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
.
1
,
Feb
r
u
ar
y
2
0
2
6
,
p
p
.
429
~
442
I
SS
N:
2
2
5
2
-
8
9
3
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijai.v
15
.i
1
.
p
p
4
2
9
-
4
4
2
429
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
Bra
in t
umo
r seg
menta
tion a
nd
cla
ss
ificatio
n using
a
rtif
icia
l
humm
ing
bird opt
imiza
tion a
lg
o
rith
m
Ra
dh
a
k
rish
na
n K
a
rt
hik
ey
a
n,
Ara
pp
a
leeswa
ra
n M
urug
a
na
nd
ha
m
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
o
n
i
c
s a
n
d
C
o
mm
u
n
i
c
a
t
i
o
n
En
g
i
n
e
e
r
i
n
g
,
R
a
j
a
r
a
j
e
sw
a
r
i
C
o
l
l
e
g
e
o
f
E
n
g
i
n
e
e
r
i
n
g
,
B
a
n
g
a
l
o
r
e
,
I
n
d
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Ap
r
1
8
,
2
0
2
4
R
ev
is
ed
No
v
1
0
,
2
0
2
5
Acc
ep
ted
Dec
1
5
,
2
0
2
5
Th
e
ti
m
e
a
n
d
m
e
d
ica
l
p
e
rso
n
n
e
l
e
x
p
e
rien
c
e
a
re
th
e
o
n
ly
fa
c
to
rs
th
a
t
d
e
term
in
e
wh
e
th
e
r
b
ra
in
t
u
m
o
rs
c
a
n
b
e
m
a
n
u
a
l
ly
id
e
n
ti
fied
fro
m
n
u
m
e
ro
u
s
m
a
g
n
e
ti
c
re
so
n
a
n
c
e
ima
g
i
n
g
(
M
RI)
p
ict
u
re
s
i
n
m
e
d
ica
l
p
ra
c
ti
c
e
.
M
a
n
y
fra
m
e
wo
rk
s
b
a
se
d
o
n
b
ra
i
n
tu
m
o
rs
a
re
d
iag
n
o
se
d
u
si
n
g
b
o
t
h
d
e
e
p
lea
rn
in
g
a
n
d
m
a
c
h
in
e
lea
rn
in
g
.
T
h
is
stu
d
y
p
ro
p
o
se
s
a
W
a
ss
e
rste
in
d
e
e
p
c
o
n
v
o
l
u
ti
o
n
a
l
g
e
n
e
ra
ti
v
e
a
d
v
e
rsa
rial
n
e
tw
o
rk
(
WDCG
AN
)
o
p
ti
m
ize
d
u
si
n
g
t
h
e
a
rti
ficia
l
h
u
m
m
in
g
b
ir
d
o
p
t
imiz
a
ti
o
n
a
l
g
o
rit
h
m
(AH
BOA
)
f
o
r
b
ra
i
n
t
u
m
o
r
se
g
m
e
n
tatio
n
a
n
d
c
las
sifica
ti
o
n
(
S
CBT)
.
F
irst
,
th
e
Bra
TS
d
a
tas
e
t
is
u
se
d
to
g
a
th
e
r
th
e
in
p
u
t
d
a
ta.
Th
e
n
it
i
s
p
re
-
p
ro
c
e
ss
e
d
c
o
n
su
m
i
n
g
a
d
a
p
ti
v
e
se
lf
-
g
u
i
d
e
d
f
il
terin
g
(A
S
G
F
)
a
n
d
th
e
r
e
su
lt
is
se
g
m
e
n
ted
u
sin
g
f
u
z
z
y
p
o
ss
ib
il
isti
c
C
-
o
rd
e
re
d
m
e
a
n
c
lu
ste
rin
g
(F
P
COMC).
Afte
r
th
a
t,
fe
a
tu
re
s
a
re
e
x
trac
ted
u
sin
g
t
h
e
d
u
a
l
tree
c
o
m
p
le
x
d
is
c
re
te
wa
v
e
let
tran
sfo
rm
(
DT
-
CD
WT
).
T
h
e
c
h
a
ra
c
teristics
o
f
fe
a
tu
re
e
x
trac
t
e
d
a
re
fe
d
to
WDCG
AN
fo
r
e
ffe
c
ti
v
e
l
y
c
a
teg
o
rize
th
e
v
a
rio
u
s
p
a
ra
m
e
ters
.
Th
e
n
t
h
e
p
r
o
p
o
se
d
M
ATLAB
is
u
se
d
t
o
imp
lem
e
n
t
th
e
tec
h
n
i
q
u
e
,
a
n
d
t
h
e
p
e
rfo
rm
a
n
c
e
m
e
a
su
re
m
e
n
ts
li
k
e
F
1
-
sc
o
re
,
a
c
c
u
ra
c
y
,
e
rro
r
ra
te,
p
re
c
isio
n
,
se
n
siti
v
it
y
,
m
e
a
n
sq
u
a
re
e
rro
r
,
re
c
e
iv
e
r
o
p
e
ra
ti
n
g
c
h
a
ra
c
teristic
(ROC)
,
a
n
d
c
o
m
p
u
tatio
n
a
l
ti
m
e
a
re
a
n
a
l
y
z
e
d
.
T
h
e
WDCG
AN
-
AH
BOA
-
S
CBT
m
e
t
h
o
d
sig
n
ifi
c
a
n
t
ly
imp
ro
v
e
s
p
re
c
isio
n
i
n
S
CBT
b
y
i
n
teg
ra
ti
n
g
a
d
a
p
ti
v
e
o
p
ti
m
iza
ti
o
n
stra
teg
ies
,
re
su
lt
in
g
in
3
2
.
1
8
,
3
2
.
7
5
,
a
n
d
3
2
.
9
0
%
h
ig
h
e
r
p
re
c
is
io
n
in
c
o
n
tras
t
to
c
u
rre
n
t
tec
h
n
i
q
u
e
s.
Th
is
d
e
m
o
n
stra
tes
th
a
t
th
e
a
p
p
r
o
a
c
h
i
s
m
o
re
a
c
c
u
ra
te
a
n
d
e
ffe
c
ti
v
e
,
m
a
k
in
g
i
t
a
re
li
a
b
le t
o
o
l
fo
r
m
e
d
ica
l
d
iag
n
o
si
s.
K
ey
w
o
r
d
s
:
Ad
ap
tiv
e
s
elf
-
g
u
id
e
d
f
ilter
in
g
Ad
v
er
s
ar
ial
n
etwo
r
k
Ar
tific
ial
h
u
m
m
in
g
b
ir
d
B
r
ain
tu
m
o
r
d
etec
tio
n
Du
al
tr
ee
co
m
p
lex
d
is
cr
ete
w
av
elet
tr
an
s
f
o
r
m
Fu
zz
y
p
o
s
s
ib
ilis
tic
C
-
o
r
d
er
ed
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
:
Ar
ap
p
alee
s
war
an
Mu
r
u
g
an
a
n
d
h
am
Dep
ar
tm
en
t o
f
E
lectr
o
n
ics an
d
C
o
m
m
u
n
icatio
n
E
n
g
in
ee
r
i
n
g
,
R
ajar
ajeswar
i Co
lleg
e
o
f
E
n
g
in
ee
r
in
g
B
an
g
alo
r
e,
I
n
d
ia
E
m
ail:
a.
m
u
r
u
g
an
an
d
h
am
@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
Am
o
n
g
th
e
m
o
s
t
h
o
r
r
if
y
i
n
g
il
ln
ess
es
o
f
th
e
m
o
d
er
n
d
a
y
is
b
r
ain
tu
m
o
r
.
T
h
e
m
o
s
t
co
m
m
o
n
r
ea
s
o
n
s
ar
e
th
e
ab
er
r
an
t
ce
lls
co
llectiv
e
b
eh
av
io
r
in
th
e
b
r
ain
[
1
]
.
I
n
b
io
lo
g
y
,
a
b
e
n
ig
n
tu
m
o
r
is
s
m
all
an
d
tin
y
in
ea
r
ly
s
tag
es
[
2
]
.
W
h
en
a
tu
m
o
r
is
co
n
s
id
er
ed
b
e
n
ig
n
in
b
io
lo
g
y
as
it
is
s
m
all
in
th
e
in
itial
s
tag
e
[
3
]
.
W
h
en
a
tu
m
o
r
r
ea
ch
es
th
e
s
ec
o
n
d
a
r
y
s
tag
e,
it
is
r
ef
er
r
ed
t
o
as
m
alig
n
an
t
s
in
ce
it
h
as
g
r
o
wn
b
ey
o
n
d
b
en
ig
n
b
o
u
n
d
ar
ies
an
d
is
g
r
ea
ter
in
s
ize
[
4
]
.
Ab
o
u
t
7
0
0
,
0
0
0
p
eo
p
le
in
th
e
USA
s
u
f
f
er
f
r
o
m
b
r
ai
n
tu
m
o
r
d
is
ea
s
e,
ac
co
r
d
in
g
to
th
e
Natio
n
al
B
r
ain
T
u
m
o
r
So
ciety
[
5
]
.
Of
th
o
s
e,
3
0
.
2
%
ar
e
m
alig
n
an
t
in
o
r
ig
in
,
an
d
th
e
r
e
m
ain
in
g
6
9
.
8
%
ar
e
b
en
ig
n
[
6
]
.
T
h
e
r
e
p
o
r
t
s
tates
th
at
o
n
ly
3
6
%
o
f
t
h
e
p
atien
ts
will
s
u
r
v
iv
e
[
7
]
.
A
b
o
u
t
8
7
,
0
0
0
in
2
0
2
0
,
in
d
iv
id
u
als
wer
e
d
iag
n
o
s
ed
with
b
r
ain
tu
m
o
r
s
[
8
]
.
T
h
er
e
wer
e
8
4
,
1
7
0
p
eo
p
le
with
b
r
ain
tu
m
o
r
s
in
2
0
2
1
,
ac
co
r
d
in
g
to
esti
m
ate
[
9
]
.
T
h
er
e
wer
e
6
9
,
9
5
0
p
er
s
o
n
s
o
v
e
r
4
0
with
a
d
ia
g
n
o
s
is
.
B
r
ain
tu
m
o
r
s
ar
e
class
i
f
ied
in
to
two
s
tag
es:
h
ig
h
-
g
r
ad
e
g
lio
m
a
(
HGG)
a
n
d
lo
w
-
g
r
a
d
e
g
lio
m
a
(
L
GG)
b
ased
o
n
t
h
eir
h
ig
h
m
o
r
tality
r
ate.
I
n
ad
d
itio
n
,
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
.
1
,
Feb
r
u
ar
y
2
0
2
6
:
429
-
4
4
2
430
co
m
p
ar
ed
to
HGG,
th
e
L
GG
s
u
r
v
iv
al
r
ate
is
q
u
ick
er
[
1
0
]
.
Sin
ce
th
e
a
v
er
ag
e
life
s
p
an
o
f
HGG
is
o
n
l
y
two
y
ea
r
s
,
p
r
o
m
p
t tr
ea
tm
en
t is n
ec
ess
ar
y
.
I
n
th
e
clin
ics,
v
ar
io
u
s
m
eth
o
d
s
ar
e
em
p
lo
y
ed
to
tr
ea
t
b
r
ain
t
u
m
o
r
s
[
1
1
]
.
R
ad
iatio
n
th
e
r
ap
y
is
h
elp
f
u
l
in
th
e
b
en
i
g
n
s
tag
e,
a
n
d
s
u
r
g
er
y
is
n
o
t
n
ec
ess
ar
y
f
o
r
th
e
p
atien
t
to
s
u
r
v
iv
e
[
1
2
]
.
C
o
n
v
e
r
s
ely
,
th
e
m
alig
n
an
t
s
tag
e
is
d
an
g
er
o
u
s
an
d
is
cu
r
ab
le
with
r
a
d
iatio
n
a
n
d
c
h
em
o
th
er
ap
y
[
1
3
]
.
As
a
r
esu
lt,
b
e
n
ig
n
tu
m
o
r
s
u
s
u
ally
s
p
r
ea
d
m
o
r
e
s
lo
wly
t
h
an
m
alig
n
an
t
o
n
es.
B
u
t
r
e
g
ar
d
less
o
f
th
e
s
itu
atio
n
,
d
ia
g
n
o
s
is
is
c
r
itical
an
d
r
eq
u
ir
es
q
u
alif
ied
r
ad
i
o
lo
g
is
ts
[
1
4
]
.
I
n
m
ed
ical
im
ag
in
g
,
m
o
r
e
co
n
tem
p
o
r
ar
y
im
ag
in
g
tech
n
o
lo
g
y
h
as
d
em
o
n
s
tr
ated
r
em
ar
k
ab
le
s
u
cc
ess
in
th
e
d
iag
n
o
s
is
an
d
d
etec
tio
n
o
f
s
er
io
u
s
h
u
m
an
d
is
ea
s
es,
co
u
n
tin
g
b
lo
o
d
ca
n
ce
r
,
lu
n
g
ca
n
ce
r
,
s
to
m
ac
h
ca
n
ce
r
,
b
r
ain
tu
m
o
r
s
,
an
d
a
h
o
s
t
o
f
o
th
e
r
co
n
d
itio
n
s
[
1
5
]
.
C
o
m
p
u
ted
to
m
o
g
r
ap
h
y
(
C
T
)
an
d
m
ag
n
etic
r
eso
n
an
ce
im
ag
in
g
(
MRI)
s
ca
n
s
ar
e
m
o
r
e
h
elp
f
u
l
im
ag
in
g
m
eth
o
d
s
f
o
r
b
r
ain
tu
m
o
r
s
.
MRI
s
ca
n
s
o
u
tp
er
f
o
r
m
s
b
etter
th
an
C
T
i
m
ag
in
g
w
h
en
it
c
o
m
es
to
lea
r
n
ab
o
u
t
th
e
tex
tu
r
e
an
d
tu
m
o
r
s
h
ap
es.
Fo
r
th
e
p
u
r
p
o
s
e
o
f
illn
ess
p
r
ev
e
n
tio
n
an
d
tr
ea
tm
en
t,
ea
r
ly
id
en
tifi
ca
tio
n
o
f
b
r
ain
ca
n
ce
r
is
cr
u
cial.
Sev
er
al
d
ee
p
lear
n
in
g
m
eth
o
d
s
ar
e
p
r
o
p
o
s
ed
to
d
etec
t
b
r
ain
tu
m
o
r
s
.
Ho
wev
er
,
th
e
cu
r
r
en
t
ap
p
r
o
a
ch
to
b
r
ain
ca
n
ce
r
d
etec
tio
n
is
n
o
t
ac
cu
r
ate
en
o
u
g
h
an
d
tak
es
lo
n
g
e
r
to
co
m
p
u
t
e.
A
n
u
m
b
er
o
f
cu
r
r
e
n
t
tech
n
i
q
u
es
ar
e
o
f
f
er
e
d
to
ad
d
r
ess
th
is
p
r
o
b
lem
wh
ile
cl
ass
if
y
in
g
b
r
ain
tu
m
o
r
s
.
Ho
we
v
er
,
th
e
cu
r
r
en
t
m
et
h
o
d
in
cr
ea
s
es
ca
lcu
latio
n
tim
e
d
u
r
in
g
jo
b
ex
ec
u
tio
n
wh
ile
p
r
o
v
id
in
g
in
s
u
f
f
icien
t
p
r
ec
is
io
n
.
T
h
e
ap
p
r
o
ac
h
o
f
th
e
c
u
r
r
e
n
t
m
eth
o
d
s
in
s
p
ir
ed
th
is
d
ev
elo
p
m
en
t
[
1
6
]
–
[
2
1
]
.
B
r
ain
tu
m
o
r
s
ar
e
am
o
n
g
th
e
m
o
s
t
alar
m
in
g
illn
ess
es
f
ac
ed
b
y
in
d
iv
i
d
u
als
to
d
a
y
,
d
e
f
in
ed
b
y
th
e
ag
g
r
eg
ate
ac
tio
n
s
o
f
ab
e
r
r
an
t
b
r
ain
ce
lls
.
C
an
ce
r
co
n
tain
e
r
i
s
class
if
ied
in
to
b
en
ig
n
an
d
m
alig
n
an
t
ca
teg
o
r
ies.
I
n
itially
,
a
tu
m
o
r
is
ter
m
ed
b
en
ig
n
w
h
en
it
is
s
m
all
an
d
c
o
n
f
in
ed
to
its
o
r
i
g
in
al
lo
ca
tio
n
.
Ho
wev
e
r
,
as
it
p
r
o
g
r
ess
es
to
a
m
alig
n
an
t
s
tag
e,
it
s
u
r
p
ass
es
its
b
en
ig
n
b
o
u
n
d
ar
ies
an
d
b
ec
o
m
es
m
o
r
e
t
h
r
ea
ten
in
g
.
HGG
is
s
ig
n
if
ican
tly
d
ea
d
lier
,
with
an
av
er
ag
e
life
s
p
an
o
f
o
n
ly
two
y
ea
r
s
af
ter
d
iag
n
o
s
is
,
u
n
d
er
s
co
r
in
g
th
e
u
r
g
en
cy
o
f
m
ed
ical
in
ter
v
en
tio
n
.
B
r
ain
tu
m
o
r
s
ar
e
tr
ea
ted
in
v
ar
io
u
s
way
s
;
b
en
ig
n
tu
m
o
r
s
u
s
u
ally
r
eq
u
ir
e
r
ad
iatio
n
th
er
ap
y
with
o
u
t
r
e
q
u
ir
in
g
u
n
d
er
g
o
in
g
s
u
r
g
e
r
y
,
w
h
er
ea
s
m
ali
g
n
an
t o
r
ca
n
ce
r
o
u
s
t
u
m
o
r
n
ee
d
s
a
co
m
b
in
atio
n
o
f
r
ad
iatio
n
a
n
d
c
h
em
o
th
e
r
ap
y
.
T
h
e
MRI
an
d
C
T
s
ca
n
m
a
ch
in
er
ies
in
tr
o
d
u
ce
d
in
to
c
o
n
tem
p
o
r
a
r
y
m
e
d
ical
im
ag
in
g
h
av
e
p
lay
ed
an
im
p
o
r
tan
t
p
ar
t
i
n
th
e
id
en
tific
atio
n
an
d
d
ia
g
n
o
s
is
o
f
s
ev
er
e
d
is
ea
s
es
in
h
u
m
a
n
s
,
s
u
ch
as
b
r
ain
tu
m
o
r
s
.
MRI
is
ex
ce
p
tio
n
ally
g
o
o
d
at
d
is
tin
g
u
is
h
i
n
g
th
e
tex
tu
r
e
an
d
s
h
ap
e
o
f
th
e
tu
m
o
r
an
d
is
th
u
s
ex
tr
em
ely
v
alu
ab
le
f
o
r
ea
r
ly
c
an
ce
r
d
etec
tio
n
an
d
in
ter
v
en
ti
o
n
.
Ma
n
y
d
ee
p
lear
n
in
g
f
r
am
ewo
r
k
s
ar
e
p
lan
n
ed
f
o
r
co
n
tr
i
b
u
tio
n
to
b
r
ain
ca
n
c
er
id
en
tific
atio
n
with
an
aim
f
o
r
h
ig
h
ac
cu
r
ac
y
a
n
d
lo
w
co
m
p
u
tatio
n
tim
e.
I
n
g
en
er
al,
e
x
is
tin
g
m
eth
o
d
s
s
ee
m
to
b
e
f
a
u
lty
,
lea
d
in
g
to
h
ig
h
er
ca
lcu
latio
n
tim
es
an
d
u
n
s
atis
f
ac
to
r
y
p
r
ec
is
io
n
d
u
r
in
g
ex
ec
u
tio
n
.
Alt
h
o
u
g
h
p
r
e
v
i
o
u
s
wo
r
k
s
h
a
v
e
p
r
ese
n
te
d
h
o
w
d
ee
p
le
ar
n
i
n
g
i
n
f
lu
e
n
ce
s
i
d
en
ti
f
y
in
g
b
r
ai
n
t
u
m
o
r
s
,
th
e
y
h
a
v
e
n
o
t e
x
p
lic
itl
y
ad
d
r
ess
e
d
h
o
w
t
h
e
o
p
t
im
i
za
ti
o
n
al
g
o
r
it
h
m
s
i
n
f
lu
e
n
ce
e
n
h
a
n
ci
n
g
t
h
e
cl
ass
if
ica
ti
o
n
a
cc
u
r
ac
y
.
Mo
s
t
o
f
t
h
e
ex
is
ti
n
g
f
r
a
m
e
w
o
r
k
s
ar
e
b
ase
d
o
n
c
o
n
v
en
ti
o
n
al
m
a
c
h
i
n
e
l
ea
r
n
i
n
g
m
o
d
els
t
h
a
t
o
f
t
e
n
la
ck
th
e
p
r
ec
is
i
o
n
r
eq
u
i
r
ed
f
o
r
a
cc
u
r
at
e
d
i
ag
n
o
s
is
.
A
d
d
it
io
n
all
y
,
c
u
r
r
e
n
t m
e
th
o
d
s
t
e
n
d
t
o
i
n
c
r
e
ase
c
o
m
p
u
t
ati
o
n
t
im
e
a
n
d
r
e
q
u
i
r
e
s
ig
n
i
f
i
ca
n
t
m
an
u
al
in
te
r
v
e
n
ti
o
n
.
Als
o
,
t
h
e
e
x
is
ti
n
g
m
e
th
o
d
s
i
d
e
n
t
if
ies
t
h
e
s
p
ec
i
f
i
c
t
u
m
o
r
s
to
p
r
e
d
ic
t
a
n
d
n
e
ed
m
o
r
e
a
lg
o
r
it
h
m
s
ar
e
n
e
e
d
ed
t
o
ca
l
c
u
la
te
t
h
e
d
if
f
e
r
e
n
t
p
a
r
a
m
et
er
s
t
o
f
in
ali
ze
.
T
o
ad
d
r
ess
th
ese
g
ap
s
,
we
p
r
o
p
o
s
e
th
e
W
ass
er
s
tein
d
ee
p
co
n
v
o
lu
tio
n
al
g
en
e
r
ativ
e
a
d
v
er
s
ar
ial
n
etwo
r
k
(
W
DC
GAN)
o
p
tim
ized
th
r
o
u
g
h
ar
tific
ial
h
u
m
m
in
g
b
ir
d
o
p
tim
izatio
n
alg
o
r
ith
m
(
AHBOA
)
f
o
r
b
r
ai
n
tu
m
o
r
s
eg
m
en
tatio
n
an
d
class
if
icatio
n
(
SC
B
T
)
.
T
h
is
r
esear
ch
aim
s
to
im
p
r
o
v
e
b
r
ain
t
u
m
o
r
class
if
icatio
n
ac
cu
r
ac
y
b
y
in
co
r
p
o
r
atin
g
cu
ttin
g
-
ed
g
e
m
ac
h
in
e
lear
n
i
n
g
m
et
h
o
d
s
with
o
p
tim
izati
o
n
s
tr
ateg
ies.
T
h
e
f
o
llo
win
g
ar
e
th
is
s
tu
d
y
'
s
p
r
im
ar
y
co
n
t
r
ib
u
tio
n
s
:
i)
Pre
-
p
r
o
ce
s
s
in
g
:
i
n
p
u
t
im
ag
e
s
f
r
o
m
th
e
B
r
aT
S
d
ataset
ar
e
in
itially
p
r
e
-
p
r
o
ce
s
s
ed
u
s
in
g
ad
ap
tiv
e
s
elf
-
g
u
id
ed
f
ilter
in
g
(
ASGF)
to
co
r
r
ec
t c
o
r
r
u
p
ted
a
n
d
b
l
u
r
r
e
d
im
ag
es.
ii)
Seg
m
en
tatio
n
:
f
u
zz
y
p
o
s
s
ib
ili
s
tic
C
-
o
r
d
er
ed
m
ea
n
clu
s
ter
in
g
(
FP
C
OM
C
)
s
eg
m
en
ts
th
e
af
f
ec
ted
ar
ea
s
f
r
o
m
n
o
is
e
-
r
em
o
v
ed
im
ag
es,
p
r
ep
ar
in
g
th
em
f
o
r
th
e
s
u
b
s
eq
u
en
t stag
es.
iii)
Featu
r
e
e
x
tr
ac
tio
n
:
d
u
al
tr
ee
c
o
m
p
lex
d
is
cr
ete
wav
elet
tr
an
s
f
o
r
m
(
DT
-
C
DW
T
)
ex
tr
ac
ts
Har
alick
tex
tu
r
e
f
ea
tu
r
es
an
d
g
r
ay
s
ca
le
s
tatis
tical
f
ea
tu
r
es
a
r
e
ex
a
m
p
les
o
f
r
ad
i
o
m
ic
f
ea
tu
r
es,
f
r
o
m
th
e
s
eg
m
en
te
d
im
ag
es.
iv
)
C
las
s
if
icatio
n
:
u
s
in
g
th
e
W
D
C
GAN
,
th
e
p
r
o
ce
s
s
ed
im
ag
e
s
ar
e
ca
teg
o
r
ize
d
in
to
g
r
o
u
p
s
lik
e
g
lio
m
a
,
m
en
in
g
io
m
a
,
p
itu
itar
y
,
an
d
n
o
tu
m
o
r
.
v)
Op
tim
izatio
n
:
AHBOA
is
u
til
i
ze
d
to
m
ax
im
ize
W
DC
GA
N
'
s
weig
h
t
p
ar
am
eter
s
,
en
s
u
r
in
g
ac
cu
r
ate
b
r
ai
n
tu
m
o
r
class
if
icatio
n
.
2.
L
I
T
E
R
AT
U
RE
R
E
VI
E
W
T
h
e
liter
atu
r
e
o
n
t
h
e
s
u
b
d
iv
is
io
n
an
d
class
if
icatio
n
o
f
b
r
ain
tu
m
o
r
s
u
s
in
g
d
ee
p
lear
n
in
g
was
f
u
ll
o
f
s
tu
d
y
in
f
o
r
m
atio
n
;
s
o
m
e
cu
r
r
en
t
attem
p
ts
wer
e
in
clu
d
ed
h
er
e.
Ag
r
awa
l
et
al
.
[
1
6
]
h
av
e
s
u
g
g
ested
th
at
3D
-
UNe
t
d
ee
p
n
eu
r
al
n
etw
o
r
k
s
(
DNN)
p
er
f
o
r
m
b
r
ain
tu
m
o
r
class
if
icatio
n
an
d
s
eg
m
en
tatio
n
.
T
h
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
B
r
a
in
tu
mo
r
s
eg
men
ta
tio
n
a
n
d
cla
s
s
ifica
tio
n
u
s
in
g
a
r
tifi
cia
l h
u
mmin
g
b
ir
d
…
(
R
a
d
h
a
kris
h
n
a
n
K
a
r
th
ikey
a
n
)
431
p
r
e
-
p
r
o
ce
s
s
in
g
m
o
d
u
le
tak
es
th
e
p
r
e
-
p
r
o
ce
s
s
ed
im
ag
es
f
ir
s
t.
T
h
e
d
is
to
r
ted
an
d
f
u
zz
y
im
a
g
es
ar
e
f
ilter
ed
b
y
th
e
m
o
d
u
le
u
n
d
e
r
p
r
esen
tatio
n
.
T
h
e
p
r
esen
ted
ar
c
h
itectu
r
e
c
o
n
s
is
ted
o
f
a
b
etter
co
n
v
o
l
u
tio
n
al
n
eu
r
al
n
etwo
r
k
(
C
NN)
to
class
if
y
MRI
im
a
g
e
s
an
d
a
b
etter
3
D
-
UNe
t
p
r
o
to
t
y
p
e
f
o
r
s
eg
m
en
tin
g
v
o
lu
m
es
i
n
th
e
d
e
v
elo
p
m
e
n
t
o
f
an
o
b
jectiv
e
ex
p
er
t sy
s
tem
to
p
r
ed
ict
b
r
ain
tu
m
o
r
s
ea
r
ly
.
D
e
v
i
et
al
.
[
1
7
]
h
a
v
e
p
r
e
s
e
n
t
e
d
b
r
a
i
n
t
u
m
o
r
s
a
r
e
c
l
as
s
i
f
i
e
d
a
n
d
s
e
g
m
e
n
t
e
d
u
s
i
n
g
a
d
a
p
t
i
v
e
k
er
n
e
l
f
u
z
z
y
C
m
e
a
n
s
c
l
u
s
t
e
r
i
n
g
w
i
t
h
i
m
m
o
b
i
l
e
t
r
a
n
s
m
is
s
i
o
n
o
f
w
a
v
e
l
e
t
p
a
c
k
e
t
s
t
h
a
t
u
s
e
h
y
b
r
i
d
d
e
e
p
l
e
a
r
n
i
n
g
.
I
t
i
s
a
d
e
e
p
l
e
a
r
n
i
n
g
s
y
s
t
e
m
h
y
b
r
i
d
t
h
a
t
i
s
u
s
e
d
t
o
d
i
a
g
n
o
s
e
a
n
d
c
l
a
s
s
i
f
y
b
r
a
i
n
t
u
m
o
r
s
.
T
h
e
i
m
a
g
e
s
w
h
i
c
h
h
a
d
b
e
e
n
p
r
o
c
e
s
s
e
d
b
e
f
o
r
e
w
e
r
e
t
a
k
e
n
a
d
v
a
n
t
a
g
e
o
f
b
y
t
h
e
f
e
a
t
u
r
e
e
x
t
r
a
ct
i
o
n
p
r
o
c
e
s
s
.
A
c
o
m
b
i
n
a
ti
o
n
o
f
h
y
b
r
i
d
a
d
a
p
t
i
v
e
b
l
a
ck
w
i
d
o
w
o
p
t
i
m
i
z
at
i
o
n
w
it
h
m
o
t
h
f
l
a
m
e
o
p
t
i
m
i
z
a
ti
o
n
(
H
AB
W
MF
O
)
w
as
u
s
e
d
t
o
s
e
l
e
c
t
t
h
e
b
e
s
t
c
h
a
r
a
c
t
e
r
is
t
i
cs
.
T
o
p
e
r
f
o
r
m
d
i
s
s
e
ct
i
o
n
,
t
h
e
a
d
a
p
t
i
v
e
k
e
r
n
e
l
f
u
z
z
y
C
m
e
a
n
s
c
l
u
s
t
e
r
in
g
t
e
c
h
n
i
q
u
e
w
a
s
a
d
j
u
s
t
e
d
.
Pre
eth
i
an
d
Ais
h
war
y
a
[
1
8
]
h
av
e
p
r
esen
ted
a
n
ef
f
ec
tiv
e
wa
v
elet
-
f
u
s
io
n
im
ag
e
f
u
s
io
n
tec
h
n
iq
u
e
o
f
MRI
im
ag
es
an
d
p
o
s
itro
n
em
is
s
io
n
to
m
o
g
r
a
p
h
y
(
PET
)
im
a
g
es
o
f
th
e
s
eg
m
e
n
tatio
n
an
d
d
etec
tio
n
o
f
b
r
ain
tu
m
o
r
s
.
T
h
e
p
r
o
p
o
s
ed
r
esear
c
h
p
r
esen
ts
an
ef
f
icien
t
f
u
s
io
n
-
b
ased
m
eth
o
d
o
f
class
if
y
in
g
a
n
d
d
etec
tin
g
b
r
ain
tu
m
o
r
s
.
T
h
is
is
d
o
n
e
b
y
f
ir
s
t
f
u
s
in
g
th
e
in
p
u
t
im
ag
e
with
th
e
h
elp
o
f
d
is
cr
ete
wav
elet
tr
a
n
s
f
o
r
m
an
d
a
s
p
ec
ial
f
u
s
io
n
r
u
le.
T
h
e
f
u
s
io
n
p
r
o
ce
s
s
r
etr
iev
ed
th
e
f
ea
tu
r
es
o
f
th
e
g
r
ay
-
lev
el
co
-
o
cc
u
r
r
en
ce
m
atr
ix
(
GL
C
M)
.
T
h
en
d
em
o
n
s
tr
ate
th
e
b
r
ain
im
a
g
es a
s
n
o
r
m
al
o
r
ab
n
o
r
m
al
with
an
id
ea
l
DNN
.
Ku
m
ar
an
d
Kar
ib
asap
p
a
[
1
9
]
h
av
e
p
r
esen
ted
an
attitu
d
e
d
ep
en
d
o
n
th
e
d
u
al
-
t
r
ee
co
m
p
lex
to
id
en
tify
b
r
ain
tu
m
o
r
s
Gab
o
r
wa
v
elet
tr
an
s
f
o
r
m
ati
o
n
(
DT
C
GW
T
)
u
s
in
g
Had
o
o
p
b
ig
d
ata
a
n
aly
s
is
alo
n
g
a
n
eu
r
al
n
etwo
r
k
.
I
n
th
e
p
r
esen
ted
s
tu
d
y
,
b
r
ain
tu
m
o
r
ca
teg
o
r
izatio
n
an
d
s
eg
m
en
tatio
n
ar
e
ac
h
iev
e
d
b
y
th
e
in
teg
r
atio
n
o
f
i
m
ag
e
p
r
o
ce
s
s
in
g
an
d
h
u
g
e
d
ata.
T
h
e
Had
o
o
p
s
y
s
tem
o
n
MA
T
L
AB
was
u
s
ed
f
o
r
im
m
en
s
e
d
ata
an
aly
s
is
o
f
th
e
b
r
ain
tu
m
o
r
im
ag
es.
T
h
e
d
ata
f
r
o
m
th
e
i
n
v
esti
g
atio
n
is
b
r
o
k
e
n
d
o
wn
u
s
in
g
th
e
r
ec
en
t
ly
cr
ea
ted
d
u
al
-
tr
ee
co
m
p
o
s
ite
Gab
o
r
wav
elet
tr
an
s
f
o
r
m
.
Qad
er
et
a
l.
[
2
0
]
h
av
e
ac
ce
s
s
ib
le
a
g
r
ea
ter
d
ee
p
co
n
v
o
l
u
tio
n
al
n
eu
r
al
n
etwo
r
k
(
DC
NN)
d
et
ec
tio
n
an
d
ca
teg
o
r
izatio
n
u
tili
zin
g
au
g
m
en
ted
MRI
im
a
g
es
b
y
em
p
lo
y
in
g
h
y
b
r
i
d
o
p
tim
izatio
n
s
tr
ateg
ies.
T
h
e
p
r
esen
te
d
wo
r
k
ad
o
p
ts
im
p
r
o
v
ed
o
p
tim
izatio
n
s
tr
ateg
ies
to
ass
u
r
e
a
d
ev
elo
p
m
en
t
in
d
ee
p
co
n
v
o
lu
tio
n
al
lear
n
in
g
.
Ad
d
itio
n
ally
,
Ots
u
th
r
esh
o
ld
i
n
g
was
r
u
m
m
ag
e
-
s
ale
to
d
iv
id
e
th
e
tu
m
o
r
to
s
ec
tio
n
s
th
at
f
o
cu
s
o
n
b
r
ain
tu
m
o
r
id
en
tific
atio
n
.
T
h
e
p
r
esen
tatio
n
o
f
th
e
p
r
esen
ted
tech
n
iq
u
e
is
ass
es
s
ed
th
r
o
u
g
h
e
x
p
er
i
m
en
tatio
n
s
o
n
2
,
0
7
3
en
h
an
ce
d
MRI
im
ag
es in
all.
Van
k
d
o
th
u
an
d
Ham
ee
d
[
2
1
]
h
av
e
s
h
o
wed
h
o
w
b
r
ain
t
u
m
o
r
s
wer
e
d
iv
id
e
d
in
th
e
m
ag
n
etic
r
eso
n
an
ce
im
a
g
es
ap
p
ly
in
g
th
e
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
(
SVM)
an
d
f
u
zz
y
class
if
ier
in
th
e
m
ac
h
in
e
lear
n
in
g
.
T
h
is
r
esear
ch
p
a
p
er
ex
p
lo
r
e
d
b
r
ain
tu
m
o
r
s
eg
m
e
n
tatio
n
b
y
th
e
u
s
e
o
f
C
T
.
W
h
ich
ev
er
p
r
e
p
r
o
ce
s
s
in
g
m
eth
o
d
o
n
e
em
p
lo
y
s
,
th
e
tech
n
iq
u
es
o
f
f
ea
tu
r
e
ex
tr
ac
tio
n
ar
e
r
e
v
is
ed
p
r
io
r
to
s
u
b
jectin
g
th
e
im
a
g
e
to
class
if
icatio
n
s
tep
s
,
wh
eth
er
b
ein
g
th
e
class
i
f
ier
u
s
ed
in
SVM
o
f
th
e
ad
ap
ti
v
e
n
eu
r
o
-
f
u
zz
y
in
f
e
r
en
ce
s
y
s
tem
(
ANFI
S)
o
r
n
o
t
to
d
eter
m
in
e
wh
et
h
er
it is
ab
n
o
r
m
al
o
r
n
o
r
m
al.
T
ab
le
1
p
r
o
v
id
es a
n
o
v
er
v
iew
o
f
th
e
r
ev
ie
wed
ap
p
r
o
ac
h
.
Alth
o
u
g
h
p
r
ev
i
o
u
s
r
esear
ch
h
as
ex
am
in
ed
h
o
w
d
ee
p
lear
n
in
g
af
f
ec
ts
th
e
id
en
tific
atio
n
o
f
b
r
ai
n
tu
m
o
r
s
,
th
ey
h
av
e
n
o
t
ex
p
licitly
ad
d
r
ess
ed
th
e
in
f
lu
en
ce
o
f
o
p
tim
izatio
n
alg
o
r
it
h
m
s
o
n
en
h
an
ci
n
g
class
if
icatio
n
ac
cu
r
ac
y
.
Ma
n
y
ex
is
tin
g
f
r
am
ewo
r
k
s
r
ely
o
n
s
tan
d
ar
d
m
ac
h
in
e
lear
n
i
n
g
m
o
d
els,
wh
ich
o
f
ten
lack
th
e
p
r
ec
is
io
n
n
ee
d
ed
f
o
r
ac
cu
r
ate
d
iag
n
o
s
is
.
Ad
d
itio
n
ally
,
cu
r
r
e
n
t
m
eth
o
d
s
ten
d
to
in
cr
ea
s
e
co
m
p
u
tatio
n
tim
e
an
d
r
eq
u
ir
e
s
ig
n
if
ican
t
m
an
u
al
in
ter
v
e
n
tio
n
.
T
o
ad
d
r
ess
th
ese
g
ap
s
,
th
is
wo
r
k
is
p
r
o
p
o
s
ed
.
T
ab
le
1
.
Ov
e
r
v
iew
o
f
t
h
e
ex
a
m
in
ed
m
eth
o
d
o
lo
g
y
R
e
f
e
r
e
n
c
e
M
e
t
h
o
d
s
O
b
j
e
c
t
i
v
e
s
G
a
p
s
A
g
r
a
w
a
l
e
t
a
l
.
[
1
6
]
,
2
0
2
2
C
N
N
D
e
v
e
l
o
p
a
n
e
x
p
e
r
t
s
y
st
e
m
f
o
r
e
a
r
l
y
b
r
a
i
n
t
u
m
o
r
p
r
e
d
i
c
t
i
o
n
,
a
c
h
i
e
v
e
p
r
e
c
i
se
seg
m
e
n
t
a
t
i
o
n
a
n
d
c
l
a
ssi
f
i
c
a
t
i
o
n
o
f
b
r
a
i
n
t
u
m
o
r
s
Li
mi
t
e
d
t
o
t
h
e
sp
e
c
i
f
i
c
st
r
u
c
t
u
r
e
o
f
3
D
-
U
N
e
t
,
may
r
e
q
u
i
r
e
si
g
n
i
f
i
c
a
n
t
c
o
m
p
u
t
a
t
i
o
n
a
l
r
e
so
u
r
c
e
s
f
o
r
p
r
o
c
e
ss
i
n
g
3
D
v
o
l
u
mes
D
e
v
i
e
t
a
l
.
[
1
7
]
,
2
0
2
2
H
A
B
W
M
F
O
En
h
a
n
c
e
b
r
a
i
n
t
u
mo
r
c
a
t
e
g
o
r
i
z
a
t
i
o
n
a
n
d
seg
m
e
n
t
a
t
i
o
n
a
c
c
u
r
a
c
y
,
i
d
e
n
t
i
f
y
o
p
t
i
m
a
l
f
e
a
t
u
r
e
s f
o
r
d
i
a
g
n
o
si
s
u
s
i
n
g
H
A
B
W
M
F
O
C
o
m
p
l
e
x
i
t
y
i
n
c
o
m
b
i
n
i
n
g
mu
l
t
i
p
l
e
o
p
t
i
m
i
z
a
t
i
o
n
st
r
a
t
e
g
i
e
s,
ma
y
n
o
t
g
e
n
e
r
a
l
i
z
e
w
e
l
l
t
o
d
i
f
f
e
r
e
n
t
d
a
t
a
se
t
s
P
r
e
e
t
h
i
a
n
d
A
i
sh
w
a
r
y
a
[
1
8
]
,
2
0
2
1
DNN
U
se
f
u
si
o
n
t
e
c
h
n
i
q
u
e
s
t
o
i
mp
r
o
v
e
seg
m
e
n
t
a
t
i
o
n
a
n
d
i
d
e
n
t
i
f
i
c
a
t
i
o
n
,
d
i
s
t
i
n
g
u
i
s
h
b
e
t
w
e
e
n
n
o
r
ma
l
a
n
d
a
b
n
o
r
ma
l
b
r
a
i
n
i
ma
g
e
s
u
si
n
g
d
e
e
p
l
e
a
r
n
i
n
g
P
o
t
e
n
t
i
a
l
l
o
ss
o
f
d
e
t
a
i
l
d
u
r
i
n
g
i
ma
g
e
f
u
si
o
n
,
l
i
m
i
t
e
d
f
o
c
u
s
o
n
sp
e
c
i
f
i
c
t
y
p
e
s
o
f
b
r
a
i
n
t
u
m
o
r
s
K
u
mar
a
n
d
K
a
r
i
b
a
sa
p
p
a
[
1
9
]
,
2
0
2
2
D
TC
G
W
T
I
n
t
e
g
r
a
t
e
b
i
g
d
a
t
a
w
i
t
h
i
ma
g
e
p
r
o
c
e
ss
i
n
g
f
o
r
a
c
c
u
r
a
t
e
c
a
t
e
g
o
r
i
z
a
t
i
o
n
,
l
e
v
e
r
a
g
e
w
a
v
e
l
e
t
t
r
a
n
sf
o
r
ms f
o
r
e
n
h
a
n
c
e
d
s
e
g
me
n
t
a
t
i
o
n
C
o
m
p
l
e
x
i
t
y
a
n
d
r
e
s
o
u
r
c
e
d
e
ma
n
d
s
o
f
b
i
g
d
a
t
a
p
r
o
c
e
ssi
n
g
,
l
i
mi
t
e
d
s
c
a
l
a
b
i
l
i
t
y
b
e
y
o
n
d
H
a
d
o
o
p
e
n
v
i
r
o
n
me
n
t
s
Q
a
d
e
r
e
t
a
l
.
[
2
0
]
,
2
0
2
2
D
C
N
N
Emp
l
o
y
h
y
b
r
i
d
o
p
t
i
m
i
z
a
t
i
o
n
t
o
i
mp
r
o
v
e
n
e
t
w
o
r
k
p
e
r
f
o
r
ma
n
c
e
,
a
c
c
u
r
a
t
e
l
y
d
e
t
e
c
t
,
a
n
d
c
a
t
e
g
o
r
i
z
e
b
r
a
i
n
t
u
m
o
r
s i
n
M
R
I
i
mag
e
s
D
e
p
e
n
d
e
n
c
e
o
n
a
u
g
m
e
n
t
e
d
i
m
a
g
e
s
m
a
y
l
i
mi
t
r
e
a
l
-
w
o
r
l
d
a
p
p
l
i
c
a
b
i
l
i
t
y
,
r
e
q
u
i
r
e
s
e
x
t
e
n
si
v
e
c
o
m
p
u
t
a
t
i
o
n
a
l
p
o
w
e
r
f
o
r
d
e
e
p
l
e
a
r
n
i
n
g
a
n
d
o
p
t
i
m
i
z
a
t
i
o
n
V
a
n
k
d
o
t
h
u
a
n
d
H
a
mee
d
[
2
1
]
,
2
0
2
2
S
V
M
-
A
N
F
I
S
U
t
i
l
i
z
e
S
V
M
a
n
d
f
u
z
z
y
c
l
a
ss
i
f
i
e
r
s
f
o
r
a
c
c
u
r
a
t
e
se
g
me
n
t
a
t
i
o
n
,
d
e
t
e
r
mi
n
e
t
h
e
a
b
n
o
r
m
a
l
i
t
y
i
n
C
T
i
m
a
g
e
s
Li
mi
t
e
d
a
c
c
u
r
a
c
y
i
n
sp
e
c
i
f
i
c
b
r
a
i
n
t
u
m
o
r
t
y
p
e
s
,
ma
y
n
o
t
h
a
n
d
l
e
n
o
i
s
e
a
n
d
a
r
t
i
f
a
c
t
s
e
f
f
e
c
t
i
v
e
l
y
i
n
C
T
sc
a
n
s
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
.
1
,
Feb
r
u
ar
y
2
0
2
6
:
429
-
4
4
2
432
3.
ME
T
H
O
D
W
DC
GAN
-
AH
B
OA
-
S
C
B
T
d
is
cu
s
s
e
s
th
e
b
r
ain
tu
m
o
r
an
al
y
s
is
in
af
f
ec
ted
in
d
iv
id
u
als
with
in
th
is
s
ec
tio
n
.
T
h
is
m
o
d
el
in
clu
d
e
s
f
o
u
r
d
if
f
e
r
en
t
p
r
o
jecte
d
b
r
ain
tu
m
o
r
c
o
n
d
itio
n
s
:
g
lio
m
as,
m
en
in
g
io
m
as,
p
itu
itar
ies,
an
d
n
o
tu
m
o
r
im
a
g
es.
T
h
e
p
r
o
to
ty
p
e
u
n
d
er
g
o
es
in
itial
d
ata
co
llectio
n
b
y
b
r
a
in
tu
m
o
r
illn
ess
es
b
ef
o
r
e
b
ein
g
f
o
r
war
d
e
d
f
o
r
ad
d
itio
n
al
p
r
o
ce
s
s
in
g
.
T
h
ese
ch
ap
ter
s
g
o
th
r
o
u
g
h
f
o
u
r
m
ain
d
ev
elo
p
m
en
ts
:
p
r
ep
r
o
ce
s
s
in
g
,
s
eg
m
en
tatio
n
,
f
ea
tu
r
e
ex
tr
ac
tio
n
,
an
d
ca
teg
o
r
izatio
n
,
in
later
s
ec
to
r
s
.
T
h
u
s
,
th
e
ex
p
lan
atio
n
is
p
r
esen
ted
ac
co
r
d
in
g
ly
,
an
d
Fig
u
r
e
1
r
ep
r
esen
ts
th
e
r
ec
o
m
m
e
n
d
ed
W
DC
GAN
-
AH
B
OA
-
S
C
B
T
f
lo
w.
I
n
p
u
t
d
a
t
a
f
r
o
m
B
r
a
T
S
D
a
t
a
s
e
t
P
r
e
-
P
r
o
c
e
s
s
i
n
g
U
s
i
n
g
A
d
a
p
t
i
v
e
S
e
l
f
-
G
u
i
d
e
d
F
i
l
t
e
r
i
n
g
I
m
a
g
e
S
e
g
m
e
n
t
a
t
i
o
n
b
y
F
u
z
z
y
P
o
s
s
i
b
i
l
i
s
t
i
c
C
-
O
r
d
e
r
e
d
M
e
a
n
C
l
u
s
t
e
r
i
n
g
F
e
a
t
u
r
e
E
x
t
r
a
c
t
i
o
n
u
s
i
n
g
D
u
a
l
T
r
e
e
C
o
m
p
l
e
x
D
i
s
c
r
e
t
e
W
a
v
e
l
e
t
T
r
a
n
s
f
o
r
m
C
l
a
s
s
i
f
i
c
a
t
i
o
n
U
s
i
n
g
W
a
s
s
e
r
s
t
e
i
n
D
e
e
p
C
o
n
v
o
l
u
t
i
o
n
a
l
G
e
n
e
r
a
t
i
v
e
A
d
v
e
r
s
a
r
i
a
l
N
e
t
w
o
r
k
O
p
t
i
m
i
z
a
t
i
o
n
U
s
i
n
g
A
r
t
i
f
i
c
i
a
l
H
u
m
m
i
n
g
B
i
r
d
O
p
t
i
m
i
z
a
t
i
o
n
A
l
g
o
r
i
t
h
m
G
l
i
o
m
a
M
e
n
i
n
g
i
o
m
a
P
i
t
u
i
t
a
r
y
N
o
T
u
m
o
r
I
m
a
g
e
Fig
u
r
e
1
.
W
o
r
k
f
lo
w
o
f
th
e
p
r
o
p
o
s
ed
W
DC
GAN
-
AH
B
OA
-
S
C
B
T
f
r
am
ewo
r
k
3
.
1
.
Da
t
a
a
cquis
it
io
n
T
h
e
d
ata
u
s
ed
h
er
e
is
b
ased
o
n
B
r
aT
S
d
ataset
[
2
2
]
.
B
r
aT
S
h
av
e
co
n
tin
u
o
u
s
ly
f
o
cu
s
ed
o
n
s
eg
m
en
tin
g
b
r
ain
tu
m
o
r
s
in
m
u
ltimo
d
al
MRI
d
ata
with
s
tate
-
of
-
th
e
-
a
r
t
m
eth
o
d
s
.
B
r
aT
S
2
0
2
0
u
s
e
s
m
u
lti
-
in
s
titu
tio
n
al
p
r
e
-
o
p
er
ativ
e
MRI
im
a
g
in
g
f
o
r
s
eg
m
en
tin
g
in
t
r
in
s
ically
h
eter
o
g
en
eo
u
s
b
r
ai
n
tu
m
o
r
s
,
s
u
ch
as
g
lio
m
as.
I
n
o
r
d
er
to
ass
ess
th
e
clin
ical
v
alid
ity
o
f
th
is
s
eg
m
en
tatio
n
t
ask
,
ev
alu
ate
th
e
o
v
er
all
p
atien
t'
s
s
u
r
v
iv
al,
an
d
q
u
an
tify
th
e
tu
m
o
r
'
s
f
alse
p
r
o
g
r
ess
io
n
ag
ain
s
t
its
ac
tu
al
r
ec
u
r
r
en
ce
,
B
r
aT
S'
2
0
also
u
s
es
m
ac
h
in
e
lear
n
in
g
alg
o
r
ith
m
s
an
d
in
teg
r
ated
an
al
y
s
es o
f
r
ad
io
m
ics f
ea
tu
r
es.
3
.
2
.
Ada
ptiv
e
s
elf
-
g
uid
ed
f
ilte
ring
-
ba
s
ed
pre
-
pro
ce
s
s
ing
ASGF
[
2
3
]
tech
n
iq
u
e
is
u
tili
ze
d
f
o
r
p
r
e
-
p
r
o
ce
s
s
in
g
th
e
b
r
ain
tu
m
o
r
im
ag
e.
Her
e,
th
e
n
o
is
es
ar
e
in
d
if
f
er
en
t
f
r
o
m
th
e
im
ag
es
wh
ile
p
r
o
ce
s
s
in
g
th
e
b
r
ain
tu
m
o
r
im
ag
es
o
f
th
e
cr
is
p
ed
g
es
an
d
it
b
ec
o
m
es
n
o
is
e
-
f
r
ee
d
u
r
in
g
th
e
f
lash
/n
o
-
f
lash
d
e
n
o
is
in
g
d
o
es
t
h
e
g
u
id
ed
f
ilter
p
e
r
f
o
r
m
v
e
r
y
well
in
ter
m
s
o
f
e
d
g
e
p
r
eser
v
atio
n
.
ASGF
m
eth
o
d
i
s
u
s
ed
to
ad
ap
tiv
ely
r
eg
u
late
th
e
r
eg
u
lar
izatio
n
p
ar
am
eter
i
n
ea
ch
s
tag
e
o
f
th
e
ad
ap
tiv
e
g
u
id
ed
f
ilter
in
g
p
r
o
c
ess
o
f
im
ag
es.
T
h
en
th
e
r
eg
u
lar
izatio
n
is
g
iv
en
as (
1
)
.
+
1
=
1
|
|
∑
+
∈
(
1
)
W
h
er
e
=
0
,
1
,
2
,
.
.
.
;
+
1
d
en
o
ted
as
th
e
g
u
id
an
ce
im
ag
e;
is
u
p
d
ated
d
y
n
a
m
ically
u
s
in
g
th
e
o
u
tco
m
es
o
f
ea
ch
c
y
cle,
u
s
in
g
t
h
e
in
itial
g
u
id
in
g
im
ag
e
as
th
e
i
n
p
u
t
im
ag
e;
d
en
o
ted
as
th
e
ce
n
t
r
al
p
i
x
el
o
f
th
e
cu
r
r
e
n
t
win
d
o
w
,
wh
ich
is
an
d
in
d
icate
s
th
e
n
th
iter
atio
n
o
f
ad
ap
tiv
e
g
u
id
ed
f
ilter
in
g
u
s
in
g
in
p
u
t
(
)
.
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
B
r
a
in
tu
mo
r
s
eg
men
ta
tio
n
a
n
d
cla
s
s
ifica
tio
n
u
s
in
g
a
r
tifi
cia
l h
u
mmin
g
b
ir
d
…
(
R
a
d
h
a
kris
h
n
a
n
K
a
r
th
ikey
a
n
)
433
I
t
is
cr
u
cial
to
r
em
em
b
e
r
th
at
in
th
e
cu
r
r
e
n
t
s
tu
d
y
,
t
h
is
r
eg
u
lar
izatio
n
v
al
u
e
wh
ic
h
is
p
r
o
v
i
d
ed
b
y
(
2
)
af
f
ec
ts
wh
eth
er
th
e
win
d
o
w
h
a
s
lar
g
e
v
ar
iatio
n
o
r
is
f
lat.
(
,
)
=
∑
(
(
+
−
)
2
+
(
)
2
)
∈
(
2)
W
h
er
e
=
1
,
2
,
3
,
.
.
.
.
;
is
th
e
an
ticip
ated
n
o
is
e
lev
el
th
at
d
eter
m
in
es
th
e
r
eg
u
lar
izatio
n
p
ar
am
eter
;
is
u
p
d
ated
d
y
n
a
m
ically
u
s
in
g
th
e
o
u
tco
m
es
o
f
ea
ch
c
y
cle,
u
s
i
n
g
th
e
in
itial
g
u
id
in
g
im
a
g
e
a
s
th
e
in
p
u
t
im
ag
e;
d
en
o
ted
as
th
e
ce
n
tr
al
p
ix
el
o
f
th
e
cu
r
r
en
t
win
d
o
w
,
wh
ich
is
;
in
d
icate
s
th
e
n
th
iter
atio
n
o
f
ad
ap
tiv
e
g
u
id
ed
f
ilter
in
g
u
s
in
g
in
p
u
t
(
)
;
an
d
d
en
o
tes
th
e
r
e
g
u
lar
izati
o
n
p
ar
a
m
eter
o
f
th
e
win
d
o
w's
f
ea
tu
r
e.
B
y
p
r
o
ce
s
s
in
g
th
is
s
y
s
tem
,
th
e
B
r
aT
S
d
ataset
'
s
in
p
u
t
d
ata
is
c
lean
s
ed
o
f
im
ag
e
n
o
is
e.
Nex
t,
th
e
s
eg
m
en
tatio
n
p
h
ase
r
ec
eiv
es th
e
p
r
e
-
p
r
o
ce
s
s
ed
d
ata.
3
.
3
.
Seg
m
ent
a
t
i
o
n us
ing
f
uz
zy
po
s
s
ibi
li
s
t
ic
C
-
o
rder
ed
m
ea
n c
lus
t
er
ing
I
n
t
h
is
s
e
cti
o
n
,
FP
C
OM
C
[
2
4
]
m
et
h
o
d
is
u
s
e
d
f
o
r
s
eg
m
en
t
in
g
t
h
e
a
f
f
ec
te
d
p
ar
t
f
r
o
m
t
h
e
n
o
is
e
r
e
m
o
v
ed
i
m
a
g
e.
T
h
e
o
r
d
er
e
d
m
et
h
o
d
is
t
y
p
ic
al
a
n
d
th
e
f
l
ex
i
b
l
e
p
o
s
s
i
b
i
lis
ti
c
C
-
o
r
d
er
ed
m
ea
n
s
m
et
h
o
d
t
h
a
t
m
o
d
er
ates
t
h
e
o
u
t
lie
r
’
s
e
f
f
e
ct.
Ass
u
m
e
th
at
a
s
et
o
f
d
a
ta
wit
h
p
o
i
n
ts
is
d
i
v
i
d
e
d
i
n
t
o
t
h
e
a
f
f
e
cte
d
p
a
r
t
o
f
n
o
is
e
r
e
m
o
v
ed
i
m
a
g
es
b
y
t
h
e
FP
C
O
MC a
l
g
o
r
it
h
m
.
T
h
e
f
o
ll
o
wi
n
g
cr
i
te
r
i
o
n
f
u
n
cti
o
n
o
f
FP
C
O
MC
is
g
i
v
en
as
(
3
)
.
∑
=
1
=
1
(
3
)
W
h
er
e
d
en
o
tes
th
e
k
th
p
o
in
t'
s
m
em
b
er
s
h
ip
d
eg
r
ee
in
r
elatio
n
to
th
e
i
th
v
alu
e
a
n
d
,
,
r
ep
r
esen
t
ed
as
th
e
p
ar
am
eter
o
f
FP
C
OM
C
.
T
h
is
co
m
p
ar
ativ
ely
h
ig
h
r
a
n
k
in
g
was
s
u
p
p
lied
b
y
th
e
tu
m
o
u
r
s
eg
m
en
tatio
n
in
s
ag
ittal v
iew
p
ictu
r
es.
T
h
en
th
e
d
is
tan
ce
wh
ich
is
m
ea
s
u
r
ed
wh
ile
s
eg
m
en
tin
g
is
g
iv
en
as (
4
)
.
=
1
1
+
[
;
]
(
4
)
W
h
er
e
d
en
o
ted
as
th
e
ty
p
icality
m
atr
ix
;
d
en
o
tes
th
e
v
ar
io
u
s
f
ac
to
r
s
in
f
lu
e
n
ce
m
em
b
er
s
h
ip
an
d
ty
p
icality
;
d
en
o
tes
th
e
d
is
tan
ce
wh
ich
is
m
ea
s
u
r
ed
u
s
in
g
th
e
lo
s
s
f
u
n
ctio
n
;
r
ep
r
esen
ts
th
e
k
th
p
o
in
t
ty
p
icality
f
o
r
th
e
i
th
v
alu
e;
an
d
,
,
r
ep
r
esen
ted
as
th
e
p
ar
a
m
eter
o
f
FP
C
OM
C
.
T
h
e
p
er
f
o
r
m
an
ce
o
f
clu
s
ter
in
g
will
b
e
im
p
ac
ted
b
y
th
e
s
en
s
itiv
ity
o
f
n
o
is
e
an
d
o
u
tlier
s
in
th
e
d
ata
s
eg
m
en
ti
n
g
th
e
af
f
ec
ted
p
ar
t
f
r
o
m
th
e
n
o
is
e
r
em
o
v
ed
im
a
g
e
.
T
h
en
th
e
af
f
ec
ted
p
ar
t o
f
im
a
g
e
is
g
iv
en
as (
5
)
.
[
]
=
[
∑
[
(
)
+
(
)
]
[
]
=
1
]
[
∑
[
(
)
+
(
)
]
[
]
=
1
]
(
5
)
W
h
er
e
d
en
o
tes
th
e
v
ar
io
u
s
f
ac
to
r
s
in
f
lu
en
ce
m
em
b
er
s
h
ip
a
n
d
ty
p
icality
;
[
]
is
eq
u
iv
alen
t
to
th
e
i
th
v
alu
e
ce
n
ter
in
th
e
s
th
iter
atio
n
;
d
en
o
tes
th
e
k
th
p
o
in
t'
s
m
em
b
er
s
h
ip
d
e
g
r
ee
in
r
elatio
n
t
o
th
e
i
th
v
al
u
e;
[
]
r
ep
r
esen
ts
th
e
p
ar
am
eter
t
h
at
is
b
ased
o
n
t
h
e
r
esid
u
al
a
n
d
t
h
e
lo
s
s
f
u
n
ctio
n
;
r
ep
r
esen
ts
th
e
k
th
p
o
i
n
t'
s
ty
p
icality
f
o
r
t
h
e
i
th
clu
s
ter
;
an
d
d
ep
en
d
e
n
t
u
p
o
n
h
o
w
ty
p
ical
th
e
p
o
i
n
t
is
ac
r
o
s
s
all
v
alu
es;
,
,
r
ep
r
esen
ted
as
th
e
p
a
r
am
eter
o
f
FP
C
OM
C
.
Fin
ally
,
th
e
af
f
e
cted
p
ar
t
im
ag
es
a
r
e
s
eg
m
e
n
ted
f
r
o
m
th
e
n
o
is
e
r
em
o
v
ed
im
a
g
e,
af
te
r
wh
ich
th
e
f
ea
tu
r
e
ex
t
r
ac
tio
n
s
ec
tio
n
r
ec
eiv
es th
e
s
eg
m
en
te
d
im
ag
e.
3
.
4
.
F
e
a
t
ure
ex
t
r
a
ct
io
n by
d
ua
l t
re
e
co
m
plex
dis
cr
et
e
wa
v
elet
t
ra
ns
f
o
r
m
s
T
h
e
s
eg
m
en
ted
af
f
ec
te
d
p
a
r
t
p
ictu
r
es
ar
e
s
en
t
t
o
f
ea
t
u
r
e
e
x
tr
ac
tio
n
a
n
d
th
e
f
ea
tu
r
es
ar
e
ex
tr
ac
ted
u
tili
zin
g
DT
-
C
DW
T
[
2
5
]
.
DT
-
C
DW
T
ex
tr
ac
ts
f
ea
tu
r
es
ac
co
r
d
in
g
ly
f
r
o
m
th
e
s
eg
m
en
ted
i
m
ag
es.
B
y
u
tili
zin
g
DT
-
C
DW
T
th
e
g
r
ay
s
ca
le
s
tatis
tic
f
ea
tu
r
es
an
d
f
ea
tu
r
es
o
f
H
ar
alick
t
ex
tu
r
e
in
clu
d
e
m
ea
n
,
s
k
ewn
ess
,
co
n
tr
ast,
an
d
h
o
m
o
g
en
e
ity
.
T
h
e
c
r
ea
tio
n
o
f
a
n
o
v
el
ex
tr
ac
tio
n
o
p
e
r
ato
r
em
p
lo
y
in
g
r
id
g
e
cu
r
v
e
id
en
tific
atio
n
is
th
e
m
ain
o
b
jectiv
e
o
f
th
e
DT
-
C
D
W
T
b
ein
g
p
r
esen
ted
.
I
n
(
6
)
e
x
p
r
ess
es it.
(
)
=
∑
,
,
(
)
,
(
6
)
W
h
er
e
d
en
o
tes
s
ca
lin
g
f
ac
to
r
th
at
d
en
o
tes
f
r
e
q
u
en
cies
ar
e
in
v
er
s
e;
,
d
en
o
ted
as
s
et
o
f
weig
h
tin
g
co
ef
f
icien
ts
;
(
)
d
en
o
tes
s
h
if
t
alo
n
g
tim
e
ax
is
in
tr
o
d
u
ce
d
b
y
th
e
d
ilatio
n
p
ar
am
eter
;
an
d
,
(
)
ar
e
a
co
llectio
n
o
f
f
u
n
d
am
en
tal
f
u
n
ctio
n
s
th
at
ca
n
b
e
ac
q
u
ir
ed
b
y
alter
in
g
a
s
ca
lin
g
f
u
n
ctio
n
.
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
.
1
,
Feb
r
u
ar
y
2
0
2
6
:
429
-
4
4
2
434
3
.
4
.
1
.
G
ra
y
s
ca
le
s
t
a
t
is
t
ic
f
ea
t
ures
I
n
th
is
p
ar
t,
th
e
w
av
elet
t
r
a
n
s
f
o
r
m
was
u
s
ed
to
elim
in
ate
g
r
ay
s
ca
le
s
tatis
tical
ch
ar
ac
te
r
is
tics
lik
e
m
ea
n
an
d
s
k
ewn
ess
.
Ad
d
itio
n
ally
,
th
e
f
o
llo
win
g
tech
n
iq
u
es
ar
e:
i)
Me
an
:
t
h
e
im
ag
e'
s
m
ea
n
co
lo
r
v
alu
e
ca
n
b
e
u
s
ed
to
d
ef
in
e
t
h
e
m
ea
n
.
I
t e
x
p
r
ess
ed
in
(
7
)
.
[
]
=
(
×
ℎ
)
[
ℎ
]
(
7
)
W
h
er
e
,
k
,
N
d
en
o
te
d
as
th
e
o
r
d
er
o
f
f
ilter
;
[
]
d
en
o
ted
as
th
e
d
ata
p
o
in
ts
o
f
m
ea
n
im
a
g
e
;
an
d
[
ℎ
]
d
en
o
ted
as a
p
p
r
o
x
im
atio
n
a
n
d
d
etail
co
ef
f
icien
t o
f
m
ea
n
im
a
g
e.
ii)
S
k
e
w
n
e
s
s
:
i
t
u
s
e
s
d
e
v
ia
t
i
o
n
as
a
m
e
a
s
u
r
e
o
f
t
h
e
d
e
g
r
e
e
o
f
as
y
m
m
e
t
r
y
i
n
t
h
e
d
i
s
t
r
i
b
u
t
i
o
n
.
I
n
(
8
)
e
x
p
r
e
s
s
es
i
t
.
[
]
=
∑
∞
=
−
∞
[
]
ℎ
[
−
]
(
8
)
W
h
er
e
j
,
k
,
N
d
en
o
ted
as
th
e
o
r
d
er
o
f
f
ilter
;
[
]
d
en
o
ted
as
th
e
d
ata
p
o
in
ts
o
f
m
ea
n
im
a
g
e;
[
ℎ
]
d
en
o
te
d
as
ap
p
r
o
x
im
ati
o
n
a
n
d
d
etail
co
e
f
f
icien
t
o
f
m
ea
n
i
m
ag
e
;
an
d
[
]
co
m
p
u
ted
b
y
r
u
n
n
in
g
it
th
r
o
u
g
h
s
ev
er
al
s
eg
m
e
n
ts
o
f
i
m
ag
e.
3
.
4
.
2
.
H
a
ra
lic
k
t
e
x
t
ure
f
ea
t
ures
I
n
th
is
s
ec
tio
n
,
th
e
wav
elet
tr
an
s
f
o
r
m
was
u
s
ed
to
elim
in
at
e
Har
alick
t
ex
tu
r
e
p
r
o
p
er
ties
in
clu
d
in
g
h
o
m
o
g
en
eity
an
d
co
n
tr
ast.
Ad
d
itio
n
ally
,
th
e
f
o
llo
win
g
tec
h
n
iq
u
es a
r
e:
i)
C
o
n
tr
ast:
it
ca
lcu
lates
th
e
d
en
s
ity
co
n
tr
ast
b
etwe
en
a
p
ix
e
l
an
d
its
s
u
r
r
o
u
n
d
in
g
p
i
x
els
th
r
o
u
g
h
o
u
t
th
e
wh
o
le
im
ag
e.
I
n
(
9
)
ex
p
lain
s
th
e
co
n
tr
ast.
[
]
=
∑
[
]
ℎ
[
2
−
]
∞
=
−
∞
(
9
)
W
h
er
e
[
]
s
y
n
th
esized
to
lo
w
-
f
r
e
q
u
en
cy
o
f
DT
-
C
DW
T
co
ef
f
icien
t;
,
,
d
en
o
te
d
as
th
e
o
r
d
er
o
f
f
ilter
;
[
]
d
en
o
ted
as
th
e
d
ata
p
o
in
ts
o
f
m
e
an
im
ag
e;
[
ℎ
]
d
e
n
o
ted
as
ap
p
r
o
x
im
atio
n
an
d
d
etail
co
ef
f
icien
t o
f
m
ea
n
im
a
g
e
;
an
d
[
]
co
m
p
u
ted
b
y
r
u
n
n
in
g
it th
r
o
u
g
h
s
ev
er
al
s
eg
m
e
n
ts
o
f
im
ag
e.
ii)
Ho
m
o
g
en
eity
:
i
t
is
em
p
lo
y
ed
t
o
q
u
an
tify
h
o
w
clo
s
e
th
e
d
is
tr
i
b
u
tio
n
o
f
GL
C
M
elem
en
ts
is
ap
p
r
o
x
im
ated
to
th
e
GL
C
M
d
iag
o
n
al.
I
n
(
1
0
)
ex
p
lain
s
th
e
h
o
m
o
g
en
eity
.
ℎ
[
]
=
∑
[
]
ℎ
[
2
−
]
∞
=
−
∞
(
1
0
)
W
h
er
e
ℎ
[
]
s
y
n
th
esized
to
h
ig
h
-
f
r
eq
u
en
cy
o
f
DT
-
C
DW
T
co
ef
f
i
cien
t;
j
,
k
,
N
d
en
o
ted
as
th
e
o
r
d
er
o
f
f
ilter
;
[
]
d
en
o
ted
as
th
e
d
ata
p
o
in
ts
o
f
m
e
an
im
ag
e;
[
ℎ
]
d
e
n
o
ted
as
ap
p
r
o
x
im
atio
n
an
d
d
etail
co
ef
f
icien
t o
f
m
ea
n
im
a
g
e
;
an
d
[
]
co
m
p
u
ted
b
y
r
u
n
n
in
g
it th
r
o
u
g
h
s
ev
er
al
s
eg
m
e
n
ts
o
f
im
ag
e.
Fin
ally
,
g
r
a
y
s
ca
le
s
tatis
tical
a
n
d
Har
alick
tex
tu
r
e
f
ea
t
u
r
es
a
r
e
ex
tr
ac
ted
f
r
o
m
th
e
i
m
ag
es
to
ca
p
tu
r
e
im
p
o
r
tan
t
in
ten
s
ity
an
d
tex
t
u
r
e
in
f
o
r
m
atio
n
.
T
h
ese
e
x
tr
ac
ted
f
ea
tu
r
es
a
r
e
th
e
n
u
s
ed
as
in
p
u
t
to
th
e
W
DC
GAN
m
o
d
el.
Usi
n
g
th
is
ap
p
r
o
ac
h
,
th
e
m
o
d
el
class
if
ies
th
e
im
ag
es
in
to
f
o
u
r
ca
teg
o
r
ies:
g
lio
m
a,
m
en
i
n
g
io
m
a
,
p
itu
itar
y
tu
m
o
r
,
an
d
n
o
tu
m
o
r
.
3
.
5
.
Cla
s
s
if
ica
t
io
n us
ing
WDCG
AN
W
DC
GAN
[
2
6
]
is
d
is
cu
s
s
ed
i
n
th
is
s
ec
tio
n
.
W
DC
GAN
g
en
er
ates
to
cr
ea
te
a
n
ew
d
ata
b
y
r
an
d
o
m
l
y
ad
d
in
g
n
o
is
e
an
d
f
itti
n
g
im
a
g
e
s
to
class
if
y
th
e
b
r
ain
tu
m
o
r
.
W
DC
GAN
ca
n
clas
s
if
y
th
e
o
u
tp
u
t
ty
p
e
b
y
ad
d
i
n
g
co
n
d
itio
n
al
d
ata
to
th
e
lab
el
o
f
th
e
g
en
er
ato
r
.
In
(
1
1
)
r
ep
r
es
en
ts
th
e
two
co
m
p
o
n
en
ts
o
f
th
e
W
D
C
GAN
's
to
ta
l
ad
d
itio
n
co
n
d
itio
n
al.
=
∑
=
1
(
1
1
)
W
h
er
e
r
ep
r
esen
t
th
e
in
p
u
t
o
f
W
DC
GAN
an
d
r
ep
r
esen
t
th
e
o
u
tp
u
ts
o
f
W
DC
GA
N
.
T
h
e
o
r
ig
in
al
im
ag
es
ar
e
class
if
ied
to
cr
ea
te
th
e
b
r
a
in
tu
m
o
r
im
ag
es,
w
h
ich
is
th
e
n
f
ed
in
to
th
e
W
DC
GAN
clas
s
if
ier
an
d
it
is
g
iv
en
as
in
(
1
2
)
.
(
)
=
−
≈
[
(
)
]
(
1
2
)
W
h
er
e
(
)
d
en
o
ted
as
th
e
tr
an
s
f
o
r
m
ed
i
m
ag
es;
[
(
)
]
d
en
o
ted
as
th
e
s
u
b
s
am
p
lin
g
lay
er
is
to
less
en
th
e
alter
ed
d
ata'
s
v
ar
ian
ce
;
an
d
≈
r
ep
r
esen
ts
th
e
v
alu
es
o
f
a
s
p
ec
if
ic
ch
ar
ac
ter
is
tic
in
a
s
ec
tio
n
o
f
th
e
in
p
u
t
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
B
r
a
in
tu
mo
r
s
eg
men
ta
tio
n
a
n
d
cla
s
s
ifica
tio
n
u
s
in
g
a
r
tifi
cia
l h
u
mmin
g
b
ir
d
…
(
R
a
d
h
a
kris
h
n
a
n
K
a
r
th
ikey
a
n
)
435
lay
er
.
T
h
e
W
DC
GAN
cla
s
s
if
i
er
u
s
es
th
e
ex
ten
d
ed
d
ata
as
its
in
p
u
t,
an
d
to
class
if
y
d
i
r
ec
tly
f
r
o
m
th
e
d
ata
an
d
it is
g
iv
en
as
in
(
1
3
)
.
(
)
=
−
≈
[
(
)
+
≈
[
(
)
]
]
(
1
3
)
W
h
er
e
(
)
d
en
o
ted
as
th
e
tr
an
s
f
o
r
m
ed
i
m
ag
es;
[
(
)
]
d
en
o
ted
as
th
e
s
u
b
s
am
p
lin
g
lay
er
is
to
less
en
th
e
alter
ed
d
ata'
s
v
ar
ian
ce
;
≈
r
ep
r
esen
ts
th
e
v
alu
es
o
f
a
s
p
ec
if
ic
c
h
ar
ac
ter
is
tic
in
a
s
ec
tio
n
o
f
th
e
in
p
u
t
lay
er
;
an
d
≈
d
en
o
te
d
as
th
e
m
a
x
-
p
o
o
lin
g
f
u
n
ctio
n
.
T
h
e
p
u
r
p
o
s
e
o
f
th
e
s
u
b
s
am
p
lin
g
la
y
er
is
to
less
en
th
e
v
ar
ian
ce
o
f
th
e
m
o
d
if
ied
d
ata
in
o
r
d
er
to
co
m
p
u
te
th
e
v
al
u
e
s
o
f
a
ce
r
tain
im
ag
es
in
th
e
s
e
ctio
n
o
f
in
p
u
t
lay
er
an
d
co
m
b
in
e
th
em
to
g
eth
e
r
.
T
h
en
th
e
v
ar
ian
ce
o
f
th
e
m
o
d
if
i
ed
d
ata
is
g
iv
en
as
in
(
1
4
)
.
(
,
)
=
−
(
|
)
(
1
4
)
W
h
er
e
r
ep
r
esen
ts
th
e
d
is
cr
im
in
ato
r
'
s
o
u
tp
u
t
an
d
(
|
)
d
en
o
ted
as
th
e
g
e
n
er
ated
d
ata
an
d
th
e
o
r
ig
in
al
d
ata
d
is
tr
ib
u
tio
n
.
Fin
ally
,
W
DC
GAN
clas
s
if
ies
b
r
ain
ca
n
ce
r
s
in
to
f
o
u
r
g
r
o
u
p
s
:
m
en
in
g
io
m
a,
p
itu
itar
y
,
g
lio
m
a,
an
d
n
o
tu
m
o
r
.
Her
e,
t
h
e
weig
h
t a
n
d
b
ias
p
ar
am
eter
s
o
f
W
DC
GA
N
ar
e
tu
n
ed
u
s
in
g
AHBOA.
3
.
6
.
O
pti
m
iza
t
io
n f
o
r
WDCG
AN
us
ing
AH
B
O
A
AHBOA
[
2
7
]
is
p
r
o
p
o
s
ed
to
en
h
an
ce
th
e
weig
h
ts
o
f
th
e
W
DC
GAN
.
T
h
e
weig
h
t
p
ar
a
m
eter
o
f
W
DC
GAN
is
o
p
tim
ized
u
s
in
g
th
e
p
r
o
p
o
s
ed
AHBOA.
AHB
OA
is
co
n
s
id
er
ed
th
e
s
m
allest
b
ir
d
s
in
th
e
wo
r
ld
,
h
u
m
m
in
g
b
ir
d
s
a
r
e
am
az
i
n
g
c
r
ea
tu
r
es.
Hu
m
m
i
n
g
b
ir
d
s
ar
e
th
e
m
o
s
t
in
tellectu
al
s
p
ec
ies
o
n
th
e
p
lan
et,
ev
e
n
h
u
m
an
s
,
if
i
n
tellig
en
ce
is
d
ete
r
m
in
ed
u
s
in
g
b
r
ain
-
to
-
b
o
d
y
r
a
tio
.
‒
Step
1
:
i
n
itializatio
n
.
AHBOA's
f
ir
s
t
p
o
p
u
latio
n
was
cr
e
ated
at
r
a
n
d
o
m
.
I
n
(
1
5
)
th
e
n
d
er
i
v
es
th
e
in
itializatio
n
.
=
+
.
(
−
)
(
15
)
W
h
er
e
‘
R
’
d
en
o
ted
as
th
e
r
an
d
o
m
v
ec
to
r
b
etwe
en
[
0
an
d
1
]
;
d
en
o
tes
th
e
i
th
f
o
o
d
s
u
p
p
l
y
f
its
in
to
th
e
s
o
lu
tio
n
o
f
a
p
ar
ticu
lar
is
s
u
e
;
an
d
b
o
u
n
d
s
,
b
o
th
u
p
p
e
r
,
lo
wer
,
f
o
r
a
d
-
d
im
e
n
s
io
n
al
p
r
o
b
lem
.
‒
Step
2
:
r
an
d
o
m
g
en
er
atio
n
.
T
h
e
in
p
u
t
weig
h
t
p
ar
am
ete
r
s
ar
e
g
en
er
ated
r
an
d
o
m
ly
f
o
llo
win
g
in
itializatio
n
u
s
in
g
th
e
AHBOA a
p
p
r
o
ac
h
.
‒
Step
3
:
f
itn
ess
f
u
n
ctio
n
.
T
h
e
in
p
u
t
weig
h
t
p
ar
a
m
eter
s
ar
e
g
en
er
ated
r
a
n
d
o
m
ly
f
o
llo
win
g
in
itializatio
n
u
s
in
g
th
e
AHBOA a
p
p
r
o
ac
h
.
=
[
]
(
1
6
)
‒
Step
4
:
ex
p
lo
r
atio
n
p
h
ase
.
T
h
e
th
r
ee
f
ly
in
g
s
k
ills
o
m
n
id
ir
ec
tio
n
al,
d
iag
o
n
al
an
d
ax
ial
f
lig
h
ts
th
at
ar
e
ad
eq
u
ately
u
tili
ze
d
d
u
r
in
g
f
o
r
ag
in
g
a
r
e
r
e
p
r
esen
ted
b
y
a
d
i
r
ec
tio
n
s
witch
v
ec
to
r
in
th
e
AHA
m
eth
o
d
.
I
n
d
-
d
im
e
n
s
io
n
s
p
ac
e,
th
is
v
ec
to
r
d
eter
m
in
es
wh
eth
er
o
n
e
o
r
m
o
r
e
d
ir
ec
tio
n
s
ex
is
t.
T
h
e
ex
p
lo
r
in
g
p
h
ase
is
th
en
g
iv
en
b
y
(
1
7
)
.
(
+
1
)
=
,
(
)
+
.
.
(
(
)
−
,
(
)
)
(
1
7
)
W
h
er
e
,
(
)
is
wh
er
e
th
e
i
th
h
u
m
m
in
g
b
ir
d
p
lan
s
to
v
is
it
wh
en
it
c
o
m
es
to
f
in
d
in
g
f
o
o
d
;
(
+
1
)
is
wh
er
e
i
th
f
o
o
d
s
u
p
p
ly
at
tim
e
T
;
d
en
o
ted
as
th
e
d
ir
ec
ted
f
ac
t
o
r
an
is
d
is
tr
ib
u
ted
n
o
r
m
ally
ac
co
r
d
in
g
to
N(
0
,
1
)
;
an
d
H
d
en
o
ted
as
th
e
p
o
ten
tial
f
o
o
d
s
u
p
p
ly
h
as
a
h
ig
h
er
r
ate
o
f
n
ec
tar
r
e
p
len
is
h
m
en
t
th
an
th
e
o
n
e
th
at
ex
is
ts
n
o
w.
‒
Step
5
:
e
x
p
lo
itatio
n
p
h
ase
f
o
r
o
p
tim
izin
g
.
Af
ter
v
is
itin
g
i
ts
tar
g
et
f
o
o
d
s
o
u
r
ce
an
d
co
n
s
u
m
in
g
th
e
n
ec
tar
f
r
o
m
f
lo
wer
s
,
a
h
u
m
m
i
n
g
b
ir
d
s
ea
r
ch
es
f
o
r
a
n
ew
f
o
o
d
s
o
u
r
ce
r
ath
er
th
an
g
o
i
n
g
t
o
an
o
th
er
.
Af
te
r
th
at,
it
tak
es
o
f
f
f
o
r
a
n
ea
r
b
y
ar
ea
with
in
its
o
wn
r
an
g
e,
wh
er
e
it
m
ig
h
t
d
is
co
v
er
a
f
r
esh
f
o
o
d
s
o
u
r
ce
o
r
s
o
m
eth
in
g
b
etter
t
h
an
wh
at
it
n
o
w
h
as.
I
n
(
1
8
)
th
en
p
r
o
v
id
es
th
e
ex
p
lo
itatio
n
p
h
ase.
(
+
1
)
=
(
)
+
.
.
(
)
(
1
8
)
W
h
er
e
(
+
1
)
d
en
o
tes
i
th
f
o
o
d
s
u
p
p
ly
at
tim
e
T
;
d
en
o
ted
as
th
e
d
ir
ec
ted
f
ac
to
r
an
is
d
is
tr
ib
u
ted
n
o
r
m
ally
ac
co
r
d
in
g
to
N(
0
,
1
)
;
H
d
en
o
te
d
as
th
e
p
o
ten
tial
f
o
o
d
s
u
p
p
ly
h
as
a
h
ig
h
er
r
a
te
o
f
n
ec
tar
r
ep
len
is
h
m
en
t
th
an
th
e
o
n
e
th
at
ex
is
ts
n
o
w
;
an
d
is
a
ter
r
ito
r
ial
co
m
p
o
n
en
t
th
at
h
as
a
m
ea
n
o
f
0
an
d
d
is
tr
ib
u
ted
ac
co
r
d
in
g
to
th
e
n
o
r
m
al
d
is
tr
ib
u
tio
n
N
0
,
1
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
1
5
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
6
:
429
-
4
4
2
436
‒
Step
6
:
t
er
m
in
ati
o
n
.
T
h
e
wei
g
h
t
p
a
r
am
eter
o
f
g
en
er
at
o
r
f
r
o
m
atten
tio
n
in
d
u
ce
d
m
u
lti
h
ea
d
C
NN
is
o
p
tim
iz
ed
b
y
u
tili
zin
g
AHBOA
;
an
d
it
wil
l
r
ep
ea
t
s
tep
3
u
n
til
it
o
b
tain
s
it
s
h
altin
g
cr
iter
ia
=
+
1
.
T
h
en
W
DC
GAN
-
AH
B
OA
-
S
C
B
T
D
m
eth
o
d
ef
f
ec
tiv
ely
class
if
ies
th
e
b
r
ain
tu
m
o
r
lik
e
g
lio
m
a,
m
en
in
g
io
m
a
,
p
itu
itar
y
,
an
d
n
o
tu
m
o
r
.
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
e
u
n
t
r
i
e
d
r
e
s
u
l
t
s
o
f
w
is
h
ed
-
f
o
r
W
DC
G
AN
-
A
HB
OA
-
SC
B
T
a
r
e
c
o
n
f
e
r
r
e
d
i
n
t
h
is
s
e
c
t
o
r
.
T
h
e
s
i
m
u
l
at
i
o
n
i
s
e
x
e
c
u
t
e
d
i
n
P
y
t
h
o
n
u
s
i
n
g
P
C
d
o
n
e
I
n
t
e
l
c
o
r
e
i
5
,
8
G
B
R
A
M
,
2
.
5
0
G
H
z
C
PU
,
W
i
n
d
o
ws
7
u
t
i
l
i
z
i
n
g
B
r
a
T
S
d
at
a
s
et
.
O
b
t
ai
n
e
d
o
u
t
c
o
m
e
o
f
t
h
e
p
r
o
p
o
s
e
d
W
D
C
GA
N
-
A
HB
OA
-
SC
B
T
a
p
p
r
o
a
c
h
i
s
a
n
a
l
y
z
e
d
wi
t
h
e
x
i
s
ti
n
g
3
D
-
U
Ne
t
-
D
NN
-
SC
B
T
[
1
6
]
,
C
NN
-
l
o
n
g
s
h
o
r
t
-
t
e
r
m
m
e
m
o
r
y
(
L
S
T
M
)
-
SC
B
T
[
1
7
]
,
a
n
d
D
N
N
–
S
C
B
T
[
1
8
]
s
y
s
tem
s
.
Fig
u
r
e
2
d
is
p
lay
s
o
u
t
p
u
t o
f
p
r
o
p
o
s
ed
W
DC
GAN
-
A
HB
OA
-
S
C
B
T
s
y
s
tem
.
Fig
u
r
e
2
.
Ou
t
p
u
t r
esu
lt o
f
p
r
o
jecte
d
W
DC
GA
N
-
AHB
OA
-
S
C
B
T
m
eth
o
d
4
.
1
.
P
er
f
o
r
m
a
nce
m
e
a
s
ures
T
h
e
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
an
d
m
ea
n
s
q
u
a
r
e
e
r
r
o
r
(
MSE
)
p
er
f
o
r
m
an
ce
m
ea
s
u
r
es
ar
e
u
s
ed
t
o
an
aly
ze
th
e
d
esire
d
s
lan
t'
s
p
er
f
o
r
m
an
ce
.
T
h
e
f
o
llo
win
g
p
ar
am
eter
s
ar
e
co
m
p
u
te
d
u
s
in
g
th
e
s
u
g
g
ested
ap
p
r
o
ac
h
to
o
b
tain
th
e
d
if
f
er
en
t
v
alu
es,
an
d
th
eir
s
ig
n
if
ican
ce
is
em
p
h
asized
b
y
c
o
m
p
ar
is
o
n
with
th
e
cu
r
r
en
t m
eth
o
d
s
.
‒
A
c
c
u
r
a
c
y
:
a
c
c
u
r
a
c
y
d
e
f
i
n
e
s
d
e
t
e
c
t
i
o
n
r
a
t
e
,
w
h
i
c
h
a
r
e
c
o
r
r
e
c
t
l
y
c
a
t
e
g
o
r
i
z
e
d
.
T
h
e
n
f
o
r
m
u
l
a
i
s
d
e
r
i
v
e
d
i
n
(
1
9
)
.
=
(
+
)
(
+
+
+
)
(
1
9
)
‒
Pre
cisi
o
n
:
p
r
ec
is
io
n
co
m
p
u
tes
th
e
co
u
n
t
o
f
tr
u
e
p
o
s
itiv
es
d
iv
id
ed
t
h
r
o
u
g
h
tr
u
e
p
o
s
itiv
es
p
lu
s
n
u
m
b
er
,
f
alse p
o
s
itiv
e
n
u
m
b
e
r
. I
t is g
iv
en
b
y
(
2
0
)
.
=
(
+
)
(
2
0
)
‒
MSE
:
it
is
m
ea
s
u
r
ed
b
y
MSE
,
a
r
is
k
f
u
n
ctio
n
in
s
tatis
tics
.
I
f
y
o
u
wis
h
to
p
en
alize
h
u
g
e
er
r
o
r
s
m
o
r
e
th
a
n
litt
le
o
n
es
in
r
eg
r
ess
io
n
an
d
y
o
u
t
h
in
k
y
o
u
r
o
b
jectiv
e
is
n
o
r
m
ally
d
is
tr
ib
u
te
d
,
u
tili
ze
MSE
.
T
h
en
t
h
e
MSE
is
d
er
iv
ed
in
(
2
1
)
.
MSE
=
1
∑
(
−
̂
)
=
1
(
2
1
)
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
B
r
a
in
tu
mo
r
s
eg
men
ta
tio
n
a
n
d
cla
s
s
ifica
tio
n
u
s
in
g
a
r
tifi
cia
l h
u
mmin
g
b
ir
d
…
(
R
a
d
h
a
kris
h
n
a
n
K
a
r
th
ikey
a
n
)
437
4
.
2
.
P
er
f
o
r
m
a
nce
a
na
ly
s
is
Fig
u
r
es
3
to
5
d
ep
icts
s
im
u
la
tio
n
o
u
tco
m
e
W
DC
GAN
-
AH
B
OA
-
S
C
B
T
m
eth
o
d
.
T
h
e
p
er
f
o
r
m
an
ce
m
etr
ics
ar
e
co
m
b
in
ed
with
e
x
is
tin
g
3
D
-
UNet
-
DNN
-
SC
B
T
,
C
NN
-
L
STM
-
SC
B
T
,
an
d
D
NN
-
SC
B
T
m
eth
o
d
s
.
Acc
u
r
ac
y
a
n
aly
s
is
is
s
h
o
wn
i
n
Fig
u
r
e
3
.
T
h
e
p
r
o
p
o
s
ed
W
DC
GAN
-
AH
B
OA
-
S
C
B
T
attain
s
3
1
.
8
2
,
3
2
.
0
4
,
a
n
d
3
2
.
3
6
%
g
r
ea
ter
ac
cu
r
ac
y
f
o
r
b
r
ain
tu
m
o
r
im
ag
e
w
h
ich
is
ex
am
in
ed
co
m
p
ar
e
d
with
ex
is
tin
g
3
D
-
UNet
-
DNN
-
SC
B
T
,
C
NN
-
L
STM
-
S
C
B
T
,
an
d
DNN
–
SC
B
T
m
eth
o
d
s
.
Fig
u
r
e
4
d
ep
icts
th
e
p
er
f
o
r
m
an
ce
an
aly
s
is
o
f
p
r
o
p
o
s
ed
m
eth
o
d
o
lo
g
y
.
T
h
e
p
r
o
p
o
s
ed
W
DC
GAN
-
AH
B
OA
-
SC
B
T
attain
s
3
1
.
1
8
,
3
2
.
7
5
,
an
d
3
2
.
9
0
%
g
r
ea
ter
f
o
r
p
r
ec
is
io
n
;
attain
s
2
9
.
1
8
,
3
0
.
5
1
,
an
d
3
1
.
7
5
%
g
r
ea
ter
f
o
r
s
en
s
itiv
ity
;
attain
s
3
2
.
5
3
,
3
2
.
7
1
,
an
d
3
2
.
8
1
%
g
r
ea
ter
f
o
r
r
ec
all
;
attain
s
3
0
.
1
5
,
3
1
.
5
0
,
a
n
d
3
2
.
9
1
%
h
i
g
h
er
f
o
r
s
en
s
itiv
ity
;
an
d
attain
s
2
8
.
9
5
,
3
1
.
2
0
,
an
d
3
2
.
4
8
% h
ig
h
er
f
o
r
F1
-
s
co
r
e
[
2
8
]
wh
ich
is
an
aly
ze
d
with
ex
is
tin
g
3
D
-
UNet
-
DNN
-
SC
B
T
,
C
NN
-
L
STM
-
SC
B
T
,
an
d
DNN
–
SC
B
T
m
eth
o
d
s
.
An
aly
s
is
o
f
MSE
is
s
h
o
wn
in
Fig
u
r
e
5
.
T
h
e
p
r
o
p
o
s
ed
W
DC
GA
N
-
AHBOA
-
S
C
B
T
attain
s
2
8
.
7
2
,
3
0
.
9
4
,
an
d
3
1
.
0
4
%
lo
wer
MSE
f
o
r
g
lio
m
a
;
attain
s
2
9
.
6
8
,
3
0
.
7
1
,
an
d
3
1
.
9
2
%
lo
wer
MSE
f
o
r
m
en
in
g
io
m
a
; a
ttain
s
3
1
.
5
3
,
3
2
.
8
3
,
an
d
3
2
.
1
1
%
lo
wer
MSE
f
o
r
p
itu
itar
y
;
an
d
attain
s
2
9
.
7
5
,
3
0
.
8
0
,
an
d
3
1
.
1
5
%
lo
wer
MSE
f
o
r
n
o
tu
m
o
r
im
a
g
e
wh
ich
is
an
aly
ze
d
with
ex
i
s
tin
g
3
D
-
UNet
-
DNN
-
SC
B
T
,
C
NN
-
L
STM
-
S
C
B
T
an
d
DNN
–
SC
B
T
m
eth
o
d
s
.
Fig
u
r
e
3
.
Acc
u
r
ac
y
a
n
aly
s
is
Fig
u
r
e
4
.
Per
f
o
r
m
an
c
e
an
aly
s
i
s
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
.
1
,
Feb
r
u
ar
y
2
0
2
6
:
429
-
4
4
2
438
Fig
u
r
e
5
.
Me
an
s
q
u
ar
e
e
r
r
o
r
a
n
aly
s
is
T
ab
le
2
p
r
esen
ts
s
o
m
e
co
m
p
ar
ativ
e
an
aly
s
es
d
o
n
e
a
m
o
n
g
v
ar
io
u
s
m
eth
o
d
s
f
o
r
s
eg
m
en
tin
g
b
r
ain
tu
m
o
r
s
an
d
ca
teg
o
r
izatio
n
,
s
h
o
win
g
s
tr
ik
in
g
v
ar
ia
n
ce
in
ce
r
tain
p
er
f
o
r
m
an
ce
m
etr
ics
s
u
ch
as
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
k
ap
p
a
s
tatis
tic
(
K
ap
p
a)
,
Ma
tth
ew
co
r
r
elatio
n
co
ef
f
icien
t
(
MCC
)
,
av
er
ag
e
Hau
s
d
o
r
f
f
d
is
tan
ce
(
AHD)
,
p
r
o
ce
s
s
in
g
tim
e
,
er
r
o
r
r
ate,
in
ter
s
ec
tio
n
o
v
er
u
n
i
o
n
(
I
o
U)
,
d
ice
s
im
ilar
ity
co
e
f
f
icien
t
,
s
p
ec
if
icity
s
co
r
e
,
an
d
Hau
s
d
o
r
f
f
d
is
tan
ce
.
T
h
e
C
NN
ap
p
r
o
ac
h
h
ad
an
a
cc
u
r
ac
y
o
f
8
9
.
5
7
%
with
p
r
ec
is
io
n
at
9
0
.
8
6
%
an
d
a
K
ap
p
a
v
alu
e
o
f
9
0
.
4
5
,
d
ep
i
ctin
g
g
o
o
d
p
er
f
o
r
m
a
n
ce
b
u
t
with
r
elativ
ely
h
ig
h
p
r
o
ce
s
s
in
g
tim
e
at
0
.
7
a
n
d
an
er
r
o
r
r
ate
o
f
0
.
6
.
Am
o
n
g
th
e
m
,
th
e
HABW
MF
O
m
eth
o
d
h
ad
a
lo
wer
ac
cu
r
ac
y
o
f
7
8
.
5
4
%
b
u
t
a
v
er
y
h
ig
h
p
r
ec
is
io
n
at
8
9
.
0
7
%.
On
th
e
o
th
er
h
an
d
,
th
e
DNN
ap
p
r
o
ac
h
ac
h
iev
ed
a
h
ig
h
er
ac
cu
r
ac
y
o
f
9
0
.
6
5
%
at
p
r
ec
is
io
n
8
8
.
9
6
%,
r
e
f
lectin
g
it
s
v
er
y
s
tr
o
n
g
p
er
f
o
r
m
a
n
ce
in
v
ar
io
u
s
m
etr
ics s
u
ch
as M
C
C
a
n
d
I
o
U.
T
ab
le
2
.
Per
f
o
r
m
an
ce
co
m
p
a
r
is
o
n
o
f
b
r
ain
tu
m
o
r
s
eg
m
e
n
tatio
n
m
eth
o
d
s
M
e
t
h
o
d
s
A
c
c
u
r
a
c
y
(
%)
P
r
e
c
i
s
i
o
n
(
%)
K
a
p
p
a
st
a
t
i
st
i
c
M
C
C
AHD
P
r
o
c
e
ss
i
n
g
t
i
m
e
Er
r
o
r
r
a
t
e
I
o
U
D
i
c
e
si
mi
l
a
r
i
t
y
S
p
e
c
i
f
i
c
i
t
y
sco
r
e
H
a
u
s
d
o
r
f
f
d
i
s
t
a
n
c
e
C
N
N
[
1
6
]
8
9
.
5
7
9
0
.
8
6
9
0
.
4
5
8
8
.
7
8
7
9
.
5
6
0
.
7
0
.
6
8
8
.
9
5
7
5
.
4
6
5
.
4
76
H
A
B
W
M
F
O
[
1
7
]
7
8
.
5
4
8
9
.
0
7
8
8
.
9
7
8
9
.
4
7
7
9
.
9
7
0
.
6
0
.
4
8
9
.
7
5
5
8
.
7
6
7
7
.
9
7
6
73
DNN
[
1
8
]
9
0
.
6
5
8
8
.
9
6
9
0
.
5
2
8
9
.
6
3
9
0
.
8
2
0
.
5
0
.
6
8
9
.
6
2
6
6
.
5
7
8
5
.
2
4
77
D
TC
G
W
T
[
1
9
]
9
0
.
4
8
8
9
.
7
4
7
8
.
5
5
8
9
.
4
2
9
3
.
5
1
0
.
4
0
.
5
8
9
.
7
2
7
5
.
4
2
6
5
.
3
76
D
C
N
N
[
2
0
]
9
0
.
2
3
8
9
.
9
8
9
0
.
5
3
9
2
.
8
2
9
0
.
9
1
0
.
6
0
.
6
9
2
.
8
2
6
7
.
8
2
7
4
.
8
1
69
S
V
M
-
A
N
F
I
S
[
2
1
]
9
0
.
8
1
8
8
.
6
2
9
0
.
8
3
8
9
.
8
2
9
0
.
7
9
0
.
5
0
.
4
8
9
.
7
2
6
6
.
4
8
5
.
6
76
W
D
C
G
A
N
-
A
H
B
O
A
-
S
C
B
T
(
p
r
o
p
o
s
e
d
)
9
9
.
9
7
9
8
.
8
2
9
9
.
8
2
9
9
.
8
3
9
9
.
8
9
0
.
1
0
.
1
9
9
.
8
3
9
5
.
8
2
9
6
.
6
25
T
h
e
SVM
-
ANFI
S d
em
o
n
s
tr
ated
th
e
ef
f
ec
tiv
en
ess
o
f
th
e
co
m
b
in
atio
n
o
f
m
ac
h
in
e
lea
r
n
in
g
with
f
u
zz
y
lo
g
ic
an
d
th
e
ac
c
u
r
ac
y
o
f
9
0
.
8
1
%
an
d
p
r
ec
is
io
n
o
f
8
8
.
6
2
%
.
T
h
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
,
W
DC
GAN
-
AH
B
OA
-
SC
B
T
,
was
f
ar
m
u
c
h
b
etter
w
ith
an
ac
c
u
r
ac
y
o
f
9
9
.
9
7
%
,
p
r
ec
is
io
n
o
f
9
8
.
8
2
%
,
Kap
p
a
o
f
9
9
.
8
2
,
an
d
MCC
o
f
9
9
.
8
3
at
th
e
e
x
p
en
s
e
o
f
all
o
t
h
er
m
eth
o
d
s
.
T
h
e
ap
p
r
o
ac
h
al
s
o
p
r
esen
ted
im
p
r
o
v
e
d
v
alu
es
o
f
AHD,
I
o
U,
an
d
d
ice
s
im
ilar
ity
co
ef
f
icien
t
wi
th
th
e
least
v
alu
es
o
f
p
r
o
ce
s
s
in
g
tim
e
an
d
e
r
r
o
r
r
ate
at
0
.
1
an
d
v
e
r
y
lo
w
Hau
s
d
o
r
f
f
d
is
tan
ce
o
f
2
5
.
I
t
i
s
th
is
in
-
d
ep
th
co
m
p
ar
is
o
n
th
at
p
r
o
v
id
es
r
ea
s
o
n
to
b
eliev
e
th
at
th
e
s
u
g
g
ested
ap
p
r
o
ac
h
is
m
u
ch
s
u
p
er
io
r
to
th
e
cu
r
r
en
t
o
n
es
in
te
r
m
s
o
f
en
h
an
ce
m
en
t
o
f
th
e
s
eg
m
en
tatio
n
an
d
ca
teg
o
r
izatio
n
o
f
ac
cu
r
ac
y
,
ef
f
icien
cy
,
an
d
r
eliab
ilit
y
.
T
ab
le
3
co
n
tr
asts
th
e
co
m
p
u
tatio
n
al
co
s
t
an
d
c
o
m
p
lex
ity
o
f
m
an
y
tech
n
i
q
u
es
f
o
r
s
eg
m
en
tin
g
an
d
class
if
y
in
g
b
r
ain
tu
m
o
r
s
.
T
h
e
co
m
p
u
tin
g
d
em
an
d
s
v
ar
y
f
o
r
e
ac
h
ap
p
r
o
ac
h
,
t
h
u
s
m
ak
in
g
a
b
ig
d
if
f
er
e
n
ce
in
th
e
ap
p
licab
ilit
y
an
d
p
e
r
f
o
r
m
an
ce
o
f
an
ap
p
licab
le
tech
n
iq
u
e.
T
h
e
C
NN
ap
p
r
o
ac
h
d
is
p
lay
s
a
m
o
d
est
ef
f
icien
cy
i
n
b
o
th
co
m
p
u
tatio
n
al
c
o
s
t
an
d
co
m
p
lex
ity
,
with
a
v
alu
e
o
f
0
.
5
a
n
d
0
.
7
,
r
esp
ec
tiv
ely
.
T
h
e
HABW
MF
O
m
eth
o
d
o
l
o
g
y
h
as
a
lo
wer
lev
e
l
o
f
co
m
p
lex
ity
,
s
tan
d
in
g
at
0
.
6
,
y
et
h
ig
h
er
o
n
c
o
m
p
u
tatio
n
al
co
s
t,
s
tan
d
in
g
at
0
.
7
,
wh
ic
h
m
ea
n
s
th
is
tech
n
iq
u
e
is
less
co
m
p
licated
t
h
an
s
o
m
e
o
th
er
tech
n
iq
u
es
b
u
t
r
eq
u
ir
es
m
o
r
e
co
m
p
u
tatio
n
al
r
eso
u
r
ce
s
.
T
h
e
DNN
ap
p
r
o
ac
h
is
r
eso
u
r
ce
-
in
ten
s
iv
e
b
u
t
h
as
a
co
m
p
ar
ativ
ely
s
im
p
ler
co
m
p
u
tatio
n
al
s
tr
u
ctu
r
e,
as
in
f
er
r
ed
f
r
o
m
its
co
m
p
u
tatio
n
al
c
o
s
t
o
f
0
.
7
a
n
d
co
m
p
le
x
ity
o
f
0
.
5
.
T
h
e
DT
C
GW
T
0
2
4
6
8
10
12
Gli
o
m
a
Me
n
in
g
io
m
a
Pi
tu
i
tar
y
No
T
u
m
o
r
MSE
Dif
f
er
en
t
b
r
ain
tu
m
o
r
im
ag
es
Pro
p
o
s
ed
E
x
is
ti
n
g
1
E
x
is
ti
n
g
2
E
x
is
ti
n
g
3
Evaluation Warning : The document was created with Spire.PDF for Python.