w
h
i
t
e
b
l
o
o
d
c
e
l
l
s
u
s
i
n
g
w
e
i
g
h
t
e
d
o
p
t
i
m
i
z
e
d
d
e
f
o
r
ma
b
l
e
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s,”
Ar
t
i
f
i
c
i
a
l
C
e
l
l
s,
N
a
n
o
m
e
d
i
c
i
n
e
,
a
n
d
B
i
o
t
e
c
h
n
o
l
o
g
y
,
v
o
l
.
4
9
,
n
o
.
1
,
p
p
.
1
4
7
–
1
5
5
,
Ja
n
.
2
0
2
1
,
d
o
i
:
1
0
.
1
0
8
0
/
2
1
6
9
1
4
0
1
.
2
0
2
1
.
1
8
7
9
8
2
3
.
[
5
]
H
.
Ü
z
e
n
a
n
d
H
.
F
ı
r
a
t
,
“
A
h
y
b
r
i
d
a
p
p
r
o
a
c
h
b
a
se
d
o
n
mu
l
t
i
p
a
t
h
sw
i
n
t
r
a
n
sf
o
r
mer
a
n
d
C
o
n
v
M
i
x
e
r
f
o
r
w
h
i
t
e
b
l
o
o
d
c
e
l
l
s
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
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
s
t
e
m
s
,
v
o
l
.
1
2
,
n
o
.
1
,
A
p
r
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
3
7
5
5
-
024
-
0
0
2
9
1
-
w.
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
5
,
No
.
3
,
J
u
n
e
2
0
2
6
:
2
5
9
5
-
2
6
0
5
2604
[
6
]
D
.
R
y
u
e
t
a
l
.
,
“
L
a
b
e
l
-
f
r
e
e
w
h
i
t
e
b
l
o
o
d
c
e
l
l
c
l
a
ssi
f
i
c
a
t
i
o
n
u
s
i
n
g
r
e
f
r
a
c
t
i
v
e
i
n
d
e
x
t
o
mo
g
r
a
p
h
y
a
n
d
d
e
e
p
l
e
a
r
n
i
n
g
,
”
BM
E
Fr
o
n
t
i
e
rs
,
v
o
l
.
2
0
2
1
,
Ja
n
.
2
0
2
1
,
d
o
i
:
1
0
.
3
4
1
3
3
/
2
0
2
1
/
9
8
9
3
8
0
4
.
[
7
]
K
.
A
l
mez
h
g
h
w
i
a
n
d
S
.
S
e
r
t
e
,
“
I
mp
r
o
v
e
d
c
l
a
ssi
f
i
c
a
t
i
o
n
o
f
w
h
i
t
e
b
l
o
o
d
c
e
l
l
s
w
i
t
h
t
h
e
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
a
n
d
d
e
e
p
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
,
”
C
o
m
p
u
t
a
t
i
o
n
a
l
I
n
t
e
l
l
i
g
e
n
c
e
a
n
d
N
e
u
ro
s
c
i
e
n
c
e
,
v
o
l
.
2
0
2
0
,
p
p
.
1
–
1
2
,
Ju
l
.
2
0
2
0
,
d
o
i
:
1
0
.
1
1
5
5
/
2
0
2
0
/
6
4
9
0
4
7
9
.
[
8
]
E.
H
.
M
o
h
a
me
d
,
W
.
H
.
E
l
-
B
e
h
a
i
d
y
,
G
.
K
h
o
r
i
b
a
,
a
n
d
J
.
Li
,
“
I
mp
r
o
v
e
d
w
h
i
t
e
b
l
o
o
d
c
e
l
l
s
c
l
a
ss
i
f
i
c
a
t
i
o
n
b
a
s
e
d
o
n
p
r
e
-
t
r
a
i
n
e
d
d
e
e
p
l
e
a
r
n
i
n
g
mo
d
e
l
s,
”
J
o
u
r
n
a
l
o
f
c
o
m
m
u
n
i
c
a
t
i
o
n
s
s
o
f
t
w
a
r
e
a
n
d
s
y
st
e
m
s
,
v
o
l
.
1
6
,
n
o
.
1
,
p
p
.
3
7
–
4
5
,
M
a
r
.
2
0
2
0
,
d
o
i
:
1
0
.
2
4
1
3
8
/
j
c
o
mss
.
v
1
6
i
1
.
8
1
8
.
[
9
]
R
.
A
.
B
a
g
i
d
o
,
M
.
A
l
z
a
h
r
a
n
i
,
a
n
d
M
.
A
r
i
f
,
“
W
h
i
t
e
b
l
o
o
d
c
e
l
l
t
y
p
e
s
c
l
a
ss
i
f
i
c
a
t
i
o
n
u
s
i
n
g
d
e
e
p
l
e
a
r
n
i
n
g
mo
d
e
l
s
,
”
I
J
C
S
N
S
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
N
e
t
w
o
rk
S
e
c
u
ri
t
y
,
v
o
l
.
2
1
,
n
o
.
9
,
2
0
2
1
,
d
o
i
:
1
0
.
2
2
9
3
7
/
I
JC
S
N
S
.
2
0
2
1
.
2
1
.
9
.
3
0
.
[
1
0
]
A
.
K
h
o
u
a
n
i
,
M
.
E.
H
.
D
a
h
o
,
S
.
A
.
M
a
h
m
o
u
d
i
,
M
.
A
.
C
h
i
k
h
,
a
n
d
B
.
B
e
n
z
i
n
e
b
,
“
A
u
t
o
ma
t
e
d
r
e
c
o
g
n
i
t
i
o
n
o
f
w
h
i
t
e
b
l
o
o
d
c
e
l
l
s
u
s
i
n
g
d
e
e
p
l
e
a
r
n
i
n
g
,
”
B
i
o
m
e
d
i
c
a
l
E
n
g
i
n
e
e
r
i
n
g
L
e
t
t
e
rs
,
v
o
l
.
1
0
,
n
o
.
3
,
p
p
.
3
5
9
–
3
6
7
,
A
u
g
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
3
5
3
4
-
0
2
0
-
0
0
1
6
8
-
3.
[
1
1
]
S
.
K
h
a
n
,
M
.
S
a
j
j
a
d
,
T
.
H
u
ss
a
i
n
,
A
.
U
l
l
a
h
,
a
n
d
A
.
S
.
I
mr
a
n
,
“
A
r
e
v
i
e
w
o
n
t
r
a
d
i
t
i
o
n
a
l
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
n
d
d
e
e
p
l
e
a
r
n
i
n
g
mo
d
e
l
s
f
o
r
W
B
C
s
c
l
a
ss
i
f
i
c
a
t
i
o
n
i
n
b
l
o
o
d
sm
e
a
r
i
ma
g
e
s
,
”
I
E
EE
Ac
c
e
ss
,
v
o
l
.
9
,
p
p
.
1
0
6
5
7
–
1
0
6
7
3
,
2
0
2
1
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
0
.
3
0
4
8
1
7
2
.
[
1
2
]
M
.
M
.
A
h
s
a
n
,
S
.
A
.
Lu
n
a
,
a
n
d
Z
.
S
i
d
d
i
q
u
e
,
“
M
a
c
h
i
n
e
-
l
e
a
r
n
i
n
g
-
b
a
s
e
d
d
i
s
e
a
se
d
i
a
g
n
o
si
s
:
a
c
o
m
p
r
e
h
e
n
si
v
e
r
e
v
i
e
w
,
”
H
e
a
l
t
h
c
a
r
e
,
v
o
l
.
1
0
,
n
o
.
3
,
M
a
r
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
h
e
a
l
t
h
c
a
r
e
1
0
0
3
0
5
4
1
.
[
1
3
]
R
.
A
h
ma
d
,
M
.
A
w
a
i
s,
N
.
K
a
u
sar,
a
n
d
T.
A
k
r
a
m
,
“
W
h
i
t
e
b
l
o
o
d
c
e
l
l
s
c
l
a
ss
i
f
i
c
a
t
i
o
n
u
s
i
n
g
e
n
t
r
o
p
y
-
c
o
n
t
r
o
l
l
e
d
d
e
e
p
f
e
a
t
u
r
e
s
o
p
t
i
m
i
z
a
t
i
o
n
,
”
D
i
a
g
n
o
s
t
i
c
s
,
v
o
l
.
1
3
,
n
o
.
3
,
Ja
n
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
d
i
a
g
n
o
st
i
c
s1
3
0
3
0
3
5
2
.
[
1
4
]
H
.
K
u
t
l
u
,
E
.
A
v
c
i
,
a
n
d
F
.
Ö
z
y
u
r
t
,
“
W
h
i
t
e
b
l
o
o
d
c
e
l
l
s
d
e
t
e
c
t
i
o
n
a
n
d
c
l
a
ssi
f
i
c
a
t
i
o
n
b
a
s
e
d
o
n
r
e
g
i
o
n
a
l
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s,”
M
e
d
i
c
a
l
H
y
p
o
t
h
e
ses
,
v
o
l
.
1
3
5
,
F
e
b
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
me
h
y
.
2
0
1
9
.
1
0
9
4
7
2
.
[
1
5
]
J.
B
a
s
n
e
t
,
A
.
A
l
sa
d
o
o
n
,
P
.
W
.
C
.
P
r
a
sad
,
S
.
A
l
A
l
o
u
ssi
,
a
n
d
O
.
H
.
A
l
sa
d
o
o
n
,
“
A
n
o
v
e
l
so
l
u
t
i
o
n
o
f
u
s
i
n
g
d
e
e
p
l
e
a
r
n
i
n
g
f
o
r
w
h
i
t
e
b
l
o
o
d
c
e
l
l
s
c
l
a
ss
i
f
i
c
a
t
i
o
n
:
e
n
h
a
n
c
e
d
l
o
ss
f
u
n
c
t
i
o
n
w
i
t
h
r
e
g
u
l
a
r
i
z
a
t
i
o
n
a
n
d
w
e
i
g
h
t
e
d
l
o
ss
(
ELFRW
L)
,
”
N
e
u
r
a
l
P
ro
c
e
ssi
n
g
L
e
t
t
e
rs
,
v
o
l
.
5
2
,
n
o
.
2
,
p
p
.
1
5
1
7
–
1
5
5
3
,
O
c
t
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
1
0
6
3
-
020
-
1
0
3
2
1
-
9.
[
1
6
]
S
.
N
a
h
z
a
t
,
F
.
B
o
z
k
u
r
t
,
a
n
d
M
.
Y
a
ğ
a
n
o
ğ
l
u
,
“
W
h
i
t
e
b
l
o
o
d
c
e
l
l
c
l
a
s
si
f
i
c
a
t
i
o
n
u
s
i
n
g
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
,
”
J
o
u
rn
a
l
o
f
S
c
i
e
n
t
i
f
i
c
T
e
c
h
n
o
l
o
g
y
a
n
d
E
n
g
i
n
e
e
r
i
n
g
R
e
se
a
rc
h
,
v
o
l
.
3
,
n
o
.
1
,
p
p
.
3
2
–
4
1
,
A
p
r
.
2
0
2
2
,
d
o
i
:
1
0
.
5
3
5
2
5
/
j
st
e
r
.
1
0
1
8
2
1
3
.
[
1
7
]
A
.
K
a
n
a
v
o
s
,
O
.
P
a
p
a
d
i
m
i
t
r
i
o
u
,
K
.
A
l
-
H
u
ssae
n
i
,
M
.
M
a
r
a
g
o
u
d
a
k
i
s,
a
n
d
I
.
K
a
r
a
mi
t
so
s
,
“
A
d
v
a
n
c
e
d
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s
f
o
r
p
r
e
c
i
se
w
h
i
t
e
b
l
o
o
d
c
e
l
l
s
u
b
t
y
p
e
c
l
a
ss
i
f
i
c
a
t
i
o
n
i
n
m
e
d
i
c
a
l
d
i
a
g
n
o
s
t
i
c
s,
”
El
e
c
t
ro
n
i
c
s
,
v
o
l
.
1
3
,
n
o
.
1
4
,
Ju
l
.
2
0
2
4
,
d
o
i
:
1
0
.
3
3
9
0
/
e
l
e
c
t
r
o
n
i
c
s1
3
1
4
2
8
1
8
.
[
1
8
]
C
.
Z
o
n
y
f
a
r
a
n
d
J.
K
i
m,
“
C
A
B
‐
N
e
t
:
c
o
n
v
o
l
u
t
i
o
n
a
l
a
t
t
e
n
t
i
o
n
B
LSTM
n
e
t
w
o
r
k
f
o
r
e
n
h
a
n
c
e
d
l
e
u
k
o
c
y
t
e
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
I
m
a
g
i
n
g
S
y
st
e
m
s a
n
d
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
3
4
,
n
o
.
2
,
M
a
r
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
2
/
i
ma.
2
3
0
6
5
.
[
1
9
]
S
.
S
h
a
r
m
a
e
t
a
l
.
,
“
D
e
e
p
l
e
a
r
n
i
n
g
m
o
d
e
l
f
o
r
t
h
e
a
u
t
o
mat
i
c
c
l
a
ssi
f
i
c
a
t
i
o
n
o
f
w
h
i
t
e
b
l
o
o
d
c
e
l
l
s
,
”
C
o
m
p
u
t
a
t
i
o
n
a
l
I
n
t
e
l
l
i
g
e
n
c
e
a
n
d
N
e
u
ro
s
c
i
e
n
c
e
,
v
o
l
.
2
0
2
2
,
p
p
.
1
–
1
3
,
Ja
n
.
2
0
2
2
,
d
o
i
:
1
0
.
1
1
5
5
/
2
0
2
2
/
7
3
8
4
1
3
1
.
[
2
0
]
C
.
C
h
e
u
q
u
e
,
M
.
Q
u
e
r
a
l
e
s
,
R
.
Le
ó
n
,
R
.
S
a
l
a
s,
a
n
d
R
.
T
o
r
r
e
s,
“
A
n
e
f
f
i
c
i
e
n
t
mu
l
t
i
-
l
e
v
e
l
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
a
p
p
r
o
a
c
h
f
o
r
w
h
i
t
e
b
l
o
o
d
c
e
l
l
s
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
D
i
a
g
n
o
s
t
i
c
s
,
v
o
l
.
1
2
,
n
o
.
2
,
J
a
n
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
d
i
a
g
n
o
s
t
i
c
s1
2
0
2
0
2
4
8
.
[
2
1
]
I
.
S
.
R
a
j
p
u
t
,
S
.
Ty
a
g
i
,
A
.
G
u
p
t
a
,
a
n
d
V
.
Ja
i
n
,
“
A
d
v
a
n
c
e
s i
n
me
d
i
c
a
l
i
ma
g
i
n
g
:
u
s
i
n
g
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s
f
o
r
w
h
i
t
e
b
l
o
o
d
c
e
l
l
i
d
e
n
t
i
f
i
c
a
t
i
o
n
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
I
m
a
g
e
,
G
r
a
p
h
i
c
s
a
n
d
S
i
g
n
a
l
Pro
c
e
ss
i
n
g
,
v
o
l
.
1
6
,
n
o
.
1
,
p
p
.
1
0
8
–
1
2
5
,
F
e
b
.
2
0
2
4
,
d
o
i
:
1
0
.
5
8
1
5
/
i
j
i
g
sp
.
2
0
2
4
.
0
1
.
0
8
.
[
2
2
]
A
.
M
.
P
a
t
i
l
,
M
.
D
.
P
a
t
i
l
,
a
n
d
G
.
K
.
B
i
r
a
j
d
a
r
,
“
W
h
i
t
e
b
l
o
o
d
c
e
l
l
s
i
m
a
g
e
c
l
a
ss
i
f
i
c
a
t
i
o
n
u
s
i
n
g
d
e
e
p
l
e
a
r
n
i
n
g
w
i
t
h
c
a
n
o
n
i
c
a
l
c
o
r
r
e
l
a
t
i
o
n
a
n
a
l
y
s
i
s,
”
I
n
n
o
v
a
t
i
o
n
a
n
d
Re
s
e
a
rc
h
i
n
Bi
o
M
e
d
i
c
a
l
e
n
g
i
n
e
e
ri
n
g
,
v
o
l
.
4
2
,
n
o
.
5
,
p
p
.
3
7
8
–
3
8
9
,
O
c
t
.
2
0
2
1
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
i
r
b
m
.
2
0
2
0
.
0
8
.
0
0
5
.
[
2
3
]
A
.
A
b
o
z
e
i
d
e
t
a
l
.
,
“
W
h
i
t
e
b
l
o
o
d
c
e
l
l
s
d
e
t
e
c
t
i
o
n
u
si
n
g
d
e
e
p
l
e
a
r
n
i
n
g
i
n
h
e
a
l
t
h
c
a
r
e
a
p
p
l
i
c
a
t
i
o
n
s
,
”
A
l
e
x
a
n
d
ri
a
E
n
g
i
n
e
e
ri
n
g
J
o
u
r
n
a
l
,
v
o
l
.
1
2
4
,
p
p
.
1
3
5
–
1
4
6
,
J
u
n
.
2
0
2
5
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
a
e
j
.
2
0
2
5
.
0
3
.
0
9
8
.
[
2
4
]
Y
.
C
h
e
n
,
Z.
C
h
e
n
,
Y
.
A
n
,
C
.
L
u
,
a
n
d
X
.
Q
i
a
o
,
“
D
A
F
F
N
e
t
:
a
d
u
a
l
a
t
t
e
n
t
i
o
n
f
e
a
t
u
r
e
f
u
s
i
o
n
n
e
t
w
o
r
k
f
o
r
c
l
a
ssi
f
i
c
a
t
i
o
n
o
f
w
h
i
t
e
b
l
o
o
d
c
e
l
l
s,
”
B
i
o
m
e
d
i
c
a
l
S
i
g
n
a
l
Pr
o
c
e
ss
i
n
g
a
n
d
C
o
n
t
ro
l
,
v
o
l
.
1
0
6
,
A
u
g
.
2
0
2
5
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
b
s
p
c
.
2
0
2
5
.
1
0
7
6
9
9
.
[
2
5
]
A
.
M
.
F
a
y
y
a
z
,
S
.
J.
A
b
d
u
l
k
a
d
i
r
,
S
.
M
.
A
l
-
S
e
l
w
i
,
E.
H
.
S
u
mi
e
a
,
S
.
I
q
b
a
l
,
a
n
d
S
.
U
l
H
a
ssa
n
,
“
A
n
o
v
e
l
C
N
N
m
o
d
e
l
w
i
t
h
e
n
t
r
o
p
y
-
c
o
d
e
d
g
e
n
e
t
i
c
a
l
g
o
r
i
t
h
m
f
o
r
b
l
o
o
d
c
e
l
l
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
J
o
u
r
n
a
l
o
f
Ad
v
a
n
c
e
d
Re
s
e
a
r
c
h
i
n
A
p
p
l
i
e
d
S
c
i
e
n
c
e
s
a
n
d
En
g
i
n
e
e
ri
n
g
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
6
3
,
n
o
.
3
,
M
a
r
.
2
0
2
5
,
d
o
i
:
1
0
.
3
7
9
3
4
/
a
r
a
s
e
t
.
6
3
.
3
.
1
3
4
1
4
8
.
[
2
6
]
P
.
H
ä
r
t
i
n
g
e
r
a
n
d
C
.
S
t
e
g
e
r
,
“
A
d
a
p
t
i
v
e
h
i
s
t
o
g
r
a
m
e
q
u
a
l
i
z
a
t
i
o
n
i
n
c
o
n
s
t
a
n
t
t
i
me,
”
J
o
u
r
n
a
l
o
f
R
e
a
l
-
T
i
m
e
I
m
a
g
e
Pr
o
c
e
ssi
n
g
,
v
o
l
.
2
1
,
n
o
.
3
,
J
u
n
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
1
5
5
4
-
0
2
4
-
0
1
4
6
5
-
1.
[
2
7
]
G
.
F
.
C
.
C
a
mp
o
s,
S
.
M
.
M
a
st
e
l
i
n
i
,
G
.
J.
A
g
u
i
a
r
,
R
.
G
.
M
a
n
t
o
v
a
n
i
,
L.
F
.
d
e
M
e
l
o
,
a
n
d
S
.
B
a
r
b
o
n
,
“
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
h
y
p
e
r
p
a
r
a
me
t
e
r
s
e
l
e
c
t
i
o
n
f
o
r
c
o
n
t
r
a
st
l
i
m
i
t
e
d
a
d
a
p
t
i
v
e
h
i
s
t
o
g
r
a
m
e
q
u
a
l
i
z
a
t
i
o
n
,
”
EU
RA
S
I
P
J
o
u
r
n
a
l
o
n
I
m
a
g
e
a
n
d
V
i
d
e
o
Pro
c
e
ssi
n
g
,
v
o
l
.
2
0
1
9
,
n
o
.
1
,
D
e
c
.
2
0
1
9
,
d
o
i
:
1
0
.
1
1
8
6
/
s1
3
6
4
0
-
0
1
9
-
0
4
4
5
-
4.
[
2
8
]
P
.
M
o
o
n
e
y
,
“
B
l
o
o
d
c
e
l
l
i
m
a
g
e
s,”
K
a
g
g
l
e
.
c
o
m
,
2
0
1
8
.
A
c
c
e
sse
d
:
D
e
c
.
0
1
,
2
0
2
4
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
w
w
w
.
k
a
g
g
l
e
.
c
o
m
/
d
a
t
a
se
t
s/
p
a
u
l
t
i
m
o
t
h
y
mo
o
n
e
y
/
b
l
o
o
d
-
c
e
l
l
s
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
Fi
r
a
s
Mu
n
e
a
m
B
a
c
h
a
y
re
c
e
iv
e
d
h
is
B.
S
c
.
a
n
d
M
.
S
c
.
(Co
m
p
u
ter
S
c
ien
c
e
)
fro
m
Ku
fa
Un
iv
e
rsity
,
Ira
q
i
n
2
0
1
0
a
n
d
2
0
2
2
,
re
sp
e
c
ti
v
e
ly
.
He
is
c
u
rre
n
tl
y
a
n
a
ss
istan
t
l
e
c
tu
re
in
De
p
a
rtme
n
t
o
f
t
h
e
S
c
ien
ti
fic
Af
fa
ir,
Un
i
v
e
rsity
o
f
K
u
fa
,
Na
jaf,
Ira
q
.
His
re
se
a
rc
h
in
c
l
u
d
e
s
d
e
e
p
lea
rn
i
n
g
,
a
rti
f
icia
l
i
n
telli
g
e
n
c
e
,
b
i
o
m
e
tri
c
,
a
n
d
ima
g
e
p
r
o
c
e
ss
in
g
.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
firas
m
.
b
a
c
h
a
y
@u
o
k
u
fa
.
e
d
u
.
iq
.
Evaluation Warning : The document was created with Spire.PDF for Python.