I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
pu
t
er
E
ng
ineering
(
I
J
E
CE
)
Vo
l.
14
,
No
.
6
,
Dec
em
b
er
20
24
,
p
p
.
6
3
3
9
~
6
3
5
0
I
SS
N:
2088
-
8
7
0
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijece.
v
14
i
6
.
pp
6
3
3
9
-
6
3
5
0
6339
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
Perf
o
rma
nce
ana
ly
sis
of hybrid bio
-
inspi
red
alg
o
rithms for
cla
ss
ifying
brain t
umo
rs in imba
la
n
ced
ma
g
netic
reso
na
nce
ima
g
ing
da
tas
ets
Ra
hu
l R
a
m
esh
Cha
k
re
1
,
Arc
ha
na
S
.
Va
id
y
a
2
,
Dipa
k
V.
P
a
t
il
2
1
S
c
h
o
o
l
o
f
C
o
m
p
u
t
a
t
i
o
n
a
l
S
c
i
e
n
c
e
s,
F
a
c
u
l
t
y
o
f
S
c
i
e
n
c
e
a
n
d
Te
c
h
n
o
l
o
g
y
,
JS
P
M
U
n
i
v
e
r
si
t
y
P
u
n
e
,
P
u
n
e
,
I
n
d
i
a
2
G
ES’
s R
.
H
.
S
a
p
a
t
C
o
l
l
e
g
e
o
f
E
n
g
i
n
e
e
r
i
n
g
,
M
a
n
a
g
e
me
n
t
S
t
u
d
i
e
s a
n
d
R
e
se
a
r
c
h
,
N
a
s
h
i
k
,
S
a
v
i
t
r
i
b
a
i
P
h
u
l
e
P
u
n
e
U
n
i
v
e
r
si
t
y
,
P
u
n
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
Sep
9
,
2
0
2
3
R
ev
is
ed
J
u
l 1
6
,
2
0
2
4
Acc
ep
ted
Au
g
6
,
2
0
2
4
M
a
g
n
e
ti
c
re
so
n
a
n
c
e
ima
g
in
g
(M
RI)
is
a
su
b
sta
n
ti
a
l
ima
g
i
n
g
p
ro
c
e
d
u
re
fo
r
d
iag
n
o
sin
g
b
ra
in
t
u
m
o
rs.
Ho
we
v
e
r,
b
ra
in
tu
m
o
r
c
las
sifica
ti
o
n
c
o
n
ti
n
u
e
s
c
h
a
ll
e
n
g
i
n
g
d
u
e
to
t
h
e
u
n
e
q
u
a
l
d
istri
b
u
ti
o
n
o
f
c
las
se
s
with
in
d
a
tas
e
ts,
c
o
m
p
li
c
a
ti
n
g
p
re
c
ise
d
ia
g
n
o
sis
a
n
d
c
las
sifica
ti
o
n
.
Th
is
re
se
a
rc
h
fo
c
u
se
s
o
n
th
e
c
las
s
imb
a
lan
c
e
in
m
e
d
ica
l
ima
g
e
d
a
tas
e
ts
b
y
p
r
o
p
o
sin
g
a
h
y
b
ri
d
b
i
o
-
in
sp
ire
d
a
lg
o
rit
h
m
f
o
r
b
ra
in
t
u
m
o
r
c
las
sifica
ti
o
n
.
A
r
id
e
r
o
p
ti
m
iza
ti
o
n
a
n
d
p
a
rti
c
le
rid
e
r
m
u
t
u
a
l
i
n
fo
rm
a
ti
o
n
-
b
a
se
d
d
e
n
d
rit
ic
-
sq
u
irrel
se
a
rc
h
a
l
g
o
rit
h
m
c
o
m
b
i
n
e
d
with
a
n
a
rti
fic
ial
imm
u
n
e
c
las
sif
ier
is
d
e
v
e
lo
p
e
d
a
n
d
tes
ted
o
n
i
m
b
a
lan
c
e
d
d
a
tas
e
ts
g
e
n
e
ra
ted
fro
m
BRATS
a
n
d
S
imB
RATS
.
Ex
p
e
rime
n
tal
o
u
t
c
o
m
e
s
a
re
e
n
c
o
u
ra
g
i
n
g
,
F
o
r
th
e
imb
a
lan
c
e
d
BRATS
d
a
tas
e
t,
t
h
e
rid
e
r
o
p
t
imiz
a
ti
o
n
-
b
a
se
d
c
las
sifier
a
c
h
iev
e
d
a
n
a
c
c
u
ra
c
y
o
f
9
4
.
8
4
%
,
se
n
siti
v
it
y
o
f
9
2
.
9
6
%
,
a
n
d
sp
e
c
ifi
c
it
y
o
f
9
4
.
9
5
%
.
Th
e
p
a
rti
c
l
e
rid
e
r
m
u
tu
a
l
in
f
o
rm
a
ti
o
n
-
b
a
se
d
c
las
sifier
o
u
t
p
e
rfo
rm
e
d
o
t
h
e
rs
with
9
6
.
2
5
%
a
c
c
u
ra
c
y
,
9
4
.
3
3
%
se
n
siti
v
it
y
,
a
n
d
9
4
.
8
5
%
sp
e
c
ifi
c
it
y
.
F
o
r
th
e
imb
a
lan
c
e
d
S
imB
RATS
d
a
tas
e
t,
th
e
rid
e
r
o
p
ti
m
iza
ti
o
n
-
b
a
se
d
c
las
si
fier
a
c
h
iev
e
d
9
4
.
9
5
%
a
c
c
u
ra
c
y
,
9
2
.
0
5
%
se
n
siti
v
it
y
,
a
n
d
9
4
.
0
4
%
sp
e
c
ifi
c
it
y
.
T
h
e
p
a
rti
c
le
rid
e
r
m
u
tu
a
l
i
n
fo
rm
a
ti
o
n
-
b
a
se
d
c
las
sifier
e
x
c
e
ll
e
d
with
9
6
.
3
5
%
a
c
c
u
ra
c
y
,
9
4
.
4
2
%
se
n
siti
v
it
y
,
a
n
d
9
5
.
4
4
%
s
p
e
c
ifi
c
it
y
.
Th
e
se
fin
d
i
n
g
s
su
g
g
e
st
t
h
a
t
th
e
p
r
o
p
o
se
d
a
lg
o
rit
h
m
e
ffe
c
ti
v
e
ly
a
d
d
re
ss
e
s
c
la
ss
imb
a
lan
c
e
in
m
e
d
ica
l
ima
g
e
d
a
t
a
se
ts,
o
ffe
rin
g
a
r
o
b
u
st
so
lu
ti
o
n
fo
r
b
ra
i
n
tu
m
o
r
c
las
sifica
ti
o
n
.
T
h
e
p
a
rti
c
l
e
rid
e
r
m
u
tu
a
l
in
f
o
rm
a
ti
o
n
-
b
a
se
d
c
las
sifier
sh
o
ws
p
r
o
m
ise
fo
r
e
n
h
a
n
c
in
g
d
iag
n
o
stic
a
c
c
u
ra
c
y
in
c
li
n
ica
l
se
tt
in
g
s,
d
e
m
o
n
st
ra
ti
n
g
t
h
e
e
ffica
c
y
o
f
h
y
b
rid
ize
d
b
io
-
i
n
sp
ire
d
a
lg
o
rit
h
m
s
in
m
a
n
a
g
in
g
imb
a
lan
c
e
d
d
a
tas
e
ts.
K
ey
w
o
r
d
s
:
B
io
-
in
s
p
ir
ed
co
m
p
u
tin
g
B
r
ain
tu
m
o
r
class
if
icatio
n
I
m
b
alan
ce
d
m
ed
ical
im
ag
e
d
ataset
I
m
m
u
n
e
c
o
m
p
u
tin
g
Swar
m
in
tellig
en
ce
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
:
R
ah
u
l Ram
esh
C
h
ak
r
e
Sch
o
o
l o
f
C
o
m
p
u
tatio
n
al
Scie
n
ce
s
,
Facu
lty
o
f
Scien
ce
a
n
d
T
ec
h
n
o
lo
g
y
,
J
SP
M
Un
iv
er
s
ity
Pu
n
e
Pu
n
e,
Ma
h
ar
ash
tr
a,
I
n
d
ia
E
m
ail:
r
ah
u
lr
ch
a
k
r
e@
g
m
ail.
c
o
m
1.
I
NT
RO
D
UCT
I
O
N
Du
e
to
th
e
u
s
e
o
f
v
ar
io
u
s
im
ag
in
g
m
o
d
alities
in
th
e
m
ed
i
ca
l
f
ield
,
th
e
a
m
o
u
n
t
o
f
m
ed
i
ca
l
im
ag
e
d
atasets
i
s
in
cr
ea
s
in
g
r
ap
id
ly
[
1
]
.
Me
d
ical
im
ag
e
an
aly
s
is
aid
s
in
d
iag
n
o
s
in
g
v
ar
i
o
u
s
illn
ess
es.
Du
e
to
it
s
s
ize
,
s
tr
u
ctu
r
e,
an
d
s
h
ap
e,
m
u
ltid
im
en
s
io
n
al
m
ed
ical
im
ag
e
an
a
ly
s
is
b
ec
o
m
es
a
cr
u
cial
ta
s
k
.
T
h
e
m
o
s
t
co
m
m
o
n
r
ea
s
o
n
f
o
r
ca
n
ce
r
-
r
elate
d
m
o
r
tality
is
a
b
r
ain
tu
m
o
r
.
E
ar
ly
d
iag
n
o
s
is
o
f
t
h
is
co
n
d
itio
n
g
r
ea
tly
in
cr
ea
s
es
th
e
lik
elih
o
o
d
th
at
t
h
e
p
atien
t
will
r
ec
eiv
e
a
s
u
cc
ess
f
u
l
co
u
r
s
e
o
f
th
er
ap
y
an
d
lo
wer
s
th
e
r
is
k
o
f
life
.
Ar
tific
ial
in
tellig
en
ce
an
d
v
a
r
io
u
s
o
p
tim
izatio
n
tech
n
iq
u
es
ar
e
u
s
ed
to
ex
tr
ac
t
u
s
ef
u
l
k
n
o
wled
g
e
f
r
o
m
th
e
lar
g
e
m
e
d
ical
im
ag
e
d
ataset
in
a
way
th
at
m
ed
ical
p
r
ac
titi
o
n
er
s
will
g
et
h
elp
to
d
iag
n
o
s
e
th
e
p
atien
ts
[
2
]
,
[
3
]
.
T
h
e
ty
p
ical
class
if
icatio
n
tech
n
iq
u
es
p
r
e
s
u
p
p
o
s
e
th
at
th
e
d
ataset
o
f
m
ed
ical
im
a
g
es
co
n
tain
s
s
am
p
les
with
a
f
air
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
e
r
20
24
:
6
3
3
9
-
6
3
5
0
6340
d
is
tr
ib
u
tio
n
.
T
h
e
ter
m
“
im
b
ala
n
ce
d
d
ataset
”
r
ef
er
s
to
s
itu
atio
n
s
in
wh
ich
th
e
d
is
tr
ib
u
tio
n
is
n
o
t
f
air
.
T
h
is
ty
p
e
o
f
d
ataset
will
ca
u
s
e
b
iased
lear
n
in
g
,
wh
ic
h
r
esu
lts
in
an
in
e
f
f
ec
tiv
e
m
o
d
el
[
4
]
,
[
5
]
.
T
h
e
i
m
b
alan
ce
d
d
ataset
is
n
o
th
in
g
m
o
r
e
t
h
an
a
n
u
n
ev
e
n
d
is
tr
ib
u
tio
n
o
f
s
am
p
les
ac
r
o
s
s
class
es
[
6
]
.
I
n
th
is
,
t
h
er
e
a
r
e
m
o
r
e
s
am
p
les o
f
o
n
e
class
th
an
an
o
th
er
.
I
n
th
e
en
d
,
th
e
class
w
ith
th
e
m
o
s
t
s
am
p
l
es
will
b
e
th
e
m
ajo
r
clas
s
,
an
d
th
e
class
w
ith
th
e
f
ewe
s
t
s
am
p
les
will
b
e
th
e
s
ec
o
n
d
ar
y
class
.
I
n
m
a
n
y
r
ea
l
-
wo
r
ld
ap
p
licatio
n
s
,
i
n
clu
d
in
g
d
et
ec
tin
g
f
r
a
u
d
,
b
r
ain
tu
m
o
r
s
,
in
tr
u
s
io
n
d
etec
tio
n
s
y
s
tem
s
,
an
d
en
tity
r
eso
lu
tio
n
,
t
h
e
im
b
alan
ce
d
d
is
tr
ib
u
tio
n
p
r
o
b
lem
is
p
r
esen
t
[
7
]
,
[
8
]
.
E
f
f
ec
tiv
e
f
ea
tu
r
e
ex
tr
ac
ti
o
n
,
f
ea
tu
r
e
s
elec
tio
n
,
an
d
f
e
atu
r
e
class
if
icatio
n
ap
p
r
o
ac
h
es
ar
e
r
eq
u
ir
ed
f
o
r
co
m
p
u
ter
-
aid
ed
d
iag
n
o
s
is
em
p
lo
y
in
g
m
u
ltid
im
en
s
io
n
al
im
ag
es.
Du
e
to
th
e
lo
w
n
u
m
b
er
o
f
im
p
o
r
tan
t
d
ata
in
s
tan
ce
s
,
im
b
ala
n
c
ed
d
ata
d
is
tr
ib
u
tio
n
p
r
o
b
lem
s
ar
is
e
wh
en
th
e
m
ajo
r
ity
class
s
eg
m
en
t
h
a
s
a
h
ig
h
er
p
r
o
p
o
r
tio
n
th
an
th
e
m
in
o
r
ity
class
s
eg
m
en
t
[
9
]
.
N
u
m
er
o
u
s
r
ea
l
-
wo
r
ld
ap
p
licatio
n
s
,
s
u
ch
as
tex
t
class
if
icatio
n
,
d
ef
ec
t
d
etec
tio
n
,
f
r
au
d
d
etec
tio
n
,
o
il
s
p
ill
d
etec
tio
n
in
s
atellite
p
ictu
r
es,
to
x
ico
lo
g
y
,
cu
ltu
r
al
m
o
d
elin
g
,
an
d
m
ed
ical
d
iag
n
o
s
tics
,
ar
e
p
lag
u
ed
b
y
th
e
im
b
alan
ce
d
d
a
ta
s
et
p
r
o
b
lem
[
1
0
]
.
T
h
e
f
o
llo
win
g
f
ac
to
r
s
co
n
tr
ib
u
te
to
th
e
cu
r
r
en
t
class
if
i
ca
tio
n
alg
o
r
ith
m
s
'
in
ad
eq
u
ate
p
er
f
o
r
m
a
n
ce
o
n
im
b
alan
ce
d
d
ata
s
ets:
i)
T
h
ey
ar
e
ac
cu
r
ac
y
-
d
r
iv
en
,
i.e
.
,
th
e
y
s
tr
iv
e
to
r
ed
u
ce
th
e
o
v
er
all
er
r
o
r
,
o
f
wh
ich
th
e
m
in
o
r
ity
class
m
ak
es
u
p
a
v
er
y
s
m
all
p
o
r
tio
n
;
ii)
T
h
ey
m
ak
e
th
e
s
u
p
p
o
s
itio
n
th
at
th
e
d
ata
i
s
d
is
tr
ib
u
ted
eq
u
ally
ac
r
o
s
s
all
clas
s
es
;
an
d
iii)
T
h
ey
lik
ewise
p
r
esu
m
e
th
at
er
r
o
r
s
o
r
ig
in
atin
g
f
r
o
m
v
a
r
io
u
s
ca
teg
o
r
ies
h
av
e
th
e
s
am
e
co
s
t
[
1
1
]
.
B
o
th
at
th
e
d
ata
an
d
alg
o
r
ith
m
ic
lev
els,
a
v
a
r
iety
o
f
ap
p
r
o
ac
h
es
to
th
e
im
b
alan
ce
d
ataset
p
r
o
b
lem
h
av
e
alr
ea
d
y
b
ee
n
p
u
t
f
o
r
th
.
T
h
ese
ap
p
r
o
ac
h
es
in
clu
d
e
a
v
ar
iety
o
f
r
e
-
s
am
p
lin
g
tech
n
iq
u
es
at
th
e
d
ata
lev
el,
in
clu
d
in
g
r
a
n
d
o
m
o
v
er
s
am
p
li
n
g
with
r
ep
lace
m
en
t,
r
a
n
d
o
m
u
n
d
er
-
s
am
p
lin
g
,
d
ir
ec
ted
o
v
e
r
s
am
p
lin
g
,
d
ir
ec
ted
u
n
d
er
-
s
am
p
lin
g
,
o
v
e
r
s
am
p
lin
g
with
in
f
o
r
m
ed
g
en
er
atio
n
o
f
n
ew
s
am
p
les,
an
d
co
m
b
in
atio
n
s
o
f
th
e
tech
n
iq
u
es.
W
h
en
wo
r
k
in
g
with
d
ec
is
io
n
tr
ee
s
,
m
o
d
i
f
y
i
n
g
t
h
e
p
r
o
b
a
b
i
l
i
ty
e
s
t
i
m
at
e
a
t
t
h
e
t
r
e
e
l
ea
f
,
a
d
ju
s
t
i
n
g
t
h
e
d
e
c
is
i
o
n
t
h
r
e
s
h
o
l
d
,
a
n
d
r
e
c
o
g
n
i
t
i
o
n
-
b
a
s
e
d
r
a
t
h
e
r
t
h
a
n
d
i
s
c
r
i
m
i
n
a
t
i
o
n
-
b
a
s
e
d
l
e
a
r
n
i
n
g
a
r
e
s
o
m
e
a
l
g
o
r
ith
m
i
c
s
t
r
at
e
g
i
es
t
h
a
t
a
d
d
r
e
s
s
t
h
e
cl
as
s
i
m
b
a
l
a
n
c
e
[
1
2
]
.
M
o
s
t
c
l
as
s
i
f
i
e
r
s
,
i
n
cl
u
d
i
n
g
d
e
c
i
s
i
o
n
t
r
e
es
a
n
d
n
e
u
r
al
n
e
two
r
k
s
,
p
e
r
f
o
r
m
w
e
ll
w
h
e
n
t
h
e
d
a
t
a
s
e
t'
s
r
e
s
p
o
n
s
e
v
ar
i
a
b
l
e
d
i
s
t
r
i
b
u
t
i
o
n
i
s
b
a
l
a
n
c
e
d
b
u
t
i
t
s
u
f
f
e
r
s
f
r
o
m
t
h
e
p
r
o
b
l
e
m
o
f
i
m
b
a
l
a
n
c
e
[
1
3
]
.
I
n
p
u
t
d
a
t
a
s
a
m
p
l
i
n
g
is
r
e
g
a
r
d
ed
a
s
a
p
o
p
u
l
a
r
m
e
t
h
o
d
o
f
c
o
r
r
e
c
t
i
n
g
d
ata
im
b
alan
ce
.
I
t
m
o
d
i
f
ies
th
e
s
tr
u
ctu
r
e
o
f
th
e
r
ea
l
d
ata
s
et
to
c
h
an
g
e
its
b
alan
cin
g
r
atio
to
th
e
d
e
s
ir
ed
lev
el
b
y
u
s
in
g
b
o
th
u
n
d
er
-
s
am
p
lin
g
an
d
o
v
er
s
am
p
lin
g
[
1
4
]
.
As
m
en
ti
o
n
ed
i
n
t
h
e
i
n
t
r
o
d
u
c
ti
o
n
,
u
n
e
q
u
al
d
is
t
r
i
b
u
ti
o
n
o
f
s
a
m
p
les
ca
u
s
es
b
iase
d
le
ar
n
i
n
g
a
n
d
d
ir
ec
tl
y
af
f
e
cts
t
h
e
wo
r
t
h
i
n
ess
o
f
t
h
e
m
o
d
el
.
H
o
we
v
er
,
w
e
c
o
u
l
d
c
o
n
c
lu
d
e
th
at
t
r
a
in
in
g
o
n
t
h
e
im
b
a
l
an
ce
d
d
at
aset
h
as
a
s
ig
n
i
f
ic
a
n
tl
y
w
ea
k
im
p
a
ct
o
n
t
h
e
p
er
f
o
r
m
a
n
ce
,
s
o
i
t is
n
ec
ess
ar
y
t
o
i
n
v
esti
g
a
te
n
ew
tec
h
n
i
q
u
es
f
o
r
v
e
r
i
f
y
in
g
t
h
e
p
e
r
f
o
r
m
a
n
c
e
o
n
t
h
e
i
m
b
al
a
n
ce
d
d
a
tas
et.
M
o
s
t
o
f
t
h
e
te
c
h
n
iq
u
es
a
r
e
a
n
al
y
ze
d
u
s
i
n
g
t
h
e
b
al
an
ce
d
d
a
tase
t
i
.
e.
m
o
r
e
o
r
less
e
q
u
a
l
d
is
tr
ib
u
t
io
n
o
f
cl
ass
es.
S
o
,
a
n
e
f
f
ic
ie
n
t
p
e
r
f
o
r
m
an
ce
o
n
a
n
i
m
b
ala
n
c
e
d
m
ag
n
e
tic
r
es
o
n
a
n
ce
im
ag
e
d
at
ase
t is
a
v
e
r
y
c
h
a
lle
n
g
i
n
g
t
a
s
k
.
A
n
al
y
s
is
o
f
m
u
lti
v
a
r
i
ate
m
ed
i
ca
l i
m
a
g
i
n
g
is
a
n
i
m
p
o
r
ta
n
t
task
,
as w
as sai
d
in
th
e
in
tr
o
d
u
cti
o
n
.
B
r
a
in
t
u
m
o
r
s
c
au
s
e
m
o
s
t
ca
n
c
er
-
r
el
ate
d
f
at
aliti
es.
T
h
e
li
k
eli
h
o
o
d
t
h
a
t
a
p
atie
n
t
w
o
u
l
d
r
e
c
ei
v
e
an
e
f
f
ec
ti
v
e
t
h
e
r
a
p
y
is
s
i
g
n
if
i
c
an
t
ly
i
n
cr
ea
s
ed
b
y
ea
r
l
y
d
i
a
g
n
o
s
is
o
f
c
er
tai
n
d
is
o
r
d
e
r
s
.
T
h
e
p
r
o
p
o
s
e
d
s
o
lu
ti
o
n
is
“
h
y
b
r
id
iza
ti
o
n
o
f
p
a
r
t
icl
e
r
i
d
e
r
m
u
t
u
al
i
n
f
o
r
m
at
io
n
a
n
d
d
en
d
r
iti
c
c
ell
s
q
u
i
r
r
el
s
ea
r
c
h
al
g
o
r
it
h
m
wit
h
ar
ti
f
ic
ial
im
m
u
n
e
c
lass
i
f
i
er
(
AI
C
)
f
o
r
b
r
ain
t
u
m
o
r
cl
ass
i
f
ic
ati
o
n
f
o
r
i
m
b
al
a
n
ce
d
m
a
g
n
eti
c
r
es
o
n
a
n
c
e
i
m
a
g
e
d
at
aset
”.
T
h
e
g
o
al
o
f
th
is
r
esear
ch
is
to
an
aly
ze
p
a
r
ticle
r
id
er
m
u
tu
a
l
in
f
o
r
m
atio
n
(
PR
MI
)
an
d
d
e
n
d
r
itic
ce
ll
s
q
u
ir
r
el
s
ea
r
ch
alg
o
r
ith
m
(
DC
A
-
SS
A)
with
an
AI
C
[
1
5
]
f
o
r
b
r
ain
tu
m
o
r
class
if
icatio
n
o
n
an
im
b
alan
ce
d
m
ag
n
etic
r
eso
n
an
ce
im
a
g
e
d
a
taset.
R
eg
io
n
o
f
in
ter
est
an
d
Gau
s
s
ian
f
ilter
s
ar
e
u
s
ed
to
e
lim
in
ate
n
o
is
e
an
d
ar
tifa
cts
f
r
o
m
th
e
im
ag
e.
T
o
a
cq
u
ir
e
s
eg
m
en
ts
,
s
p
ar
s
e
f
u
zz
y
C
-
m
ea
n
s
clu
s
ter
in
g
(
Sp
a
r
s
e
FC
M)
[
1
6
]
h
as
also
b
ee
n
u
tili
ze
d
.
E
ac
h
s
eg
m
e
n
t'
s
s
tatis
tical
an
d
tex
tu
r
e
d
ata
ar
e
co
m
b
in
ed
to
f
o
r
m
a
f
ea
tu
r
e
v
ec
to
r
.
T
h
e
PR
MI
[
1
7
]
,
wh
ich
was
cr
ea
ted
b
y
f
u
s
in
g
th
e
p
ar
ticle
s
war
m
o
p
tim
izatio
n
(
PS
O)
[
1
8
]
,
r
id
er
o
p
ti
m
izatio
n
alg
o
r
ith
m
(
R
OA)
[
1
9
]
,
an
d
m
u
tu
al
in
f
o
r
m
atio
n
(
MI
)
,
is
u
tili
ze
d
f
o
r
f
ea
tu
r
e
s
elec
tio
n
.
AI
C
in
d
u
ce
s
th
e
ch
o
s
en
f
ea
tu
r
es to
id
en
tify
th
e
tu
m
o
r
o
u
s
r
eg
io
n
s
[
2
0
]
.
T
h
e
ar
tific
ial
im
m
u
n
e
class
if
ier
is
tr
ain
ed
u
s
in
g
th
e
DC
A
-
SSA
m
eth
o
d
[
2
1
]
,
wh
ich
co
m
b
in
es
th
e
s
q
u
ir
r
el
s
ea
r
ch
tech
n
i
q
u
e
(
SS
A)
[
2
2
]
a
n
d
th
e
d
en
d
r
itic
ce
ll
alg
o
r
ith
m
(
DC
A)
[
2
3
]
.
T
h
e
r
ec
o
m
m
en
d
ed
DC
A
-
SSA
m
eth
o
d
is
u
s
ed
to
im
p
r
o
v
e
th
e
a
r
tific
ial
im
m
u
n
e
class
if
ier
p
ar
am
eter
s
[
2
4
]
.
2.
M
E
T
H
O
D
As
s
h
o
wn
in
Fig
u
r
e
1
,
th
is
p
ap
er
o
f
f
er
s
p
er
f
o
r
m
an
ce
a
n
aly
s
is
f
o
r
th
e
b
r
ain
tu
m
o
r
class
if
icati
o
n
m
eth
o
d
n
am
ed
d
en
d
r
itic
ce
ll
-
s
q
u
ir
r
el
s
ea
r
ch
alg
o
r
ith
m
-
b
ased
class
if
ier
o
n
im
b
alan
ce
d
m
e
d
ical
im
ag
e
d
ataset
.
An
ex
p
er
im
en
tal
r
esu
lt
s
h
o
ws
th
at
th
e
p
r
o
p
o
s
ed
al
g
o
r
ith
m
g
i
v
es
p
r
o
m
is
in
g
r
esu
lts
o
n
th
e
im
b
alan
ce
d
m
e
d
ical
im
ag
e
d
ataset
.
Her
e,
th
e
p
r
ep
r
o
ce
s
s
in
g
o
f
th
e
i
n
p
u
t
MR
im
ag
e
is
ca
r
r
ied
o
u
t
with
th
e
h
e
lp
o
f
th
e
r
eg
io
n
o
f
in
ter
est
an
d
Gau
s
s
ian
f
ilter
.
T
h
e
p
r
ep
r
o
ce
s
s
ed
im
ag
e
is
g
iv
en
to
war
d
s
th
e
s
p
ar
s
e
f
u
zz
y
C
-
m
ea
n
s
f
o
r
th
e
s
eg
m
en
tatio
n
task
.
Statis
tical
f
ea
tu
r
es
an
d
lo
ca
l
d
e
r
iv
ativ
e
p
atter
n
s
(
L
DP)
ar
e
ex
tr
ac
ted
f
r
o
m
th
e
s
eg
m
en
te
d
im
ag
e.
Fo
llo
win
g
th
at,
th
e
f
ea
tu
r
e
s
elec
tio
n
p
r
o
ce
s
s
is
ac
co
m
p
lis
h
ed
b
y
co
m
b
in
in
g
th
r
ee
alg
o
r
ith
m
s
:
m
u
tu
al
in
f
o
r
m
atio
n
,
r
id
er
o
p
tim
iza
tio
n
tech
n
iq
u
e
,
an
d
PSO
.
T
h
e
i
n
teg
r
atio
n
o
f
th
e
s
q
u
ir
r
el
s
ea
r
ch
alg
o
r
ith
m
with
d
en
d
r
itic c
ell
alg
o
r
ith
m
co
m
p
l
etes th
e
class
if
icatio
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
P
erfo
r
ma
n
ce
a
n
a
lysi
s
o
f h
yb
r
id
b
io
-
in
s
p
ir
ed
a
l
g
o
r
ith
ms fo
r
cla
s
s
ifyin
g
b
r
a
in
…
(
R
a
h
u
l R
a
mesh
C
h
a
kre
)
6341
Fig
u
r
e
1
.
B
lo
ck
d
iag
r
am
o
f
h
y
b
r
id
class
if
icatio
n
m
eth
o
d
2
.
1
.
I
np
ut
bra
i
n
M
R
im
a
g
e
Su
p
p
o
s
e
th
e
im
b
alan
ce
d
d
atas
et
with
th
e
to
tal
n
u
m
b
er
o
f
s
am
p
les.
T
h
e
e
x
p
r
ess
io
n
f
o
r
th
e
d
ataset
with
th
e
q
u
an
tity
o
f
im
ag
es is
s
h
o
wn
(
1
)
.
=
{
1
,
2
,
.
.
.
,
.
.
.
}
(
1
)
H
er
e,
s
tates
th
e
ℎ
MR
im
ag
e
i
n
th
e
d
ataset
an
d
th
e
to
tal
n
u
m
b
er
o
f
s
am
p
les
in
th
e
d
atase
t
is
r
ep
r
esen
ted
as
.
2
.
2
.
P
re
-
pro
ce
s
s
ing
o
f
M
R
im
a
g
e
T
h
e
r
em
o
v
al
o
f
u
n
d
esira
b
le
d
i
s
to
r
tio
n
s
f
r
o
m
th
e
im
ag
e
d
u
r
in
g
p
r
e
-
p
r
o
ce
s
s
in
g
,
wh
ich
is
a
c
r
u
cial
s
tep
in
class
if
y
in
g
b
r
ain
t
u
m
o
r
s
,
is
p
er
f
o
r
m
ed
u
s
in
g
th
e
r
e
g
io
n
o
f
in
ter
est
an
d
Gau
s
s
ian
f
ilter
.
a.
R
eg
io
n
o
f
in
ter
est
B
y
co
n
s
id
er
in
g
s
u
cc
ess
iv
e
m
a
s
k
in
g
,
th
e
r
eg
i
o
n
o
f
in
ter
est
is
f
r
eq
u
e
n
tly
d
ete
r
m
in
ed
f
r
o
m
th
e
v
alu
es
o
f
p
ix
el
in
ten
s
ities
.
T
h
e
d
eter
m
in
atio
n
o
f
co
n
tin
u
o
u
s
p
ix
els
h
av
in
g
a
v
alu
e
o
f
1
an
d
ex
cl
u
d
in
g
th
e
v
alu
e
o
f
0
is
a
v
ital
s
tep
in
th
e
r
eg
io
n
o
f
i
n
ter
e
s
t
ex
tr
ac
tio
n
p
r
o
ce
s
s
.
T
h
e
v
alu
es
o
f
t
h
e
r
e
g
io
n
o
f
in
ter
est
i
n
ten
s
ity
ar
e
c
o
m
m
o
n
ly
r
ef
er
r
ed
to
as d
en
s
ity
s
lices.
b.
Gau
s
s
ian
f
ilter
T
h
e
Gau
s
s
ian
f
ilter
is
f
r
eq
u
en
tly
u
s
ed
t
o
r
e
m
o
v
e
n
o
is
e
a
n
d
it
ca
n
p
r
o
v
i
d
e
a
s
m
o
o
th
tr
an
s
i
tio
n
.
T
h
e
Gau
s
s
ian
two
-
d
im
en
s
io
n
al
d
is
tr
ib
u
tio
n
is
as
(
2
)
.
(
,
)
=
1
√
2
2
−
2
+
2
2
2
(
2
)
H
er
e,
s
tates
th
e
s
tan
d
ar
d
d
ev
ia
tio
n
wh
ile
an
d
ch
ar
ac
ter
ize
th
e
d
im
en
s
io
n
o
f
th
e
k
er
n
el.
T
h
e
p
r
e
-
p
r
o
ce
s
s
ed
o
u
tp
u
t a
c
q
u
ir
ed
a
f
ter
f
ilter
in
g
th
e
im
ag
e
is
id
en
tifie
d
as
.
2
.
3
.
Seg
m
ent
a
t
io
n o
f
prepro
ce
s
s
ed
im
a
g
e
T
h
e
p
r
e
-
p
r
o
ce
s
s
ed
o
u
t
p
u
t
is
g
iv
en
in
p
u
t
to
th
e
s
eg
m
en
tatio
n
p
r
o
ce
s
s
,
wh
ich
is
p
er
f
o
r
m
ed
u
s
in
g
s
p
ar
s
e
f
u
zz
y
C
-
m
ea
n
s
cl
u
s
ter
in
g
,
a
n
d
it
is
d
ev
elo
p
e
d
b
y
th
e
m
o
d
if
icatio
n
o
f
ty
p
ical
f
u
z
zy
C
-
m
ea
n
s
clu
s
ter
in
g
.
Utilizin
g
s
p
ar
s
e
f
u
zz
y
C
-
m
ea
n
s
clu
s
ter
in
g
o
f
f
er
s
h
ig
h
d
im
e
n
s
io
n
al
d
ata
cl
u
s
ter
in
g
,
w
h
ich
is
th
e
m
ain
b
e
n
ef
it.
T
h
e
f
o
llo
win
g
ac
tio
n
s
ar
e
tak
e
n
wh
en
u
s
in
g
s
p
ar
s
e
f
u
zz
y
C
-
m
ea
n
s
clu
s
ter
in
g
:
a.
I
n
itializatio
n
:
I
n
itially
,
th
e
f
ea
tu
r
e
weig
h
ts
ar
e
s
tim
u
lated
an
d
ar
e
r
ep
r
esen
ted
as
=
1
=
2
=
.
.
.
=
=
1
√
.
T
h
e
f
ea
tu
r
es a
ttain
ed
f
r
o
m
p
i
x
el
p
o
s
itio
n
ar
e
ass
u
m
ed
as
=
2
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
e
r
20
24
:
6
3
3
9
-
6
3
5
0
6342
b.
Up
d
atio
n
o
f
p
ar
titi
o
n
m
at
r
ix
:
At
f
ir
s
t,
th
e
attr
ib
u
te
weig
h
ts
an
d
clu
s
ter
ce
n
ter
wh
en
(
ℵ
)
is
m
i
n
im
ized
,
=
{
1
;
=
0
=
{
1
:
=
0
}
0
;
≠
0
=
0
,
≠
1
∑
(
)
(
1
−
1
)
=
1
;
ℎ
(
3
)
c.
Up
d
ate
clu
s
ter
ce
n
ter
K
: L
et
u
s
ass
u
m
e
th
at
an
d
b
e
f
ix
ed
an
d
(
)
is
m
in
im
ized
if
=
{
0
;
=
0
∑
.
=
1
∑
=
1
;
≠
0
(
4
)
Her
e,
th
e
weig
h
t c
o
m
p
o
n
en
t is r
ep
r
esen
ted
as
an
d
.
d.
E
v
a
l
u
a
t
e
t
h
e
c
l
as
s
:
T
h
e
c
l
a
s
s
v
al
u
e
i
s
e
v
a
l
u
a
t
e
d
wi
t
h
c
o
n
s
t
a
n
t
c
lu
s
t
e
r
s
=
{
1
,
3
,
…
,
,
…
}
a
n
d
m
e
m
b
e
r
s
h
i
p
.
e.
T
er
m
in
ate:
T
h
e
s
tep
s
ar
e
co
n
tin
u
ed
u
n
til
it
r
ea
ch
es
th
e
s
atis
f
ied
cr
iter
io
n
.
T
h
e
o
u
tco
m
e
o
f
s
p
ar
s
e
f
u
zz
y
C
-
m
ea
n
s
clu
s
ter
in
g
d
ep
icted
t
h
e
s
eg
m
en
ts
r
ec
eiv
ed
f
r
o
m
th
e
p
r
e
-
p
r
o
ce
s
s
ed
im
ag
e
an
d
it is
s
h
o
wn
(
5
)
.
=
{
1
,
3
,
…
,
,
…
}
(
5
)
2
.
4
.
F
ea
t
ure
ex
t
ra
c
t
io
n
T
o
r
ef
in
e
th
e
r
elev
a
n
t
f
ea
tu
r
e
s
f
o
r
s
u
b
s
eq
u
e
n
t
p
r
o
ce
s
s
in
g
,
f
ea
tu
r
e
ex
tr
ac
tio
n
is
a
cr
u
cial
s
tep
.
Her
e,
s
tatis
t
ical
f
ea
tu
r
es
m
ea
n
(
1
)
,
v
ar
i
an
ce
(
2
)
,
s
tan
d
ar
d
d
ev
iatio
n
(
3
)
,
k
u
r
to
s
is
(
4
)
,
s
k
ewn
ess
(
5
)
,
an
d
en
er
g
y
(
6
)
s
u
cc
ess
f
u
lly
ex
tr
ac
ted
alo
n
g
with
th
is
tex
tu
r
e
f
ea
tu
r
e
L
DP
(
7
)
is
also
ex
tr
ac
ted
.
Fo
r
m
atio
n
o
f
f
ea
tu
r
e
v
ec
to
r
.
=
{
1
,
2
,
3
,
4
,
5
,
6
,
7
}
(
6
)
2
.
5
.
F
ea
t
ure
s
elec
t
io
n
T
h
e
f
ea
t
u
r
e
v
ec
to
r
s
p
ec
if
ied
i
n
(
6
)
is
s
u
b
jecte
d
t
o
th
e
f
ea
t
u
r
e
s
elec
tio
n
s
tag
e.
Her
e,
PR
MI
is
u
s
ed
to
s
elec
t
th
e
m
o
s
t
r
elev
a
n
t
f
ea
tu
r
e.
T
h
is
alg
o
r
ith
m
is
d
ev
elo
p
e
d
b
y
th
e
co
m
b
in
atio
n
o
f
PSO
,
R
OA
,
an
d
m
u
tu
al
in
f
o
r
m
a
tio
n
.
T
h
e
s
tep
s
f
o
r
th
e
f
ea
tu
r
e
s
elec
tio
n
ar
e
as f
o
llo
ws:
a.
I
n
itializatio
n
:
I
n
itializatio
n
is
ca
r
r
ied
o
u
t
b
y
f
o
u
r
g
r
o
u
p
s
o
f
r
id
er
s
,
s
ca
tter
ed
o
v
er
a
d
im
en
s
i
o
n
al
s
p
ac
e,
an
d
s
h
o
wn
as
(
7
)
.
=
{
1
,
2
,
…
,
,
…
}
(
7
)
H
er
e,
in
d
icate
s
th
e
to
tal
n
u
m
b
er
o
f
s
o
lu
tio
n
s
an
d
r
ep
r
esen
ts
ℎ
s
o
lu
tio
n
.
b.
E
s
tab
lis
h
th
e
f
itn
ess
f
u
n
ctio
n
T
h
e
r
ec
en
tly
d
ev
elo
p
ed
o
b
j
ec
tiv
e
f
u
n
ctio
n
p
r
im
a
r
ily
r
el
ies
o
n
MI
[
1
6
]
,
w
h
ich
ca
lc
u
lates
th
e
r
elatio
n
s
h
ip
b
etwe
en
ch
ar
ac
te
r
is
tics
an
d
class
lab
els.
T
h
e
n
o
tio
n
o
f
"m
u
tu
al
in
f
o
r
m
atio
n
"
is
p
r
o
v
id
ed
b
y
(
,
)
.
T
h
e
d
escr
ip
tio
n
o
f
MI
b
etwe
en
two
f
ea
tu
r
es
an
d
class
lab
el
an
d
wh
o
s
e
jo
in
t
d
is
tr
ib
u
tio
n
is
d
escr
ib
ed
b
y
(
,
)
is
ex
p
r
ess
ed
as
(
8
)
.
(
;
)
=
∑
∑
(
,
)
(
,
)
(
)
.
(
)
∈
∈
(
8
)
wh
er
e,
(
)
an
d
(
)
in
d
icate
m
ar
g
in
al
d
is
tr
ib
u
tio
n
s
o
f
an
d
g
en
er
ated
b
y
th
e
m
a
r
g
in
aliza
tio
n
p
r
o
ce
s
s
.
Her
e,
th
e
in
d
icate
s
th
e
f
ea
tu
r
e
s
an
d
r
ep
r
esen
t th
e
class
lab
els.
c.
E
v
alu
atio
n
o
f
weig
h
ts
Acc
o
r
d
in
g
t
o
th
e
R
OA
,
th
e
a
ttack
er
'
s
lo
ca
tio
n
is
u
p
d
ated
to
ch
o
o
s
e
th
e
b
est
co
u
r
s
e
o
f
ac
tio
n
.
B
y
ass
u
m
in
g
th
e
ac
ti
o
n
s
o
f
ea
ch
r
id
er
,
as
d
ef
in
e
d
b
elo
w,
th
e
r
i
d
er
u
p
g
r
a
d
es
its
lo
ca
tio
n
.
T
h
e
p
o
s
itio
n
o
f
th
e
f
o
llo
wer
'
s
im
p
r
o
v
em
en
t
d
ep
en
d
s
o
n
th
e
lead
i
n
g
r
i
d
er
'
s
ab
ilit
y
to
r
ea
ch
th
e
g
o
al
an
d
is
as
(
9
)
.
+
1
(
,
)
=
(
,
)
+
[
(
,
∗
(
,
)
∗
)
]
(
9
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
P
erfo
r
ma
n
ce
a
n
a
lysi
s
o
f h
yb
r
id
b
io
-
in
s
p
ir
ed
a
l
g
o
r
ith
ms fo
r
cla
s
s
ifyin
g
b
r
a
in
…
(
R
a
h
u
l R
a
mesh
C
h
a
kre
)
6343
wh
er
e,
Φ
is
co
o
r
d
in
ate
s
elec
to
r
,
d
en
o
te
th
e
p
o
s
itio
n
o
f
th
e
le
ad
in
g
r
id
e
r
,
s
p
ec
if
ies
th
e
in
d
ex
o
f
th
e
lead
in
g
r
id
e
r
,
,
in
d
icate
s
th
e
s
teer
in
g
an
g
le
o
f
ℎ
r
id
er
i
n
ℎ
co
o
r
d
i
n
ate,
an
d
is
th
e
d
is
tan
ce
.
T
h
e
u
p
d
ated
p
o
s
itio
n
o
f
th
e
o
v
er
tak
er
is
u
s
ed
in
th
e
u
p
d
ate
p
r
o
ce
s
s
f
o
r
m
ax
im
izin
g
th
e
s
u
cc
ess
r
ate
b
y
d
eter
m
i
n
in
g
th
e
p
o
s
itio
n
o
f
th
e
o
v
er
tak
er
a
n
d
is
g
iv
en
b
y
(1
0
)
.
+
1
(
,
)
=
(
,
)
+
[
∗
(
)
∗
(
,
)
]
(1
0
)
w
h
er
e,
∗
(
)
r
ep
r
esen
t th
e
d
ir
ec
tio
n
in
d
icato
r
.
T
h
e
b
y
p
ass
r
id
er
s
f
o
llo
w
a
co
m
m
o
n
p
ath
with
o
u
t
tr
ac
k
i
n
g
th
e
lead
in
g
r
id
er
.
I
n
th
is
co
n
tex
t,
th
e
u
p
d
ate
r
u
le
o
f
t
h
e
b
y
p
ass
r
id
er
s
is
ex
h
ib
ited
in
wh
ich
th
e
s
tan
d
ar
d
b
y
p
ass
r
id
er
is
g
iv
en
as
(1
1
)
.
+
1
(
,
)
=
[
(
,
)
×
∆
(
)
+
(
,
)
∗
[
1
−
∆
(
)
]
]
(1
1
)
T
h
e
p
r
o
p
o
s
ed
PR
MI
'
s
f
in
al
u
p
g
r
ad
e
eq
u
atio
n
is
as
(1
2
)
.
+
1
(
,
)
=
(
(
,
)
(
1
1
+
2
2
−
1
)
−
(
,
)
−
1
1
(
,
)
)
(
1
+
(
,
)
)
+
2
2
2
2
−
1
−
,
(1
2
)
d.
T
er
m
in
ate:
T
h
e
s
tag
es a
r
e
r
e
p
ea
ted
u
n
til th
e
b
est s
o
lu
tio
n
is
d
is
co
v
er
ed
.
2
.
6
.
B
ra
in t
um
o
r
cla
s
s
if
ica
t
io
n us
ing
a
rt
if
icia
l im
m
un
e
cla
s
s
if
ier
-
dend
rit
ic
s
qu
irre
l sea
rc
h
a
lg
o
rit
hm
T
h
is
s
ec
tio
n
ex
p
lain
s
th
e
cla
s
s
if
icatio
n
o
f
b
r
ain
tu
m
o
r
s
,
wh
ich
is
d
o
n
e
u
s
in
g
an
AI
C
an
d
d
en
d
r
itic
s
q
u
ir
r
el
s
ea
r
ch
alg
o
r
ith
m
to
f
in
e
-
tu
n
e
th
e
class
if
ier
's
weig
h
ts
.
W
h
en
d
en
d
r
itic
ce
ll
alg
o
r
ith
m
an
d
s
q
u
ir
r
el
s
ea
r
ch
alg
o
r
ith
m
a
r
e
co
m
b
in
ed
,
th
e
r
esu
lt
is
th
e
p
r
o
d
u
ce
d
d
en
d
r
itic
-
SS
A.
T
h
e
r
esu
lt
o
f
th
is
class
if
icatio
n
p
r
o
ce
s
s
is
r
ef
er
r
ed
to
as.
T
h
e
m
ain
id
ea
b
eh
in
d
t
h
is
AI
C
is
th
at
it
co
n
s
id
er
s
th
e
im
m
u
n
e
s
y
s
tem
'
s
b
ac
k
g
r
o
u
n
d
n
etwo
r
k
wh
en
teac
h
in
g
th
e
id
ea
s
o
f
ar
tific
ial
in
tellig
en
ce
s
y
s
tem
(
AI
S),
wh
ich
is
d
ef
in
ed
as
a
m
ac
h
in
e
lear
n
in
g
m
ec
h
an
is
m
.
T
h
e
m
ai
n
ad
v
a
n
t
ag
e
o
f
AI
C
is
th
at
n
o
ad
d
itio
n
al
s
u
p
p
o
r
t
is
n
ee
d
ed
f
o
r
p
a
r
a
m
eter
o
p
tim
izatio
n
.
Ho
wev
er
,
th
e
m
ain
p
r
o
b
lem
with
co
n
v
e
n
tio
n
al
AI
C
s
is
an
tib
o
d
y
p
o
p
u
latio
n
p
r
o
d
u
ctio
n
.
T
h
e
s
tep
s
ar
e
as
f
o
llo
ws,
a.
I
n
itializatio
n
: T
h
e
ac
tiv
atio
n
o
f
weig
h
ts
an
d
th
eir
ass
o
ciate
d
p
ar
am
eter
s
is
th
e
f
ir
s
t stag
e.
=
{
1
,
2
,
…
,
,
.
.
}
(1
3
)
H
er
e,
is
th
e
lo
ca
tio
n
o
f
ℎ
s
o
lu
tio
n
an
d
s
h
o
ws th
e
o
v
er
all
s
o
lu
tio
n
s
.
b.
E
r
r
o
r
d
eter
m
in
atio
n
:
T
h
e
m
ea
n
s
q
u
ar
ed
er
r
o
r
(
MSE
)
,
d
is
p
lay
ed
b
elo
w,
is
th
e
b
est
s
o
lu
tio
n
d
is
co
v
er
ed
b
ased
o
n
th
e
f
itn
ess
m
ea
s
u
r
e.
=
1
∑
[
−
]
2
=
1
(1
4
)
H
er
e,
th
e
ex
p
ec
te
d
r
esu
lt a
n
d
th
e
o
u
tco
m
e
ac
h
iev
e
d
em
p
lo
y
i
n
g
AI
C
is
s
ig
n
if
ied
as
an
d
,
r
e
s
p
ec
tiv
ely
.
c.
E
s
tim
atio
n
o
f
th
e
m
o
d
if
ied
eq
u
atio
n
:
Utilizin
g
cr
ea
ted
d
en
d
r
itic
-
SS
A,
th
e
p
ar
am
eter
s
o
f
AI
C
ar
e
p
r
o
p
er
ly
s
et
to
id
en
tify
th
e
tim
o
r
o
u
s
r
e
g
io
n
o
f
th
e
b
r
ain
.
E
ac
h
an
tig
e
n
is
ap
p
r
o
p
r
iately
d
etec
ted
in
t
h
is
alg
o
r
ith
m
b
y
an
an
tib
o
d
y
,
a
n
d
ea
ch
co
n
n
ec
tio
n
o
f
a
n
tib
o
d
y
-
u
s
in
g
ag
en
ts
i
s
u
p
d
ated
as
(1
5
)
.
,
(
+
1
)
=
1
−
−
[
,
(
+
1
)
−
(
)
(
1
−
)
1
−
]
(1
5
)
d.
Re
-
ev
alu
atin
g
th
e
s
o
lu
tio
n
g
iv
en
th
e
e
r
r
o
r
:
T
h
e
er
r
o
r
is
ca
lc
u
lated
u
s
in
g
(
1
5
)
a
n
d
th
e
alg
o
r
ith
m
g
e
n
er
ates
less
er
r
o
r
an
d
is
u
tili
ze
d
f
o
r
tr
a
in
in
g
AI
C
.
e.
E
n
d
:
On
ce
th
e
o
p
tim
al
b
iase
s
h
av
e
b
ee
n
f
o
u
n
d
,
th
e
p
r
o
c
ess
is
r
ep
ea
ted
u
n
til
th
e
r
eq
u
ir
ed
n
u
m
b
er
o
f
iter
atio
n
s
h
as b
ee
n
r
ea
c
h
ed
.
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
is
s
ec
tio
n
elab
o
r
ates
o
n
th
e
p
er
f
o
r
m
a
n
ce
an
aly
s
is
o
f
ex
is
tin
g
th
e
PS
O+
NSA,
P
SO+CS
A,
AC
O+
NS
A,
AC
O+
C
SA,
AB
C
+N
SA,
A
B
C
+CS
A
tech
n
iq
u
es
alo
n
g
with
p
r
o
p
o
s
ed
R
OA
+
d
en
d
r
itic
-
SS
A
AI
C
[
9
]
an
d
p
ar
ticle
r
i
d
er
MI
+
d
en
d
r
itic
ce
ll
-
SSA
-
b
ased
AI
C
tech
n
iq
u
es
u
tili
zin
g
an
im
b
alan
ce
d
d
ataset
f
o
r
lear
n
i
n
g
an
d
class
if
icatio
n
o
f
b
r
ain
tu
m
o
r
s
u
s
in
g
ac
cu
r
ac
y
,
s
en
s
itiv
ity
,
an
d
s
p
ec
if
icity
as
p
er
f
o
r
m
an
ce
m
ea
s
u
r
es
alo
n
g
with
tim
e
co
m
p
lex
ity
a
n
aly
s
is
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
e
r
20
24
:
6
3
3
9
-
6
3
5
0
6344
T
h
e
s
tu
d
y
is
ca
r
r
ied
o
u
t
u
s
in
g
i
n
cr
ea
s
in
g
s
am
p
le
s
ize
tr
ai
n
in
g
d
ata.
I
n
p
u
t
d
ata
s
am
p
lin
g
is
th
e
co
m
m
o
n
s
o
lu
tio
n
f
o
r
t
h
e
im
b
alan
ce
d
im
ag
e
d
ataset,
it
ad
ju
s
ts
th
e
s
tr
u
c
tu
r
e
o
f
th
e
r
ea
l
d
ata
s
et
t
o
ch
an
g
e
its
b
alan
ce
r
atio
to
th
e
d
esire
d
lev
el
b
y
b
o
th
u
n
d
er
a
n
d
o
v
e
r
-
s
am
p
lin
g
[
1
0
]
.
Ad
d
itio
n
ally
,
th
e
ef
f
icien
c
y
o
f
p
ar
ticle
r
id
er
MI
+
d
en
d
r
itic
ce
ll
-
SSA
-
b
ased
AI
C
f
o
r
im
b
alan
ce
d
B
R
AT
S
an
d
SimBR
AT
S
d
atasets
is
also
ex
am
in
e
d
.
As
m
en
tio
n
ed
ea
r
lier
,
in
th
e
p
r
el
im
in
ar
y
in
v
esti
g
atio
n
,
we
h
av
e
ap
p
lied
d
ata
s
am
p
lin
g
o
n
im
p
licitly
g
en
er
ated
im
b
alan
ce
d
B
R
AT
S
an
d
SimBR
AT
S
MRI
d
ataset
.
Sam
p
lin
g
h
as
b
ee
n
d
o
n
e
i
n
p
e
r
ce
n
tag
e
s
f
r
o
m
4
0
%
to
9
0
%
wh
ile
in
cr
ea
s
in
g
th
e
s
am
p
le
s
ize
b
y
1
0
%.
T
h
e
im
b
alan
ce
d
d
ataset
g
en
er
ated
is
u
s
ed
f
o
r
ex
p
er
im
en
tatio
n
with
s
am
p
lin
g
an
d
with
o
u
t sam
p
lin
g
.
T
h
e
d
etailed
d
ataset
d
esc
r
ip
tio
n
is
g
iv
en
in
t
h
e
n
ex
t sectio
n
.
3
.
1
.
Da
t
a
s
et
des
cr
iptio
n
Usi
n
g
im
b
alan
ce
d
B
R
AT
S
an
d
s
im
u
lated
B
R
AT
S d
atase
ts
,
th
e
p
ar
ticle
r
id
er
MI
+
d
en
d
r
itic
ce
ll
-
SSA
-
b
ased
AI
C
is
test
ed
in
ter
m
s
o
f
ac
cu
r
ac
y
,
s
en
s
itiv
ity
,
an
d
s
p
ec
if
icity
.
T
h
e
d
atasets
f
o
r
B
R
AT
S
an
d
s
im
u
lated
B
R
AT
S
wer
e
tak
en
f
r
o
m
B
R
A
T
S
2
0
1
5
[
2
5
]
,
wh
ich
p
r
esen
ts
im
ag
es
with
v
ar
ied
d
eg
r
ee
s
o
f
s
ev
er
ity
.
3
0
p
atien
ts
'
b
r
ain
MRIs
ar
e
tak
en
in
to
co
n
s
id
er
atio
n
f
r
o
m
th
e
B
R
AT
S
d
at
ab
ase.
E
ac
h
p
atien
t'
s
im
ag
e
is
g
ath
er
ed
h
e
r
e
u
s
in
g
f
o
u
r
d
if
f
e
r
en
t
m
o
d
alities
,
in
clu
d
in
g
T
1
,
T
1
C
,
T
2
,
an
d
FLAI
R
.
Fo
r
th
e
s
tu
d
y
,
1
30
–
1
7
6
b
r
ain
s
lices
ar
e
p
r
o
d
u
ce
d
b
y
ea
ch
m
o
d
ality
.
T
h
is
d
ataset
o
f
MR
im
ag
es
h
as
b
ee
n
p
u
r
p
o
s
ely
im
b
alan
ce
d
.
A
to
tal
o
f
1
,
7
2
8
b
r
ain
MR
im
ag
es
with
v
ar
io
u
s
d
e
g
r
ee
s
o
f
b
r
ai
n
tu
m
o
r
s
ev
er
ity
wer
e
ta
k
en
f
r
o
m
th
e
B
R
AT
S
d
ataset
f
o
r
3
0
p
atien
ts
,
o
f
wh
ich
1
,
2
5
6
wer
e
tu
m
o
r
o
u
s
an
d
4
7
2
wer
e
n
o
t.
L
i
k
e
th
is
,
th
e
SimBR
AT
S
d
ataset
co
n
tain
s
in
f
o
r
m
atio
n
o
n
5
0
p
atien
ts
,
in
clu
d
in
g
1
2
4
MR
im
ag
es
wit
h
v
ar
ie
d
d
e
g
r
ee
s
o
f
b
r
ai
n
tu
m
o
r
s
ev
er
ity
,
o
f
wh
ich
9
6
ar
e
t
u
m
o
r
o
u
s
an
d
2
8
ar
e
non
-
tu
m
o
r
o
u
s
.
3
.
2
.
E
v
a
lua
t
io
n m
et
rics
T
h
e
p
er
f
o
r
m
a
n
ce
o
f
b
o
th
e
x
is
tin
g
an
d
p
r
o
p
o
s
ed
tec
h
n
iq
u
es
is
th
o
r
o
u
g
h
l
y
ex
p
l
o
r
ed
b
y
u
tili
zin
g
ac
cu
r
ac
y
,
s
en
s
itiv
ity
,
an
d
s
p
ec
if
icity
as
s
ig
n
if
ican
t
p
er
f
o
r
m
an
ce
m
ea
s
u
r
es.
T
h
is
e
v
alu
atio
n
is
s
u
p
er
v
is
ed
ex
p
licitly
o
n
th
e
im
b
alan
ce
d
B
R
AT
S
an
d
SimB
R
AT
S
d
ata
s
et
s
,
wh
ich
ar
e
co
n
f
r
o
n
tin
g
d
u
e
to
th
eir
f
u
n
d
a
m
en
tal
class
im
b
alan
ce
s
.
B
y
co
n
ce
n
tr
atin
g
o
n
th
ese
m
etr
ics,
th
e
ev
alu
atio
n
s
p
ec
if
ies
an
in
clu
s
iv
e
id
en
tific
atio
n
o
f
h
o
w
ef
f
ec
tiv
ely
ea
ch
m
eth
o
d
p
er
f
o
r
m
s
in
h
an
d
li
n
g
th
e
c
o
m
p
lex
iti
es o
f
th
ese
d
atasets
.
3
.
3
.
Co
m
pa
ra
t
iv
e
a
na
ly
s
is
3
.
3
.
1
.
E
x
perim
ent
a
t
io
ns
us
in
g
s
a
m
pled da
t
a
T
h
e
co
m
p
ar
ativ
e
an
aly
s
is
o
f
ex
is
tin
g
tec
h
n
iq
u
es
n
am
el
y
PS
O+
NSA,
PS
O+
C
SA,
A
C
O+
NSA,
AC
O+
C
SA,
AB
C
+N
SA,
AB
C
+C
SA
alo
n
g
with
p
r
o
p
o
s
ed
R
OA
+
d
en
d
r
itic
SS
A
AI
C
an
d
p
a
r
ticle
r
id
er
MI
+
d
en
d
r
itic
ce
ll
-
SSA
-
b
ased
AI
C
with
ac
cu
r
ac
y
,
s
en
s
itiv
ity
,
an
d
s
p
ec
if
icity
p
a
r
am
eter
s
is
ev
alu
ated
f
o
r
th
e
s
am
p
led
im
b
alan
ce
d
d
ataset
wh
ich
is
m
e
n
t
i
o
n
e
d
e
a
r
l
i
e
r
i
n
t
h
e
e
x
p
e
r
i
m
e
n
ta
l
r
es
u
l
ts
a
n
d
d
i
s
c
u
s
s
i
o
n
s
e
ct
i
o
n
.
A
s
s
h
o
w
n
i
n
T
a
b
l
e
1
,
p
r
o
p
o
s
e
d
a
p
p
r
o
a
c
h
e
s
,
i
n
cl
u
d
i
n
g
P
SO
+
DC
A
,
R
O
A+
d
e
n
d
r
i
t
i
c
-
S
SA
b
a
s
e
d
A
I
C
,
a
n
d
P
R
M
I
+
d
e
n
d
r
i
t
i
c
-
S
SA
b
a
s
e
d
AI
C
,
g
e
t
e
x
c
e
l
l
e
n
t
a
c
c
u
r
a
c
y
s
c
o
r
e
s
(
9
1
.
7
9
%
t
o
9
6
.
2
5
%
)
t
o
w
a
r
d
s
t
h
e
h
i
g
h
e
r
e
n
d
o
f
t
h
e
t
r
a
i
n
i
n
g
d
at
a
p
e
r
c
e
n
t
a
g
e
s
(
7
0
%
t
o
9
0
%
)
,
d
e
m
o
n
s
t
r
a
ti
n
g
th
e
i
r
u
s
e
f
u
l
n
es
s
.
T
h
e
a
cc
u
r
a
c
y
r
a
t
i
n
g
s
o
f
t
h
e
f
o
u
r
a
p
p
r
o
a
c
h
e
s
,
P
SO
+C
S
A
a
n
d
AC
O
+CS
A
,
v
ar
y
s
ig
n
if
ican
tly
w
h
en
c
o
m
p
ar
e
d
to
v
a
r
io
u
s
tr
ai
n
in
g
d
ata
p
e
r
ce
n
tag
es.
T
h
is
s
h
o
ws th
at
th
e
am
o
u
n
t
o
f
tr
ain
in
g
d
ata
m
ay
h
a
v
e
a
g
r
ea
ter
im
p
ac
t o
n
h
o
w
well
th
ese
s
tr
ateg
ies p
er
f
o
r
m
.
As
s
h
o
wn
in
T
ab
le
2
,
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
es,
n
o
tab
ly
PS
O+
DC
A,
R
OA+
d
en
d
r
itic
-
SS
A
b
ased
AI
C
,
an
d
PR
MI
+d
en
d
r
itic
-
SS
A
b
ased
AI
C
,
p
r
o
d
u
ce
h
ig
h
s
en
s
itiv
ity
v
alu
es (
8
9
.
9
5
% to
9
4
.
3
3
%
)
at
t
h
e
h
ig
h
er
e
n
d
o
f
th
e
tr
ain
in
g
d
ata
p
er
ce
n
tag
es
(
7
0
%
to
9
0
%).
T
h
is
s
u
g
g
ests
th
at
th
ese
tech
n
iq
u
es
ar
e
p
ar
ticu
lar
ly
ef
f
ec
tiv
e
f
o
r
d
etec
tin
g
p
o
s
itiv
e
ca
s
es
in
th
e
B
R
AT
S
d
ataset.
P
R
MI
an
d
P
SO
with
DC
A
an
d
d
en
d
r
itic
-
SS
A
b
ased
o
n
AI
C
,
th
ese
two
tech
n
iq
u
es
c
o
n
s
is
ten
tly
ex
h
ib
it
e
x
ce
p
tio
n
al
s
en
s
itiv
ity
o
v
er
all
tr
ain
in
g
d
ata
p
r
o
p
o
r
ti
o
n
s
,
m
ak
i
n
g
th
em
co
n
ten
d
er
s
f
o
r
m
o
r
e
ex
p
l
o
r
atio
n
o
r
ap
p
licatio
n
in
cir
c
u
m
s
tan
ce
s
wh
er
e
s
en
s
itiv
ity
is
cr
u
cial.
T
ab
le
1
.
Acc
u
r
ac
y
an
al
y
s
is
f
o
r
b
r
ain
tu
m
o
r
class
if
icatio
n
o
n
t
h
e
B
R
AT
S d
ataset
u
s
in
g
ex
is
ti
n
g
an
d
p
r
o
p
o
s
ed
tech
n
iq
u
es with
v
ar
y
in
g
tr
ai
n
in
g
d
ata
I
mb
a
l
a
n
c
e
d
B
R
A
TS
d
a
t
a
se
t
t
r
a
i
n
i
n
g
d
a
t
a
i
n
%
4
0
%
5
0
%
6
0
%
7
0
%
8
0
%
9
0
%
M
e
t
h
o
d
s
A
c
c
u
r
a
c
y
f
o
r
4
0
%
A
c
c
u
r
a
c
y
f
o
r
5
0
%
A
c
c
u
r
a
c
y
f
o
r
6
0
%
A
c
c
u
r
a
c
y
f
o
r
7
0
%
A
c
c
u
r
a
c
y
f
o
r
8
0
%
A
c
c
u
r
a
c
y
f
o
r
9
0
%
P
S
O
+
N
S
A
7
3
.
3
1
7
4
.
4
5
7
5
.
2
1
7
6
.
4
3
7
9
.
9
2
8
1
.
3
2
P
S
O
+
C
S
A
7
2
.
8
6
7
3
.
1
6
7
4
.
4
7
7
7
.
0
7
7
9
.
7
6
8
6
.
5
A
C
O
+
N
S
A
7
3
.
3
8
7
8
.
5
8
0
.
4
2
8
2
.
1
2
8
5
.
4
6
8
8
.
9
9
A
C
O
+
C
S
A
7
5
.
6
8
8
0
.
5
7
8
0
.
7
2
8
1
.
7
9
8
6
.
7
1
8
9
.
0
3
A
B
C
+
N
S
A
7
7
.
6
2
8
1
.
7
5
8
4
.
5
2
8
5
.
4
5
8
8
.
0
9
9
1
.
3
5
A
B
C
+
C
S
A
7
9
.
2
8
3
.
6
4
8
6
.
4
5
8
8
.
2
8
8
9
.
0
2
9
2
.
5
5
P
S
O
+
D
C
A
8
2
.
4
1
8
4
.
5
3
9
0
.
6
6
9
1
.
7
9
9
2
.
9
2
9
4
.
0
6
R
O
A
+
d
e
n
d
r
i
t
i
c
-
SSA
-
b
a
se
d
A
I
C
8
5
.
5
8
7
.
6
3
9
0
.
7
5
9
1
.
8
8
9
3
.
0
2
9
4
.
8
5
P
R
M
I
+
d
e
n
d
r
i
t
i
c
-
SSA
-
b
a
se
d
A
I
C
8
8
.
6
8
9
.
7
2
9
0
.
8
5
9
1
.
9
8
9
3
.
1
1
9
6
.
2
5
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
P
erfo
r
ma
n
ce
a
n
a
lysi
s
o
f h
yb
r
id
b
io
-
in
s
p
ir
ed
a
l
g
o
r
ith
ms fo
r
cla
s
s
ifyin
g
b
r
a
in
…
(
R
a
h
u
l R
a
mesh
C
h
a
kre
)
6345
T
ab
le
2
.
Sen
s
itiv
ity
an
aly
s
is
f
o
r
b
r
ai
n
tu
m
o
r
class
if
icatio
n
o
n
th
e
B
R
AT
S d
ataset
u
s
in
g
ex
is
tin
g
an
d
p
r
o
p
o
s
ed
tech
n
iq
u
es with
v
ar
y
in
g
tr
ai
n
in
g
d
ata
I
mb
a
l
a
n
c
e
d
B
R
A
TS
d
a
t
a
se
t
t
r
a
i
n
i
n
g
d
a
t
a
i
n
%
4
0
%
5
0
%
6
0
%
7
0
%
8
0
%
9
0
%
M
e
t
h
o
d
s
S
e
n
s
i
t
i
v
i
t
y
f
o
r
4
0
%
S
e
n
s
i
t
i
v
i
t
y
f
o
r
5
0
%
S
e
n
s
i
t
i
v
i
t
y
f
o
r
6
0
%
S
e
n
s
i
t
i
v
i
t
y
f
o
r
7
0
%
S
e
n
s
i
t
i
v
i
t
y
f
o
r
8
0
%
S
e
n
s
i
t
i
v
i
t
y
f
o
r
9
0
%
P
S
O
+
N
S
A
7
1
.
8
4
7
2
.
9
6
7
3
.
7
1
7
4
.
9
7
8
.
3
2
7
9
.
6
9
P
S
O
+
C
S
A
7
1
.
4
1
7
1
.
7
7
2
.
9
8
7
5
.
5
3
7
8
.
1
7
8
4
.
7
7
A
C
O
+
N
S
A
7
1
.
9
1
7
6
.
9
3
7
8
.
8
1
8
0
.
4
8
8
3
.
7
5
8
7
.
2
1
A
C
O
+
C
S
A
7
4
.
1
7
7
8
.
9
6
7
9
.
1
8
0
.
1
5
8
4
.
9
7
8
7
.
2
5
A
B
C
+
N
S
A
7
6
.
0
7
8
0
.
1
1
8
2
.
8
3
8
3
.
7
4
8
6
.
3
3
8
9
.
5
3
A
B
C
+
C
S
A
7
7
.
6
2
8
1
.
9
6
8
4
.
7
2
8
6
.
5
2
8
7
.
2
4
9
0
.
7
P
S
O
+
D
C
A
8
0
.
7
6
8
2
.
8
4
8
8
.
8
5
8
9
.
9
5
9
1
.
0
6
9
2
.
1
7
R
O
A
+
d
e
n
d
r
i
t
i
c
-
SSA
-
b
a
se
d
A
I
C
8
3
.
7
9
8
5
.
8
7
8
8
.
9
4
9
0
.
0
5
9
1
.
1
6
9
2
.
9
6
P
R
M
I
+
d
e
n
d
r
i
t
i
c
-
SSA
-
b
a
se
d
A
I
C
8
6
.
8
2
8
7
.
9
3
8
9
.
0
3
9
0
.
1
4
9
1
.
2
5
9
4
.
3
3
As
s
h
o
wn
in
T
ab
le
3
,
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
es,
n
o
tab
ly
PS
O+
DC
A,
R
OA+
d
en
d
r
itic
-
SS
A
b
ased
AI
C
,
an
d
PR
MI
+d
en
d
r
itic
-
SS
A
b
ased
A
I
C
p
r
o
d
u
ce
elev
ated
s
p
ec
if
icit
y
v
alu
es (
9
1
.
8
8
% to
9
4
.
8
5
%
)
at
th
e
h
ig
h
er
en
d
o
f
th
e
tr
ain
in
g
d
ata
p
e
r
ce
n
tag
es (
7
0
%
to
9
0
%).
T
h
ese
tech
n
iq
u
es
ar
e
co
n
s
is
ten
t
f
o
r
d
etec
tin
g
tr
u
e
n
e
g
ativ
e
ca
s
es.
PR
MI
an
d
R
OA
an
d
d
en
d
r
itic
-
SS
A
b
ased
AI
C
.
T
h
ese
ap
p
r
o
ac
h
es
co
n
s
is
ten
tly
ex
h
ib
it
ex
c
ep
tio
n
al
s
p
ec
if
icity
in
all
tr
ain
in
g
d
ata
p
r
o
p
o
r
tio
n
s
,
ac
co
r
d
in
g
to
A
I
C
,
m
ak
in
g
th
em
s
u
itab
le
o
p
tio
n
s
f
o
r
ap
p
licatio
n
s
wh
er
e
s
p
ec
if
icity
is
im
p
o
r
tan
t.
T
ab
le
3
.
Sp
ec
if
icity
a
n
aly
s
is
f
o
r
b
r
ai
n
tu
m
o
r
class
if
icatio
n
o
n
th
e
B
R
AT
S d
ataset
u
s
in
g
ex
is
tin
g
an
d
p
r
o
p
o
s
ed
tech
n
iq
u
es with
v
ar
y
in
g
tr
ai
n
in
g
d
ata
I
mb
a
l
a
n
c
e
d
B
R
A
TS
d
a
t
a
se
t
t
r
a
i
n
i
n
g
d
a
t
a
i
n
%
4
0
%
5
0
%
6
0
%
7
0
%
8
0
%
9
0
%
M
e
t
h
o
d
s
S
p
e
c
i
f
i
c
i
t
y
f
o
r
4
0
%
S
p
e
c
i
f
i
c
i
t
y
f
o
r
5
0
%
S
p
e
c
i
f
i
c
i
t
y
f
o
r
6
0
%
S
p
e
c
i
f
i
c
i
t
y
f
o
r
7
0
%
S
p
e
c
i
f
i
c
i
t
y
f
o
r
8
0
%
S
p
e
c
i
f
i
c
i
t
y
f
o
r
9
0
%
P
S
O
+
N
S
A
7
3
.
3
8
7
4
.
5
3
7
5
.
2
9
7
6
.
5
1
7
9
.
9
9
8
1
.
4
P
S
O
+
C
S
A
7
2
.
9
4
7
3
.
2
4
7
4
.
5
4
7
7
.
1
5
7
9
.
8
4
8
6
.
5
8
A
C
O
+
N
S
A
7
3
.
4
5
7
8
.
5
8
8
0
.
5
8
2
.
2
8
5
.
5
5
8
9
.
0
8
A
C
O
+
C
S
A
7
5
.
7
6
8
0
.
6
5
8
0
.
8
8
1
.
8
7
8
6
.
7
9
8
9
.
1
2
A
B
C
+
N
S
A
7
7
.
7
8
1
.
8
3
8
4
.
6
8
5
.
5
4
8
8
.
1
8
9
1
.
4
4
A
B
C
+
C
S
A
7
9
.
2
8
8
3
.
7
2
8
6
.
5
3
8
8
.
3
7
8
9
.
1
1
9
2
.
6
4
P
S
O
+
D
C
A
8
2
.
4
9
8
4
.
6
2
9
0
.
7
5
9
1
.
8
8
9
3
.
0
1
9
4
.
1
5
R
O
A
+
d
e
n
d
r
i
t
i
c
-
SSA
-
b
a
se
d
A
I
C
8
5
.
5
9
8
7
.
7
1
9
0
.
8
4
9
1
.
9
8
9
3
.
1
1
9
4
.
9
5
P
R
M
I
+
d
e
n
d
r
i
t
i
c
-
SSA
-
b
a
se
d
A
I
C
8
8
.
6
8
8
9
.
8
1
9
0
.
9
4
9
2
.
3
7
9
3
.
1
3
9
4
.
8
5
As
s
h
o
wn
in
T
ab
le
4
,
t
h
e
ac
cu
r
ac
y
p
er
ce
n
tag
e
f
o
r
d
if
f
er
e
n
t
m
eth
o
d
s
ap
p
lied
t
o
th
e
i
m
b
alan
ce
d
SimB
R
AT
S
d
ataset
w
ith
tr
ain
in
g
d
ata
r
an
g
in
g
f
r
o
m
4
0
%
to
9
0
%.
"PR
MI
+
d
en
d
r
itic
-
SSA
-
b
ased
AI
C
"
attain
s
s
u
p
er
io
r
ac
c
u
r
ac
y
,
attain
m
en
t
f
r
o
m
9
6
.
3
5
%
at
9
0
%
tr
ain
i
n
g
d
ata.
"PSO+NSA"
s
tead
ily
h
as
th
e
lo
wer
m
o
s
t
ac
cu
r
ac
y
,
r
ef
in
in
g
f
r
o
m
7
0
.
3
8
% to
7
8
.
4
% a
s
tr
ain
in
g
d
ata
in
cr
ea
s
es.
As
s
h
o
wn
in
T
ab
le
5
,
th
e
s
en
s
itiv
ity
p
er
ce
n
tag
e
f
o
r
m
eth
o
d
s
o
n
th
e
im
b
alan
ce
d
SimBR
A
T
S
d
ataset
with
tr
ain
in
g
d
ata
is
f
r
o
m
4
0
%
to
9
0
%.
"PR
MI
+d
en
d
r
itic
-
SSA
-
b
ased
AI
C
"
attain
s
s
u
p
er
io
r
s
en
s
itiv
ity
,
r
ea
ch
in
g
9
4
.
4
2
%
with
9
0
%
tr
ain
in
g
d
at
a.
"ROA+d
en
d
r
itic
-
SSA
-
ba
s
e
d
AI
C
"
also
p
er
f
o
r
m
s
h
ig
h
ly
,
with
s
en
s
itiv
ity
r
i
s
in
g
f
r
o
m
8
2
.
8
8
%
to
9
2
.
0
5
%.
"AB
C
+C
SA"
an
d
"ABC
+N
S
A"
ex
h
ib
it
r
o
b
u
s
t
s
en
s
itiv
ity
ad
v
an
ce
s
,
r
ea
ch
in
g
u
p
to
9
0
.
7
9
%
a
n
d
8
8
.
6
2
%,
co
r
r
esp
o
n
d
in
g
l
y
,
at
9
0
%
tr
ain
in
g
d
a
ta.
"PSO+DC
A"
an
d
"ACO+
C
SA"
al
s
o
ex
h
ib
it
p
r
o
m
in
e
n
t sen
s
itiv
ity
g
ain
s
,
r
ea
lizin
g
9
2
.
2
7
% a
n
d
8
7
.
3
4
% a
t 9
0
% tr
ain
in
g
d
ata.
"PSO+NS
A"
co
n
tin
u
es a
s
th
e
least sen
s
it
iv
e
m
eth
o
d
,
p
r
o
g
r
e
s
s
in
g
f
r
o
m
6
9
.
9
7
% to
7
9
.
8
3
%
with
m
o
r
e
tr
ain
in
g
d
ata.
As s
h
o
wn
in
T
ab
le
6
,
th
e
s
p
ec
i
f
icity
p
er
ce
n
tag
e
f
o
r
d
if
f
e
r
en
t
m
eth
o
d
s
o
n
t
h
e
im
b
ala
n
ce
d
SimBR
AT
S
d
ataset,
with
tr
ain
in
g
d
ata
r
an
g
in
g
f
r
o
m
4
0
%
to
9
0
%.
"P
R
MI
+d
en
d
r
itic
-
SSA
-
b
ased
AI
C
"
s
tead
ily
at
tain
s
s
u
p
er
io
r
s
p
ec
if
icity
,
r
ea
ch
in
g
9
5
.
4
4
%
at
9
0
%
tr
ain
in
g
d
ata
.
"ROA+d
en
d
r
itic
-
SSA
-
b
ased
AI
C
"
co
m
p
r
eh
en
d
s
in
ten
tly
,
with
s
p
ec
if
icity
in
cr
ea
s
in
g
f
r
o
m
8
5
.
6
7
%
to
9
4
.
0
4
%
as
tr
ain
in
g
d
ata
e
x
p
an
d
s
.
"ABC
+
C
SA"
an
d
"ABC
+
NSA"
also
s
h
o
w
co
n
s
id
er
ab
ly
s
tr
o
n
g
p
er
f
o
r
m
an
ce
,
with
s
p
ec
if
icity
g
o
in
g
u
p
to
9
1
.
7
3
%
an
d
9
0
.
5
4
%,
co
r
r
esp
o
n
d
i
n
g
ly
,
at
9
0
%
tr
ain
i
n
g
d
ata.
"PSO+DCA"
ex
h
ib
its
s
u
b
s
tan
tial
ad
v
an
ce
m
en
t,
ac
c
o
m
p
lis
h
in
g
9
2
.
2
4
%
s
p
ec
if
icity
at
9
0
%
tr
ain
in
g
d
at
a.
"ACO+C
SA"
an
d
"ACO+N
SA"
r
ea
ch
u
p
to
9
0
.
4
7
%
an
d
8
9
.
1
6
%
s
p
ec
if
icity
,
co
r
r
esp
o
n
d
in
g
ly
.
"PSO+NSA
"
h
as
m
in
im
al
s
p
ec
if
icity
b
u
t
s
till
a
d
ju
s
ts
f
r
o
m
7
0
.
4
5
%
to
7
8
.
4
8
%
as
tr
ain
in
g
d
ata
in
cr
ea
s
es.
T
h
e
r
esu
lts
s
h
o
w
a
ty
p
ical
in
clin
atio
n
o
f
in
cr
ea
s
in
g
s
p
ec
if
icity
as
th
e
p
er
ce
n
tag
e
o
f
tr
ai
n
in
g
d
ata
in
cr
ea
s
es
f
o
r
all
m
eth
o
d
s
.
T
h
is
im
p
lies
th
at
ea
ch
m
et
h
o
d
is
m
o
r
e
p
r
ec
is
e
in
ap
p
r
o
p
r
iately
id
en
tif
y
in
g
n
eg
ativ
e
in
s
tan
ce
s
with
m
o
r
e
tr
ain
in
g
d
ata.
Me
th
o
d
s
esp
ec
ially
in
teg
r
atin
g
d
en
d
r
itic
-
SSA
-
b
ased
AI
C
,
s
u
ch
as
"PR
MI
+d
en
d
r
itic
-
SSA
-
b
ased
AI
C
"
an
d
"ROA+d
en
d
r
itic
-
SSA
-
b
ased
AI
C
,
"
s
tead
ily
o
u
tp
er
f
o
r
m
o
th
er
s
,
em
p
h
asizin
g
th
eir
r
o
b
u
s
tn
ess
in
h
an
d
li
n
g
th
e
im
b
alan
ce
d
Si
m
B
R
AT
S d
ata
s
et.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
e
r
20
24
:
6
3
3
9
-
6
3
5
0
6346
T
ab
le
4
.
Acc
u
r
ac
y
an
al
y
s
is
f
o
r
b
r
ain
tu
m
o
r
class
if
icatio
n
o
n
t
h
e
Sim
B
R
AT
S d
ata
s
et
u
s
in
g
ex
is
tin
g
an
d
p
r
o
p
o
s
ed
tech
n
iq
u
es with
v
ar
y
in
g
tr
ain
in
g
d
ata
I
mb
a
l
a
n
c
e
d
S
i
mB
R
A
TS
d
a
t
a
s
e
t
t
r
a
i
n
i
n
g
d
a
t
a
i
n
%
4
0
%
5
0
%
6
0
%
7
0
%
8
0
%
9
0
%
M
e
t
h
o
d
s
A
c
c
u
r
a
c
y
f
o
r
4
0
%
A
c
c
u
r
a
c
y
f
o
r
5
0
%
A
c
c
u
r
a
c
y
f
o
r
6
0
%
A
c
c
u
r
a
c
y
f
o
r
7
0
%
A
c
c
u
r
a
c
y
f
o
r
8
0
%
A
c
c
u
r
a
c
y
f
o
r
9
0
%
P
S
O
+
N
S
A
7
0
.
3
8
7
1
.
5
3
7
2
.
2
9
7
3
.
5
7
6
.
9
9
7
8
.
4
P
S
O
+
C
S
A
7
2
.
9
4
7
3
.
2
4
7
4
.
5
4
7
7
.
1
5
7
9
.
8
4
8
4
.
5
8
A
C
O
+
N
S
A
7
3
.
4
5
7
8
.
5
8
8
0
.
5
8
2
.
2
8
5
.
5
5
8
9
.
0
8
A
C
O
+
C
S
A
7
5
.
7
6
8
0
.
6
5
8
0
.
8
8
1
.
8
7
8
6
.
7
9
8
9
.
1
2
A
B
C
+
N
S
A
7
7
.
7
8
1
.
8
3
8
4
.
6
8
5
.
5
4
8
8
.
1
8
9
1
.
4
5
A
B
C
+
C
S
A
7
9
.
2
8
8
3
.
7
2
8
6
.
5
3
8
8
.
3
7
8
9
.
1
1
9
2
.
6
4
P
S
O
+
D
C
A
8
2
.
4
9
8
4
.
6
2
9
0
.
7
5
9
1
.
8
8
9
3
.
0
1
9
4
.
1
5
R
O
A
+
d
e
n
d
r
i
t
i
c
-
SSA
-
b
a
se
d
A
I
C
8
5
.
5
9
8
7
.
7
2
9
0
.
8
5
9
1
.
9
8
9
3
.
1
1
9
4
.
9
5
P
R
M
I
+
d
e
n
d
r
i
t
i
c
-
SSA
-
b
a
se
d
A
I
C
8
8
.
6
8
8
9
.
8
1
9
0
.
9
4
9
2
.
0
7
9
3
.
2
1
9
6
.
3
5
T
ab
le
5
.
Sen
s
itiv
ity
an
aly
s
is
f
o
r
b
r
ai
n
tu
m
o
r
class
if
icatio
n
o
n
th
e
Sim
B
R
AT
S d
ataset
u
s
in
g
ex
is
tin
g
an
d
p
r
o
p
o
s
ed
tech
n
iq
u
es with
v
ar
y
in
g
tr
ain
in
g
d
ata
I
mb
a
l
a
n
c
e
d
S
i
mB
R
A
TS
d
a
t
a
s
e
t
t
r
a
i
n
i
n
g
d
a
t
a
i
n
%
4
0
%
5
0
%
6
0
%
7
0
%
8
0
%
9
0
%
M
e
t
h
o
d
s
S
e
n
s
i
t
i
v
i
t
y
f
o
r
4
0
%
S
e
n
s
i
t
i
v
i
t
y
f
o
r
5
0
%
S
e
n
s
i
t
i
v
i
t
y
f
o
r
6
0
%
S
e
n
s
i
t
i
v
i
t
y
f
o
r
7
0
%
S
e
n
s
i
t
i
v
i
t
y
f
o
r
8
0
%
S
e
n
s
i
t
i
v
i
t
y
f
o
r
9
0
%
P
S
O
+
N
S
A
6
9
.
9
7
7
0
.
6
7
1
.
8
4
7
2
.
0
3
7
4
.
4
5
7
9
.
8
3
P
S
O
+
C
S
A
7
0
.
4
8
7
1
.
7
7
7
2
.
0
5
7
4
.
6
1
7
7
.
2
5
8
3
.
8
9
A
C
O
+
N
S
A
7
4
.
9
8
7
6
.
0
1
7
8
.
3
9
7
9
.
5
6
8
1
.
8
4
8
6
.
2
9
A
C
O
+
C
S
A
7
4
.
2
4
7
9
.
0
4
7
9
.
1
8
8
0
.
2
3
8
5
.
0
6
8
7
.
3
4
A
B
C
+
N
S
A
7
8
.
1
5
7
9
.
1
9
8
2
.
5
1
8
3
.
8
3
8
6
.
4
2
8
8
.
6
2
A
B
C
+
C
S
A
7
7
.
6
9
8
2
.
0
5
8
4
.
8
8
6
.
6
8
7
.
3
3
9
0
.
7
9
P
S
O
+
D
C
A
7
9
.
8
4
8
1
.
9
3
8
6
.
9
4
8
8
.
0
4
9
0
.
1
5
9
2
.
2
7
R
O
A
+
d
e
n
d
r
i
t
i
c
-
SSA
-
b
a
se
d
A
I
C
8
2
.
8
8
8
4
.
9
6
8
8
.
0
3
8
9
.
1
4
9
0
.
2
5
9
2
.
0
5
P
R
M
I
+
d
e
n
d
r
i
t
i
c
-
SSA
-
b
a
se
d
A
I
C
8
6
.
9
1
8
7
.
0
1
8
8
.
1
2
9
0
.
2
3
9
1
.
3
4
9
4
.
4
2
T
ab
le
6
.
Sp
ec
if
icity
a
n
aly
s
is
f
o
r
b
r
ai
n
tu
m
o
r
class
if
icatio
n
o
n
th
e
Sim
B
R
AT
S d
ataset
u
s
in
g
ex
is
tin
g
an
d
p
r
o
p
o
s
ed
tech
n
iq
u
es with
v
ar
y
in
g
tr
ain
in
g
d
ata
I
mb
a
l
a
n
c
e
d
S
i
mB
R
A
TS
d
a
t
a
s
e
t
t
r
a
i
n
i
n
g
d
a
t
a
i
n
%
4
0
%
5
0
%
6
0
%
7
0
%
8
0
%
9
0
%
M
e
t
h
o
d
s
S
p
e
c
i
f
i
c
i
t
y
f
o
r
4
0
%
S
p
e
c
i
f
i
c
i
t
y
f
o
r
5
0
%
S
p
e
c
i
f
i
c
i
t
y
f
o
r
6
0
%
S
p
e
c
i
f
i
c
i
t
y
f
o
r
7
0
%
S
p
e
c
i
f
i
c
i
t
y
f
o
r
8
0
%
S
p
e
c
i
f
i
c
i
t
y
f
o
r
9
0
%
P
S
O
+
N
S
A
7
0
.
4
5
7
1
.
6
7
2
.
3
6
7
3
.
5
8
7
7
.
0
7
7
8
.
4
8
P
S
O
+
C
S
A
7
3
.
0
1
7
3
.
3
1
7
4
.
6
2
7
7
.
2
3
7
9
.
9
2
8
4
.
6
6
A
C
O
+
N
S
A
7
3
.
5
3
7
8
.
6
6
8
0
.
5
8
8
2
.
2
8
8
5
.
6
3
8
9
.
1
6
A
C
O
+
C
S
A
7
5
.
8
4
8
0
.
7
3
8
0
.
8
8
8
1
.
9
5
8
6
.
8
8
9
0
.
4
7
A
B
C
+
N
S
A
7
7
.
7
8
8
1
.
9
1
8
4
.
6
9
8
5
.
6
2
8
8
.
2
7
9
0
.
5
4
A
B
C
+
C
S
A
7
9
.
3
6
8
3
.
8
8
6
.
6
2
8
8
.
4
6
8
9
.
2
9
1
.
7
3
P
S
O
+
D
C
A
8
2
.
5
8
8
4
.
7
9
0
.
8
4
9
1
.
9
7
9
3
.
1
1
9
2
.
2
4
R
O
A
+
d
e
n
d
r
i
t
i
c
-
SSA
-
b
a
se
d
A
I
C
8
5
.
6
7
8
7
.
8
9
0
.
9
4
9
2
.
0
7
9
3
.
2
9
4
.
0
4
P
R
M
I
+
d
e
n
d
r
i
t
i
c
-
SSA
-
b
a
se
d
A
I
C
8
8
.
7
7
8
9
.
9
9
1
.
0
3
9
2
.
1
6
9
3
.
3
9
5
.
4
4
3
.
3
.
2
.
E
x
perim
ent
a
t
io
ns
us
in
g
K
–
f
o
ld cr
o
s
s
v
a
lid
a
t
io
n o
n
im
ba
la
nced
da
t
a
s
et
T
a
b
le
7
s
h
o
ws
t
h
e
o
u
tc
o
m
e
s
o
f
1
0
-
f
o
l
d
an
al
y
s
es
f
o
r
d
i
f
f
er
e
n
t
t
ec
h
n
i
q
u
es
o
n
th
e
B
R
AT
S
a
n
d
SimBR
AT
S
d
atas
ets
.
F
o
r
e
ac
h
ap
p
r
o
ac
h
o
n
b
o
th
d
at
asets
,
t
h
e
ac
c
u
r
a
cy
(
AC
C
)
,
s
e
n
s
it
iv
it
y
(
S
E
N)
,
an
d
s
p
ec
i
f
i
cit
y
(
SP
E
)
p
e
r
c
en
t
a
r
e
in
cl
u
d
e
d
i
n
th
is
an
al
y
s
is
.
T
h
e
1
0
-
f
o
ld
an
a
ly
s
is
,
o
u
tco
m
es
r
ev
ea
l
th
at
th
e
PR
MI
+
d
en
d
r
itic
-
SSA
-
b
ased
AI
C
m
eth
o
d
r
eliab
ly
o
u
tp
er
f
o
r
m
s
o
th
e
r
s
o
n
b
o
th
th
e
im
b
alan
ce
d
B
R
AT
S
an
d
im
b
alan
ce
d
SimB
R
AT
S
d
atasets
,
attain
i
n
g
th
e
m
ax
im
u
m
ac
c
u
r
ac
y
(
9
6
.
2
5
%
an
d
9
6
.
3
5
%,
r
esp
e
ctiv
ely
)
alo
n
g
with
ad
m
ir
ab
le
s
en
s
itiv
ity
an
d
s
p
e
cif
icity
(
ar
o
u
n
d
9
4
-
9
5
%).
Oth
er
m
eth
o
d
s
lik
e
PS
O+
DC
A
a
n
d
R
OA+
d
en
d
r
itic
-
SSA
-
b
ased
AI
C
al
s
o
d
em
o
n
s
tr
ate
r
o
b
u
s
t
p
e
r
f
o
r
m
an
ce
with
a
cc
u
r
ac
ies
o
v
er
9
4
%,
b
u
t
th
e
y
f
all
s
h
o
r
t
co
m
p
a
r
ed
to
PR
MI
+
d
en
d
r
itic
-
SS
A.
Me
th
o
d
s
s
u
ch
as
AB
C
+CS
A
an
d
AC
O+
C
SA
ex
h
ib
it
ex
ten
s
iv
e
en
h
an
ce
m
e
n
ts
o
v
er
PS
O+
NS
A
an
d
PS
O+
C
SA,
w
ith
ac
cu
r
ac
ies
v
ar
y
in
g
f
r
o
m
8
9
%
to
o
v
e
r
9
2
%,
in
d
icatin
g
r
o
b
u
s
t
class
if
icatio
n
co
m
p
eten
cies.
Ov
er
all,
t
h
e
PR
MI
+
d
en
d
r
itic
-
SSA
-
b
ased
AI
C
m
eth
o
d
p
r
o
v
es
to
b
e
a
d
v
an
ce
d
,
ex
h
i
b
itin
g
s
u
p
er
io
r
class
if
icatio
n
ac
cu
r
ac
y
an
d
s
en
s
ib
le
s
en
s
itiv
ity
an
d
s
p
ec
if
icity
th
r
o
u
g
h
b
o
th
d
atas
ets.
Her
e
is
a
s
u
m
m
ar
y
o
f
th
e
f
in
d
in
g
s
:
T
h
e
ap
p
r
o
ac
h
es
r
eg
u
lar
ly
p
er
f
o
r
m
well
in
ter
m
s
o
f
ac
cu
r
ac
y
,
s
en
s
itiv
ity
,
an
d
s
p
ec
if
icity
f
o
r
th
e
B
R
AT
S
d
ata
s
et.
W
i
th
a
9
6
.
2
5
%
ac
cu
r
ac
y
r
ate,
PR
MI
+
d
en
d
r
itic
-
SSA
-
b
ased
AI
C
h
as
th
e
b
est
ac
cu
r
ac
y
.
T
h
e
m
ax
im
u
m
s
en
s
itiv
ity
is
PR
MI
+
d
en
d
r
itic
-
SSA
-
b
ased
AI
C
,
with
v
alu
es
r
an
g
in
g
f
r
o
m
7
9
.
6
9
%
to
9
4
.
3
3
%.
T
h
e
r
an
g
e
o
f
s
p
ec
if
icity
v
alu
es
is
8
1
.
4
%
to
9
4
.
8
5
%,
with
th
e
h
ig
h
e
s
t
s
p
ec
if
icity
b
ein
g
PR
MI
+
d
en
d
r
itic
-
SSA
-
b
ased
AI
C
.
T
h
e
m
eth
o
d
s
o
n
th
e
Sim
B
R
AT
S
d
ataset
p
er
f
o
r
m
well
r
eg
ar
d
in
g
ac
cu
r
ac
y
,
s
en
s
itiv
ity
,
an
d
s
p
ec
if
icity
,
ju
s
t
as
th
e
B
R
AT
S
d
ataset.
On
th
is
d
ataset
as
well,
P
R
MI
+
d
en
d
r
itic
-
SSA
-
b
ased
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
P
erfo
r
ma
n
ce
a
n
a
lysi
s
o
f h
yb
r
id
b
io
-
in
s
p
ir
ed
a
l
g
o
r
ith
ms fo
r
cla
s
s
ifyin
g
b
r
a
in
…
(
R
a
h
u
l R
a
mesh
C
h
a
kre
)
6347
AI
C
o
b
tain
s
an
ex
ce
llen
t
ac
cu
r
ac
y
o
f
9
6
.
2
5
%.
T
h
e
m
ax
im
u
m
s
en
s
itiv
ity
i
s
P
R
MI
+
d
en
d
r
itic
-
SSA
-
b
ased
AI
C
,
with
v
alu
es
r
an
g
in
g
f
r
o
m
7
9
.
6
9
%
to
9
4
.
3
3
%.
T
h
e
r
an
g
e
o
f
s
p
ec
if
icity
v
alu
es
is
8
1
.
4
%
to
9
4
.
8
5
%,
with
th
e
h
ig
h
est s
p
ec
if
icity
b
ein
g
PR
MI
+
d
en
d
r
itic
-
SSA
-
b
ased
AI
C
.
T
ab
le
7
.
K
-
f
o
ld
an
al
y
s
is
o
n
im
b
alan
ce
d
B
R
AT
S a
n
d
SimBR
AT
S
d
ataset
wh
en
K=
1
0
f
o
r
A
C
C
(
ac
cu
r
ac
y
)
,
SEN
(
s
en
s
itiv
ity
)
an
d
SP
E
(
s
p
ec
if
icity
)
as p
er
f
o
r
m
an
ce
m
ea
s
u
r
es
10
-
F
o
l
d
a
n
a
l
y
si
s
/
me
t
h
o
d
s
I
mb
a
l
a
n
c
e
d
B
R
A
TS
d
a
t
a
se
t
I
mb
a
l
a
n
c
e
d
S
i
mB
R
A
TS
d
a
t
a
s
e
t
K
=
1
0
A
C
C
(
%)
S
EN
(
%)
S
P
E
(
%)
A
C
C
(
%)
S
EN
(
%)
S
P
E
(
%)
P
S
O
+
N
S
A
8
1
.
3
2
7
9
.
6
9
8
1
.
4
7
8
.
4
0
7
9
.
8
3
7
8
.
4
8
P
S
O
+
C
S
A
8
6
.
5
8
4
.
7
7
8
6
.
5
8
8
4
.
5
8
8
3
.
8
9
8
4
.
6
6
A
C
O
+
N
S
A
8
8
.
9
9
8
7
.
2
1
8
9
.
0
8
8
9
.
0
8
8
6
.
2
9
8
9
.
1
6
A
C
O
+
C
S
A
8
9
.
0
3
8
7
.
2
5
8
9
.
1
2
8
9
.
1
2
8
7
.
3
4
9
0
.
4
7
A
B
C
+
N
S
A
9
1
.
3
5
8
9
.
5
3
9
1
.
4
4
9
1
.
4
5
8
8
.
6
2
9
0
.
5
4
A
B
C
+
C
S
A
9
2
.
5
5
9
0
.
7
9
2
.
6
4
9
2
.
6
4
9
0
.
7
9
9
1
.
7
3
P
S
O
+
D
C
A
9
4
.
0
6
9
2
.
1
7
9
4
.
1
5
9
4
.
1
5
9
2
.
2
7
9
2
.
2
4
R
O
A
+
d
e
n
d
r
i
t
i
c
-
SSA
-
b
a
se
d
A
I
C
9
4
.
8
5
9
2
.
9
6
9
4
.
9
5
9
4
.
9
5
9
2
.
0
5
9
4
.
0
4
P
R
M
I
+
d
e
n
d
r
i
t
i
c
-
SSA
-
b
a
se
d
A
I
C
9
6
.
2
5
9
4
.
3
3
9
4
.
8
5
9
6
.
3
5
9
4
.
4
2
9
5
.
4
4
3
.
3
.
3
.
Co
m
pa
r
a
t
iv
e
dis
cus
s
io
n
Fo
r
th
is
ex
p
er
im
en
tatio
n
,
th
e
im
b
alan
ce
d
d
ataset
is
g
en
er
ated
f
r
o
m
B
R
AT
S
an
d
SimB
R
AT
S.
T
h
e
im
b
alan
ce
d
d
ataset
g
en
e
r
ated
is
u
s
ed
f
o
r
e
x
p
er
im
e
n
tatio
n
with
s
am
p
lin
g
an
d
with
o
u
t
s
am
p
lin
g
.
As
a
p
r
im
ar
y
in
v
esti
g
atio
n
o
n
an
im
b
alan
ce
d
im
ag
e
d
ata
s
et;
to
v
alid
ate
wh
eth
er
tr
ain
in
g
d
ata
s
ize
af
f
e
cts
th
e
class
if
icatio
n
p
er
f
o
r
m
an
ce
,
we
h
a
v
e
r
an
d
o
m
ly
s
am
p
led
th
e
im
ag
es
f
r
o
m
th
e
d
ataset
in
to
4
0
%
s
am
p
les
to
1
0
0
%
s
am
p
les
wh
ile
in
cr
ea
s
in
g
th
e
s
am
p
le
s
ize
b
y
1
0
%
at
ea
ch
s
tep
.
T
h
e
r
esu
lts
ar
e
p
r
esen
ted
in
T
ab
les
1
t
o
6
an
d
g
r
ap
h
ically
r
ep
r
esen
ted
in
Fig
u
r
es
2
an
d
3
.
I
t
is
f
o
u
n
d
th
at
as
th
e
s
ize
o
f
th
e
d
ata
s
am
p
le
is
in
cr
ea
s
e
d
th
e
class
if
icatio
n
p
er
f
o
r
m
an
ce
also
in
cr
ea
s
es.
So
,
it
is
co
n
clu
d
ed
t
h
at
we
n
ee
d
to
u
s
e
all
av
ailab
le
s
am
p
les
in
an
im
b
alan
ce
d
im
ag
e
d
ata
s
et
f
o
r
class
if
icatio
n
.
As
a
f
in
al
in
v
esti
g
atio
n
o
f
o
u
r
p
r
o
p
o
s
ed
m
eth
o
d
to
h
a
n
d
le
an
im
b
alan
ce
d
im
ag
e
d
ataset,
we
h
av
e
ap
p
lied
a
co
m
p
letely
im
b
alan
ce
d
d
ataset
(
with
o
u
t
s
am
p
lin
g
)
to
ex
is
tin
g
as
well
as
p
r
o
p
o
s
ed
tech
n
iq
u
es.
T
h
e
r
esu
lts
ar
e
s
h
o
wn
in
T
a
b
le
7
with
1
0
-
f
o
l
d
c
r
o
s
s
-
v
a
l
i
d
a
ti
o
n
.
T
a
b
l
e
8
i
l
l
u
s
t
r
a
t
es
t
h
e
r
e
s
u
l
ts
m
a
i
n
l
y
u
s
i
n
g
t
h
e
p
r
o
p
o
s
e
d
R
OA
+
d
e
n
d
r
i
t
ic
c
e
ll
-
SSA
-
b
as
e
d
A
I
C
a
n
d
p
a
r
t
icl
e
r
i
d
e
r
M
I
+
d
e
n
d
r
i
t
ic
c
e
l
l
-
S
SA
-
b
a
s
e
d
A
I
C
f
o
r
i
m
b
a
l
a
n
c
e
d
B
R
A
T
S
a
n
d
S
im
B
R
A
T
S
d
a
t
as
e
t
.
T
h
e
r
es
u
l
ts
u
s
i
n
g
e
x
is
t
i
n
g
t
e
c
h
n
i
q
u
e
s
n
a
m
e
l
y
PS
O
+N
S
A
,
P
S
O+
C
S
A
,
AC
O
+
NS
A
,
AC
O+
C
S
A
,
A
B
C
+N
SA
,
a
n
d
AB
C
+
C
SA
a
r
e
a
ls
o
p
r
e
s
e
n
t
e
d
i
n
T
ab
l
e
8
.
I
t
i
s
f
o
u
n
d
t
h
at
t
h
e
p
e
r
f
o
r
m
a
n
c
e
o
f
e
x
is
t
i
n
g
t
ec
h
n
i
q
u
e
s
n
a
m
e
l
y
P
SO
+
NS
A
,
P
S
O+
C
S
A
,
AC
O
+
NS
A
,
AC
O+
C
S
A
,
A
B
C
+N
SA
,
a
n
d
A
B
C
+CS
A
i
s
l
o
we
r
o
n
i
m
b
a
l
a
n
c
e
d
d
a
t
a
s
et
w
h
e
r
e
a
s
t
h
e
p
r
o
p
o
s
e
d
m
e
t
h
o
d
s
R
O
A
+
d
e
n
d
r
i
ti
c
S
S
A
A
I
C
[
9
]
a
n
d
p
a
r
t
i
c
l
e
r
i
d
e
r
M
I
+
d
e
n
d
r
it
i
c
c
el
l
-
SSA
-
b
as
e
d
A
I
C
t
h
e
d
e
n
d
r
i
t
i
c
c
e
l
l
-
SS
A
-
A
I
C
h
as
t
h
e
b
e
t
t
e
r
p
er
f
o
r
m
a
n
c
e
i
n
t
e
r
m
s
o
f
a
c
c
u
r
a
c
y
,
s
e
n
s
i
t
i
v
it
y
,
a
n
d
s
p
e
c
i
f
i
c
it
y
w
h
e
n
u
s
e
d
wi
t
h
MRI
f
o
r
b
r
ain
tu
m
o
r
class
if
icatio
n
.
Fig
u
r
e
2
.
An
al
y
s
is
o
f
ex
is
tin
g
an
d
p
r
o
p
o
s
ed
m
eth
o
d
s
u
s
in
g
i
m
b
alan
ce
d
B
R
AT
S d
ataset
70
75
80
85
90
95
10
0
Acc
ura
cy
v
a
lues
E
x
is
t
ing
a
nd
pro
po
s
ed
m
et
ho
ds
A
ccur
acy
on
40%
A
ccur
acy
on
50%
A
ccur
acy
on
60%
A
ccur
acy
on
70%
A
ccu
racy
o
n
8
0%
A
ccur
acy
on
90%
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
e
r
20
24
:
6
3
3
9
-
6
3
5
0
6348
Fig
u
r
e
3
.
An
al
y
s
is
o
f
ex
is
tin
g
an
d
p
r
o
p
o
s
ed
m
eth
o
d
s
u
s
in
g
i
m
b
alan
ce
d
SimBR
AT
S d
atase
t
T
ab
le
8
.
An
aly
s
is
o
f
v
ar
io
u
s
p
er
f
o
r
m
a
n
ce
m
ea
s
u
r
es u
s
in
g
v
a
r
io
u
s
m
eth
o
d
s
f
o
r
b
r
ain
tu
m
o
r
im
ag
e
class
if
icatio
n
o
n
im
b
alan
ce
d
B
R
AT
S a
n
d
SimBR
AT
S d
atasets
M
e
t
h
o
d
s
I
mb
a
l
a
n
c
e
d
B
R
A
TS
d
a
t
a
se
t
I
mb
a
l
a
n
c
e
d
S
i
mB
R
A
TS
d
a
t
a
s
e
t
A
C
C
(
%)
S
EN
(
%)
S
P
E
(
%)
A
C
C
(
%)
S
EN
(
%)
S
P
E
(
%)
R
O
A
+
d
e
n
d
r
i
t
i
c
-
SSA
-
b
a
se
d
A
I
C
9
4
.
8
5
9
2
.
9
6
9
4
.
9
5
9
4
.
9
5
9
2
.
0
5
9
4
.
0
4
P
R
M
I
+
d
e
n
d
r
i
t
i
c
-
SSA
-
b
a
se
d
A
I
C
9
6
.
2
5
9
4
.
3
3
9
4
.
8
5
9
6
.
3
5
9
4
.
4
2
9
5
.
4
4
T
h
e
m
o
s
t
en
h
a
n
ce
d
class
if
icatio
n
p
er
f
o
r
m
a
n
ce
o
f
ar
o
u
n
d
9
6
.
2
5
%
ac
cu
r
ac
y
is
ac
h
iev
ed
d
u
e
to
tr
ain
in
g
AI
C
class
if
ier
s
with
d
en
d
r
itic
ce
ll
-
SS
A
with
p
ar
ticle
r
id
er
MI
,
th
is
m
eth
o
d
h
as
a
h
ig
h
c
o
n
v
er
g
en
ce
m
o
m
en
t
u
m
an
d
ac
cu
r
ac
y
.
Fu
r
th
er
,
th
e
co
m
p
u
tatio
n
al
co
m
p
lex
i
ty
o
f
s
eg
m
en
tatio
n
b
ased
o
n
s
p
ar
s
e
FC
M
is
r
ed
u
ce
d
.
T
h
e
p
ar
ticle
r
id
er
MI
'
s
f
ea
tu
r
e
s
elec
tio
n
m
eth
o
d
s
elec
ts
th
e
m
o
s
t
in
f
o
r
m
ativ
e
f
ea
tu
r
es,
r
esu
ltin
g
in
th
e
h
ig
h
est
class
if
icatio
n
ac
cu
r
ac
y
in
alm
o
s
t
all
m
eth
o
d
s
.
Fu
r
th
er
m
o
r
e,
th
e
AI
C
class
if
ier
is
s
elf
-
o
r
g
a
n
izin
g
an
d
n
ec
ess
itates
g
o
o
d
ch
ar
ac
ter
is
tics
.
As
a
r
esu
lt,
th
e
n
ew
s
tr
ateg
y
o
f
ta
k
in
g
t
h
e
im
b
alan
ce
d
d
ataset
im
p
r
o
v
es
o
v
er
all
s
p
ee
d
wh
ile
r
ed
u
cin
g
er
r
o
r
s
.
Similar
ly
,
th
e
s
ec
o
n
d
p
r
o
p
o
s
ed
m
eth
o
d
R
OA
+d
en
d
r
itic
SS
A
AI
C
al
s
o
g
iv
es
s
im
ilar
r
esu
lts
to
p
ar
ticl
e
r
id
er
MI
+d
e
n
d
r
itic
C
ell
-
SS
A
with
class
if
ica
tio
n
ac
cu
r
ac
y
o
f
ar
o
u
n
d
9
4
.
8
5
%.
W
h
ile
th
e
ex
is
tin
g
7
m
eth
o
d
s
p
r
o
v
id
e
a
m
in
im
u
m
o
f
8
1
.
3
2
to
a
m
ax
im
u
m
o
f
9
4
.
0
6
p
e
r
ce
n
t
class
if
icatio
n
p
er
f
o
r
m
an
ce
in
te
r
m
s
o
f
ac
c
u
r
ac
y
As
s
h
o
wn
in
T
ab
le
8
,
o
n
b
o
th
im
b
alan
ce
d
d
atasets
,
b
o
th
m
eth
o
d
s
PR
MI
+d
en
d
r
itic
-
SS
A
b
ased
AI
C
an
d
R
OA+
d
en
d
r
itic
-
SS
A
b
ase
d
AI
C
d
em
o
n
s
tr
ate
s
tr
o
n
g
p
e
r
f
o
r
m
an
ce
.
W
h
ile
b
o
t
h
tech
n
iq
u
es
m
ain
tain
h
ig
h
lev
els
o
f
s
en
s
itiv
ity
,
PR
MI
+d
en
d
r
itic
-
SSA
-
b
ased
AI
C
ten
d
s
to
h
a
v
e
a
m
in
o
r
ad
v
a
n
tag
e
in
ter
m
s
o
f
ac
cu
r
ac
y
an
d
s
p
ec
if
icity
.
3
.
3
.
4
.
Ana
ly
s
is
ba
s
ed
o
n t
im
e
T
h
e
m
eth
o
d
s
o
n
th
e
u
n
b
alan
c
ed
B
R
AT
S
d
ataset
h
av
e
a
wid
e
r
an
g
e
o
f
ex
ec
u
tio
n
tim
es.
E
x
ec
u
tio
n
tim
es
f
o
r
PS
O+
NSA,
AB
C
+N
SA,
an
d
AB
C
+CS
A
r
an
g
e
f
r
o
m
ab
o
u
t
9
.
3
6
s
ec
o
n
d
s
t
o
1
0
.
6
4
s
ec
o
n
d
s
,
wh
ich
is
co
m
p
ar
ativ
ely
lo
n
g
.
E
x
ec
u
tio
n
tim
es
f
o
r
PS
O+
C
SA,
AC
O+
N
SA,
an
d
AC
O+
C
SA
ar
e
r
ea
s
o
n
ab
le,
r
an
g
in
g
f
r
o
m
5
.
6
4
to
6
.
5
0
s
ec
o
n
d
s
.
T
h
e
e
x
e
cu
tio
n
tim
es o
f
th
e
alg
o
r
ith
m
s
PS
O+
D
C
A,
R
OA+
d
en
d
r
itic
-
S
SA b
ased
AI
C
,
an
d
PR
MI
+
d
en
d
r
itic
-
SSA
b
ased
AI
C
ar
e
all
im
p
r
ess
iv
ely
f
ast
—
u
n
d
er
1
s
ec
o
n
d
.
T
h
e
al
g
o
r
it
h
m
s
'
ex
ec
u
tio
n
tim
es
o
n
th
e
im
b
alan
ce
d
SimBR
AT
S
d
ataset
v
ar
y
as
well.
E
x
e
cu
tio
n
tim
es
f
o
r
PS
O+
NSA
an
d
AB
C
+N
SA
ar
e
r
elativ
ely
lo
n
g
,
at
a
b
o
u
t
8
.
2
1
an
d
7
.
1
9
s
ec
o
n
d
s
,
r
esp
ec
tiv
e
ly
.
T
h
e
h
ig
h
est
ex
e
cu
tio
n
tim
e
o
f
AB
C
+C
SA
is
7
.
8
4
s
ec
o
n
d
s
.
AC
O+
NSA
e
x
ec
u
tes
in
a
r
esp
ec
tab
le
7
.
0
2
s
ec
o
n
d
s
.
E
x
ec
u
tio
n
tim
i
n
g
s
f
o
r
PS
O+
C
S
A,
AC
O+
C
SA,
PS
O+
DC
A,
an
d
R
OA+
d
en
d
r
itic
-
SSA
-
b
ased
AI
C
r
an
g
e
f
r
o
m
2
.
8
8
s
ec
o
n
d
s
t
o
5
.
0
0
s
ec
o
n
d
s
.
T
h
e
f
astes
t
ex
ec
u
tio
n
tim
e
f
o
r
th
is
d
ataset's
P
R
MI
+
d
en
d
r
itic
-
SSA
-
b
ased
AI
C
is
0
.
8
8
s
ec
o
n
d
s
.
T
ab
le
9
illu
s
tr
ates
th
e
an
aly
s
is
b
ased
o
n
th
e
r
eq
u
ir
ed
tim
e
f
o
r
th
e
ex
ec
u
tio
n
wh
il
e
u
s
in
g
th
e
im
b
alan
ce
d
d
ataset.
I
t
is
o
b
s
er
v
e
d
th
a
t
th
e
two
p
r
o
p
o
s
ed
m
et
h
o
d
s
R
OA
d
en
d
r
itic
ce
ll
-
SS
A
b
ased
AI
C
an
d
p
ar
ticle
r
id
er
MI
+
d
en
d
r
itic
ce
ll
-
SS
A
b
ased
AI
C
g
iv
e
b
etter
r
esu
lts
in
ter
m
s
o
f
lo
wer
ex
ec
u
tio
n
tim
e
wh
e
n
co
m
p
a
r
ed
with
ex
is
tin
g
m
eth
o
d
s
.
70
75
80
85
90
95
100
Acc
ura
cy
v
a
lues
E
x
is
t
ing
a
nd
pro
po
s
ed
m
et
ho
ds
A
ccur
acy
on
40%
A
ccur
acy
on
50%
A
ccur
acy
on
60%
A
ccur
acy
on
70%
A
ccur
acy
on
80%
A
ccur
acy
on
90%
Evaluation Warning : The document was created with Spire.PDF for Python.