Co
m
pu
t
er
Science
a
n
d In
f
o
r
m
a
t
io
n T
ec
hn
o
lo
g
ies
Vo
l.
2
,
No
.
3
,
No
v
em
b
er
2
0
2
1
,
p
p
.
1
21
~
1
31
I
SS
N:
2722
-
3221
,
DOI
: 1
0
.
1
1
5
9
1
/
csit
.
v
2
i
3
.p
1
21
-
1
31
121
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ia
e
s
p
r
ime.
co
m/in
d
ex
.
p
h
p
/csi
t
Cla
ss
ificatio
n of
m
a
mm
o
g
ra
m
s
ba
sed o
n f
ea
t
ures ex
traction
techniqu
es using
supp
o
rt
v
ector
ma
chine
E
na
s
M
o
ha
m
m
ed
H
u
s
s
ei
n S
a
ee
d
1
,
H
a
y
der
Adna
n Sa
leh
2
,
a
nd
E
na
m
Aze
z
K
ha
lel
3
1,
2
De
p
a
rtm
e
n
t
o
f
Co
m
p
u
ter S
c
ien
c
e
,
M
u
sta
n
siriy
a
Un
iv
e
rsit
y
,
Ba
g
h
d
a
d
,
Ira
q
3
S
p
e
c
iali
st Rad
io
lo
g
y
De
p
a
rt
m
e
n
t
-
On
c
o
lo
g
y
Tea
c
h
in
g
Ho
sp
it
a
ls,
M
e
d
ica
l
Cit
y
Co
m
p
lex
,
Ba
g
h
d
a
d
,
Ira
q
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
O
ct
1
8
,
2
0
2
0
R
ev
i
s
ed
D
ec
9
,
2
0
2
0
A
cc
ep
ted
M
ar
8
,
2
0
2
1
No
w
m
a
m
m
o
g
ra
p
h
y
c
a
n
b
e
d
e
f
in
e
d
a
s
th
e
m
o
st
re
li
a
b
le
m
e
th
o
d
f
o
r
e
a
rly
b
re
a
st
c
a
n
c
e
r
d
e
tec
ti
o
n
.
T
h
e
m
a
i
n
g
o
a
l
o
f
th
is
stu
d
y
is
to
d
e
sig
n
a
c
las
si
f
ier
m
o
d
e
l
to
h
e
lp
ra
d
i
o
lo
g
ists
t
o
p
ro
v
i
d
e
a
se
c
o
n
d
v
iew
to
d
iag
n
o
se
m
a
m
m
o
g
r
a
m
s.
In
t
h
e
p
r
o
p
o
se
d
sy
ste
m
m
e
d
iu
m
f
i
lt
e
r
a
n
d
b
i
n
a
ry
i
m
a
g
e
w
it
h
a
g
lo
b
a
l
th
re
sh
o
l
d
h
a
v
e
b
e
e
n
a
p
p
li
e
d
f
o
r
re
m
o
v
in
g
th
e
n
o
ise
a
n
d
sm
a
ll
a
rti
f
a
c
ts
i
n
th
e
p
re
-
p
ro
c
e
ss
in
g
sta
g
e
.
S
e
c
o
n
d
ly
,
in
t
h
e
se
g
m
e
n
tatio
n
p
h
a
se
,
a
Hy
b
rid
Bo
u
n
d
i
n
g
Bo
x
a
n
d
Re
g
io
n
G
ro
w
in
g
(HBBRG
)
a
l
g
o
rit
h
m
a
re
u
ti
li
z
in
g
to
re
m
o
v
e
p
e
c
to
ra
l
m
u
sc
le
s,
a
n
d
th
e
n
a
g
e
o
m
e
tri
c
m
e
th
o
d
h
a
s
b
e
e
n
a
p
p
li
e
d
to
c
u
t
th
e
larg
e
st
p
o
ss
ib
le
sq
u
a
re
th
a
t
c
a
n
b
e
o
b
tai
n
e
d
f
ro
m
a
m
a
m
m
o
g
ra
m
w
h
ich
re
p
re
se
n
ts
th
e
ROI.
In
t
h
e
f
e
a
tu
re
s
e
x
tra
c
ti
o
n
p
h
a
se
th
re
e
m
e
th
o
d
w
a
s
u
se
d
to
p
re
p
a
re
tex
tu
re
f
e
a
tu
re
s
to
b
e
a
su
it
a
b
le
in
tr
o
d
u
c
ti
o
n
to
th
e
c
las
sif
ica
ti
o
n
p
ro
c
e
ss
a
re
f
irs
t
Ord
e
r
(sta
ti
stica
l
fe
a
tu
re
s),
L
o
c
a
l
Bin
a
r
y
P
a
tt
e
rn
s
(L
BP
),
a
n
d
G
ra
y
-
L
e
v
e
l
Co
-
Oc
c
u
rre
n
c
e
M
a
tri
x
(
GL
CM
),
F
in
a
ll
y
,
S
V
M
h
a
s
b
e
e
n
a
p
p
li
e
d
in
t
w
o
-
lev
e
l
to
c
las
si
fy
m
a
m
m
o
g
ra
m
i
m
a
g
e
s
in
th
e
f
irst
lev
e
l
to
n
o
rm
a
l
o
r
a
b
n
o
rm
a
l,
a
n
d
t
h
e
n
th
e
c
las
sif
ica
ti
o
n
o
f
a
b
n
o
rm
a
l
o
n
c
e
in
th
e
se
c
o
n
d
lev
e
l
t
o
t
h
e
b
e
n
ig
n
o
r
m
a
li
g
n
a
n
t
im
a
g
e
.
T
h
e
s
y
st
e
m
w
a
s
tes
ted
o
n
t
h
e
M
A
IS
th
e
M
a
m
m
o
g
ra
m
i
m
a
g
e
a
n
a
ly
sis
S
o
c
iet
y
(M
IA
S
)
d
a
tab
a
se
,
in
a
d
d
it
i
o
n
to
t
h
e
im
a
g
e
f
ro
m
th
e
T
e
a
c
h
in
g
On
c
o
lo
g
y
Ho
sp
it
a
l,
M
e
d
ica
l
Cit
y
in
Ba
g
h
d
a
d
,
w
h
e
re
th
e
re
s
u
lt
s
s
h
o
w
e
d
a
c
h
iev
in
g
a
n
a
c
c
u
ra
c
y
o
f
9
5
.
4
5
4
%
f
o
r
th
e
f
irst
lev
e
l
a
n
d
9
7
.
2
6
0
%
f
o
r
th
e
se
c
o
n
d
lev
e
l,
a
lso
,
t
h
e
re
su
lt
s
o
f
a
p
p
ly
in
g
th
e
p
r
o
p
o
se
d
sy
ste
m
to
th
e
M
IA
S
d
a
tab
a
se
a
lo
n
e
w
e
re
a
c
h
iev
in
g
a
n
a
c
c
u
ra
c
y
o
f
9
3
.
1
0
5
%
f
o
r
th
e
f
irst
lev
e
l
a
n
d
9
4
.
5
9
f
o
r
t
h
e
se
c
o
n
d
lev
e
l.
K
ey
w
o
r
d
s
:
B
r
ea
s
t c
an
ce
r
C
las
s
i
f
icatio
n
C
u
t
t
h
e
lar
g
es
t
p
o
s
s
ib
le
s
q
u
a
r
e
Featu
r
es e
x
tr
ac
tio
n
Dig
ital
m
a
m
m
o
g
r
ap
h
y
Su
p
p
o
r
t v
ec
to
r
m
ac
h
i
n
e
(
SV
M)
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
:
E
n
as M
o
h
a
m
m
ed
Hu
s
s
ei
n
Sae
ed
De
p
a
rtme
n
t
o
f
C
o
m
p
u
ter
Scien
ce
Mu
s
ta
n
s
ir
i
y
a
U
n
i
v
er
s
it
y
B
ag
h
d
ad
,
I
r
aq
E
m
ail: e
n
as.saee
d
3
8
@
g
m
ail.
c
o
m
1.
I
NT
RO
D
UCT
I
O
N
B
r
ea
s
t
ca
n
ce
r
is
th
e
m
o
s
t
w
id
esp
r
ea
d
d
is
ea
s
e
th
at
a
f
f
ec
t
s
f
e
m
ales
d
u
r
in
g
t
h
eir
lif
e
ti
m
e,
an
d
th
e
2
n
d
lead
in
g
d
ea
t
h
ca
u
s
e
f
o
r
w
o
m
e
n
g
lo
b
all
y
[
1
]
ac
co
r
d
in
g
to
th
e
m
o
s
t r
ec
e
n
t stat
is
tical
d
ata
th
at
h
as b
ee
n
r
elea
s
ed
in
2
0
1
5
in
t
h
e
U
S
A
,
b
r
ea
s
t
ca
n
ce
r
o
cc
u
p
ies
2
9
%
o
f
th
e
n
e
w
c
ases
o
f
ca
n
ce
r
a
n
d
1
5
%
o
f
th
e
ca
n
ce
r
d
ea
th
ca
s
e
s
[
2
]
.
E
ar
ly
d
etec
tio
n
is
th
e
o
p
ti
m
al
o
p
tio
n
f
o
r
in
cr
ea
s
i
n
g
tr
ea
t
m
en
t
o
p
tio
n
s
,
as
t
h
er
e
ar
e
m
a
n
y
s
cr
ee
n
in
g
m
et
h
o
d
s
f
o
r
b
r
ea
s
t
ca
n
ce
r
d
etec
tio
n
s
u
ch
as
b
io
p
s
y
,
MRI,
u
ltra
s
o
u
n
d
,
an
d
m
a
m
m
o
g
r
ap
h
y
[
3
]
.
Ma
m
m
o
g
r
ap
h
y
i
s
co
n
s
id
er
ed
a
co
m
m
o
n
m
et
h
o
d
f
o
r
d
etec
ti
ng
ab
n
o
r
m
al
ities
i
n
th
e
ea
r
l
y
s
ta
g
es.
Ma
m
m
o
g
r
ap
h
y
is
a
lo
w
-
d
o
s
e
X
-
r
ay
ap
p
licatio
n
allo
w
i
n
g
to
v
i
s
u
alize
t
h
e
i
n
n
er
b
r
ea
s
t
s
tr
u
c
t
u
r
e
[
4
]
.
I
t’
s
u
s
u
all
y
co
n
tai
n
ed
m
an
y
ar
t
i
f
ac
ts
an
d
n
o
is
es
t
h
at
m
a
k
e
th
e
m
d
i
f
f
ic
u
l
t
to
u
n
d
er
s
ta
n
d
in
t
h
e
i
n
itial
s
t
ag
es.
T
h
er
ef
o
r
e,
s
tan
d
ar
d
i
zin
g
i
m
a
g
e
q
u
alit
y
an
d
ex
tr
ac
ti
n
g
R
OI
is
n
ec
es
s
ar
y
to
r
ed
u
ce
t
h
e
s
ea
r
c
h
f
o
r
d
is
to
r
tio
n
s
[
5
]
.
Vi
s
u
al
ex
a
m
i
n
atio
n
o
f
m
a
m
m
o
g
r
a
m
s
b
y
a
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2722
-
3221
C
o
m
p
u
t.
Sci.
I
n
f
.
T
ec
h
n
o
l.
,
Vo
l.
2
,
No
.
3
,
No
v
em
b
er
20
2
1
:
1
21
–
1
31
122
r
ad
io
lo
g
is
t
to
d
etec
t
b
r
ea
s
t
ca
n
ce
r
v
er
y
co
m
m
a
n
d
b
u
t
s
o
m
e
ti
m
e
s
lead
s
to
less
ac
cu
r
ate
d
iag
n
o
s
is
,
d
u
e
to
th
e
s
tr
ess
o
f
t
h
e
r
ad
io
lo
g
is
t
a
n
d
l
o
w
i
m
a
g
e
q
u
ali
t
y
.
T
h
e
s
t
u
d
ie
s
in
d
icate
th
at
th
e
er
r
o
r
r
ate
o
f
1
0
%
to
3
0
%
f
o
r
d
iag
n
o
s
e
s
o
f
m
a
lig
n
a
n
t
m
as
s
es
b
y
a
r
ad
io
lo
g
i
s
t
w
h
o
u
tili
ze
s
v
i
s
u
al
i
n
s
p
ec
tio
n
.
T
h
e
er
r
o
r
r
ate
d
ec
r
ea
s
es
s
ig
n
i
f
ica
n
tl
y
b
y
u
tili
zi
n
g
t
h
e
c
o
m
p
u
ter
-
a
id
ed
d
etec
tio
n
C
A
D
e
s
y
s
te
m
,
w
h
ic
h
p
r
o
v
id
es
t
h
e
s
u
p
p
o
r
t
o
f
th
e
f
i
n
al
d
ec
is
io
n
a
n
d
is
co
n
s
id
er
ed
as
a
s
ec
o
n
d
o
p
in
io
n
w
i
th
r
ad
io
lo
g
y
s
p
ec
iali
s
ts
d
iag
n
o
s
is
to
cla
s
s
i
f
y
b
r
ea
s
t
t
u
m
o
r
s
[6
]
,
[
7
]
.
No
r
m
all
y
,
t
h
er
e
ar
e
f
iv
e
p
h
a
s
es
o
f
t
h
e
C
A
De
s
y
s
te
m
f
o
r
b
r
ea
s
t
ca
n
ce
r
d
etec
tio
n
as
ca
n
b
e
s
ee
n
i
n
Fig
u
r
e.
1
,
[
8
]
.
F
ig
u
re
1
.
G
e
n
e
ra
l
p
h
a
se
f
o
r
CA
D
e
s
y
ste
m
[
8
]
T
o
ac
h
iev
e
th
is
g
o
al,
t
h
e
m
ed
i
u
m
f
i
lter
w
as
ap
p
lied
to
r
e
m
o
v
e
n
o
i
s
e,
an
d
lo
ca
l
co
n
tr
as
t,
an
d
th
e
n
t
h
e
b
in
ar
y
i
m
a
g
e
w
it
h
a
g
lo
b
al
th
r
esh
o
ld
ap
p
lied
f
o
r
s
m
all
ar
ti
f
ac
t
s
r
e
m
o
v
al.
I
n
t
h
e
s
e
g
m
en
ta
tio
n
p
h
ase,
an
HB
B
R
G
alg
o
r
ith
m
i
s
u
t
ilizi
n
g
to
r
e
m
o
v
e
th
e
p
ec
to
r
al
m
u
s
cle
s
[
9
]
,
an
d
th
en
w
e
f
i
n
d
t
h
e
lar
g
e
s
t
p
o
s
s
i
b
le
s
q
u
ar
e
ar
ea
th
at
ca
n
b
e
o
b
tain
ed
f
r
o
m
th
e
b
r
ea
s
t,
w
h
ic
h
r
ep
r
esen
ts
th
e
R
OI
[
1
0
]
.
I
n
th
e
f
ea
t
u
r
es
ex
tr
ac
tio
n
th
r
ee
f
ea
t
u
r
e
ex
tr
a
ctio
n
tec
h
n
iq
u
es
w
er
e
u
s
ed
ar
e
f
ir
s
t
Or
d
er
,
GL
C
M,
a
n
d
L
B
P
,
to
ac
q
u
ir
e
s
tr
o
n
g
tex
tu
r
e
f
ea
tu
r
e
s
th
a
t
w
er
e
en
ter
ed
in
to
th
e
SVM
clas
s
i
f
ie
r
at
th
e
class
i
f
icat
io
n
p
h
a
s
e.
T
o
co
n
d
u
ct
th
i
s
s
ea
r
c
h
,
in
v
esti
g
at
e
it,
an
d
ev
alu
a
tin
g
all
its
p
h
ase
s
,
a
s
p
ec
ial
d
atab
ase
w
a
s
estab
li
s
h
ed
t
h
at
r
elies
m
ai
n
l
y
o
n
MI
AS
w
h
ich
h
as
b
ee
n
p
r
o
p
o
s
ed
b
y
t
h
e
U.
K.
n
atio
n
al
p
r
o
g
r
a
m
o
f
b
r
ea
s
t
s
cr
ee
n
i
n
g
.
T
h
is
d
atab
ase
in
clu
d
es
3
2
2
d
ig
itized
m
a
m
m
o
g
r
a
m
s
,
1
1
4
ab
n
o
r
m
al
,
an
d
2
0
8
n
o
r
m
al
w
h
ich
s
u
b
-
d
iv
id
in
g
to
5
1
m
alig
n
a
n
t
an
d
6
3
B
en
ig
n
.
A
1
0
2
4
x
1
0
2
4
p
ix
el
im
ag
e
w
i
th
a
"
P
GM
"
f
o
r
m
at
[
1
4
]
.
T
h
e
d
atab
ase
is
av
ailab
le
o
n
t
h
e
w
eb
s
ite
h
ttp
://p
eip
a.
ess
ex
.
ac
.
u
k
/
in
f
o
/
m
ias.
h
t
m
l
[
12
]
.
A
ls
o
in
co
o
r
d
in
atio
n
w
it
h
t
h
e
T
ea
ch
in
g
On
co
lo
g
y
Ho
s
p
ital
/
Me
d
ica
l
C
it
y
/
B
ag
h
d
ad
,
a
s
et
o
f
i
m
ag
es
w
a
s
o
b
tain
ed
an
d
ad
d
ed
to
th
e
d
a
tab
ase.
T
h
e
p
r
e
-
p
r
o
ce
s
s
i
n
g
o
f
t
h
ese
i
m
a
g
es
h
as
al
s
o
b
ee
n
p
er
f
o
r
m
ed
to
b
e
in
t
h
e
s
a
m
e
MI
AS
d
atab
ase
i
m
ag
e
s
p
ec
if
icat
io
n
.
T
h
eo
r
etica
l c
o
n
s
id
er
atio
n
So
m
e
m
an
y
tech
n
o
lo
g
ies
an
d
alg
o
r
ith
m
s
h
av
e
b
ee
n
in
clu
d
ed
in
th
e
s
tag
es
o
f
th
e
C
A
D
s
y
s
tem
in
th
i
s
p
ar
t.
W
e
w
il
l
lo
o
k
at
s
o
m
e
o
f
th
e
tech
n
iq
u
es
th
at
w
er
e
u
s
ed
in
o
u
r
p
r
o
p
o
s
ed
m
eth
o
d
to
ex
tr
ac
t
an
d
s
elec
t
f
ea
tu
r
es.
First o
r
d
er
(
s
tatis
tical)
f
ea
t
u
r
e
s
Statis
tical
tex
tu
r
e
an
aly
s
es
h
av
e
b
ee
n
b
ased
u
p
o
n
th
e
s
tatis
tical
ch
ar
ac
ter
is
tics
o
f
th
e
in
ten
s
i
t
y
h
is
to
g
r
am
w
ith
n
o
co
n
s
id
er
atio
n
o
f
th
e
s
p
atial
d
ep
en
d
en
ce
s
.
T
h
e
im
ag
e
h
is
to
g
r
am
h
as
g
iv
en
a
s
tatis
tica
l
in
f
o
r
m
atio
n
s
u
m
m
ar
y
co
n
ce
r
n
in
g
th
is
im
ag
e.
T
h
e
1
s
t
o
r
d
er
s
tatis
tical
im
ag
e
in
f
o
r
m
atio
n
m
ay
b
e
p
r
o
d
u
ce
d
w
i
t
h
th
e
u
s
e
o
f
an
im
ag
e
h
is
to
g
r
am
[1
3
]
.
I
t
is
a
g
r
o
u
p
o
f
th
e
u
s
ef
u
l
f
ea
tu
r
es
th
at
m
ay
b
e
d
ir
ec
tly
ex
tr
ac
ted
f
r
o
m
th
e
s
p
atial
d
o
m
ain
o
f
th
e
im
ag
e
h
is
to
g
r
am
b
ased
o
n
p
ix
el
v
alu
es
o
n
ly
,
in
clu
d
in
g
m
ea
n
,
SD,
v
ar
ian
ce
,
Ku
r
to
s
is
,
an
d
Sk
ew
n
ess
[1
4
]
.
T
h
ese
f
ea
tu
r
es h
av
e
b
ee
n
ca
lcu
lated
u
s
in
g
th
e
f
o
llo
w
in
g
eq
u
atio
n
s
[
15
]
,
[
16
]
:
Me
an
=
g
∑
P
(
g
)
L
−
1
g
=
0
(
1
)
W
h
er
e
g
is
g
r
e
y
lev
el
in
th
e
im
ag
e
(
0
to
2
5
5
)
Stan
d
ar
d
Dev
iatio
n
=
√
∑
(
g
−
g
̅
)
2
P
(
g
)
L
−
1
g
=
0
2
(
2
)
W
h
er
e
g
̅
is
m
ea
n
o
f
g
r
ay
lev
el
in
th
e
im
ag
e
Evaluation Warning : The document was created with Spire.PDF for Python.
C
o
m
p
u
t.
Sci.
I
n
f
.
T
ec
h
n
o
l.
C
la
s
s
i
fica
tio
n
o
f m
a
mmo
g
r
a
m
s
b
a
s
ed
o
n
fea
t
u
r
es e
xtra
ctio
n
tech
n
iq
u
es
… (
E
n
a
s
Mo
h
a
mme
d
Hu
s
s
ein
S
a
ee
d
)
1
23
Va
r
ia
n
c
e
=
∑
(
g
−
g
̅
)
2
P
(
g
)
L
−
1
g
=
0
(
3
)
Sk
e
wn
e
s
s
=
∑
(
g
−
g
̅
)
3
P
(
g
)
L
−
1
g
=
0
(
4
)
Kur
tosis
=
∑
(
g
−
g
̅
)
4
P
(
g
)
L
−
1
g
=
0
(
5
)
L
o
ca
l
b
in
ar
y
p
atter
n
s
(
L
B
P
)
f
ea
tu
r
es
T
h
is
tex
tu
r
e
d
escr
ip
to
r
is
v
er
y
in
ter
esti
n
g
it
is
tex
tu
r
e
d
escr
ip
to
r
is
co
n
s
id
er
v
er
y
in
ter
esti
n
g
ly
b
ec
au
s
e
p
ar
ticu
lar
ly
s
u
itab
le
f
o
r
r
ea
l
-
tim
e
q
u
ality
co
n
tr
o
llin
g
ap
p
licatio
n
s
b
ec
au
s
e
it
i
s
f
ast
an
d
ea
s
y
to
im
p
lem
en
t
.
I
t
w
as
ex
p
an
d
ed
f
o
r
th
e
co
lo
r
im
ag
e
b
y
Ma
en
p
aa
an
d
P
ietik
¨
ain
en
an
d
u
s
ed
in
n
u
m
er
o
u
s
ap
p
licatio
n
s
to
class
i
f
y
p
r
o
b
lem
s
b
ased
o
n
co
lo
r
im
ag
es
[1
7
]
,
[
18
]
.
L
B
P
in
d
icate
s
a
r
elatio
n
s
h
ip
b
etw
ee
n
a
ce
n
tr
al
p
ix
el
an
d
its
ad
j
ac
en
t
p
ix
els
in
Mic
r
o
p
atter
n
L
B
P
ex
am
in
ed
w
in
d
o
w
to
ce
lls
(
f
o
r
ex
am
p
le
8
x
8
an
d
1
6
x
1
6
p
ix
els
f
o
r
ea
ch
ce
ll).
o
r
ev
er
y
o
n
e
o
f
th
e
p
ix
els
in
a
ce
ll,
co
m
p
r
ess
io
n
to
all
p
ix
els
o
f
its
eig
h
t
n
eig
h
b
o
r
s
(
o
n
th
e
u
p
p
er
lef
t,
m
id
d
le
lef
t,
lo
w
er
lef
t,
r
ig
h
t
to
p
,
an
d
s
o
o
n
)
as
s
h
o
w
n
in
F
ig
u
r
e
2
w
h
ich
s
h
o
w
s
a
n
ex
am
p
le
o
f
th
e
o
r
ig
in
al
o
p
er
ato
r
o
f
th
e
L
B
P
.
L
P
B
f
o
llo
w
p
ix
els
alo
n
g
th
e
cir
cle,
th
at
is
,
co
u
n
ter
clo
ck
w
is
e
o
r
clo
ck
w
is
e.
F
ig
u
re
2.
A
n
e
x
a
m
p
le o
f
th
e
o
rig
in
a
l
o
p
e
ra
to
r
o
f
th
e
L
B
P
[
19
]
T
h
e
s
tr
in
g
‘
1000
0
111
’
is
g
etti
n
g
f
o
r
3
*
3
b
lo
ck
w
ith
th
e
ce
n
t
r
al
p
ix
el
5
.
T
h
e
b
in
ar
y
f
o
r
m
is
tr
an
s
f
o
r
m
ed
to
its
135
d
u
p
licate
s
in
a
d
ec
im
al
f
o
r
m
.
L
B
P
h
is
to
g
r
am
s
ar
e
cr
ea
ted
f
r
o
m
all
m
icr
o
p
atter
n
s
d
ep
en
d
o
n
a
d
ec
im
al
v
alu
e.
A
s
s
u
m
e
th
at
I
is
an
im
ag
e
in
ten
s
ity
an
d
r
=
(
x
,
y
)
ᵀ
is
a
v
ec
to
r
o
f
p
o
s
itio
n
in
I
.
T
h
e
L
B
P
b
(
r
∈
R
ᴺ
ⁿ)
is
k
n
o
w
n
as in
th
e
f
o
llo
w
in
g
d
escr
ip
tio
n
:
B
i(
r
)
=1
:
if
I
(
r
)
<I
(
r
+Δ
s
i)
,
0
:
o
th
er
w
is
e,
(
i
=1
Nn
)
.
Nn
r
ep
r
esen
ts
th
e
n
u
m
b
er
o
f
th
e
n
eig
h
b
o
r
in
g
p
ix
els,
an
d
Δ
s
i
is
v
ec
to
r
s
of
d
is
p
lace
m
en
t
f
r
o
m
th
e
p
o
s
itio
n
o
f
ce
n
ter
p
ix
el
r
to
n
eig
h
b
o
r
in
g
p
ix
els
[
1
9
]
.
Gr
a
y
lev
el
co
-
o
cc
u
r
r
en
ce
s
m
at
r
ix
(
GL
C
M)
f
ea
t
u
r
es
GL
C
M
is
also
a
s
tatis
tica
l
m
e
th
o
d
th
at
tak
es
i
n
to
ac
co
u
n
t
t
h
e
s
p
atial
r
elatio
n
s
h
ip
b
et
w
ee
n
p
ix
els.
I
t
h
as
b
ee
n
e
m
p
lo
y
ed
w
id
el
y
in
n
u
m
er
o
u
s
ap
p
licatio
n
s
b
a
s
ed
o
n
tex
t
u
r
e
an
al
y
s
is
i
n
th
e
ar
ea
o
f
tex
tu
r
e
an
a
l
y
s
is
[
20
]
.
T
h
e
p
r
o
b
ab
ilit
y
o
f
a
p
ix
e
l
w
it
h
a
g
r
a
y
lev
el
o
f
I
o
cc
u
r
r
in
g
i
n
ter
m
s
o
f
a
p
ar
ticu
lar
s
p
atial
r
elatio
n
s
h
ip
to
an
o
th
er
p
ix
el
j
ca
n
al
w
a
y
s
b
e
ca
lcu
lated
w
it
h
G
L
C
M.
T
h
e
s
ize
o
f
th
e
G
L
C
M
i
s
lo
ca
ted
b
y
h
o
w
m
an
y
p
ix
el
s
ar
e
f
o
u
n
d
i
n
an
i
m
a
g
e.
T
h
e
GL
C
M
f
ea
t
u
r
e
is
ap
p
lied
b
ased
o
n
f
o
u
r
a
n
g
le
s
(
0
°,
4
5
°,
9
0
°,
1
3
5
°
d
eg
r
ee
s
)
an
d
d
is
p
lace
m
e
n
t
d
is
ta
n
ce
as
s
h
o
wn
i
n
F
i
g
u
r
e
3
[
21
]
.
I
n
th
is
p
ap
er
,
f
o
u
r
f
ea
t
u
r
es
e
x
tr
ac
ted
f
r
o
m
GL
C
M
ar
e
e
n
er
g
y
,
co
n
tr
ast,
co
r
r
elatio
n
,
a
n
d
h
o
m
o
g
en
eit
y
.
T
h
o
s
e
f
ea
t
u
r
es
h
a
v
e
b
ee
n
ca
lcu
lated
u
tili
ze
d
by
t
h
e
f
o
llo
w
in
g
(
6
)
.
E
n
e
r
gy
=
∑
p
(
i
,
j
)
2
n
i
,
j
=
1
(
6
)
W
h
er
e
n
is
t
h
e
Co
-
Occ
u
r
r
en
ce
m
atr
ix
d
i
m
e
n
s
io
n
,
p
is
t
h
e
p
r
o
b
ab
ilit
y
o
f
G
L
C
M
.
C
on
tr
a
s
t
=
∑
|
i
−
j
|
p
(
i
,
j
)
n
i
,
j
=
1
(
7
)
C
or
r
e
l
a
tion
=
∑
(
i
−
μ
i
)
(
j
−
μ
j
)
p
(
i
,
j
)
σ
i
σ
j
n
i
,
j
=
1
(
8
)
W
h
er
e
μ
x
,
μ
y
,
σ
x
an
d
σ
y
r
ep
r
esen
t
t
h
e
m
ea
n
a
n
d
s
tan
d
ar
d
d
ev
iatio
n
v
al
u
es.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2722
-
3221
C
o
m
p
u
t.
Sci.
I
n
f
.
T
ec
h
n
o
l.
,
Vo
l.
2
,
No
.
3
,
No
v
em
b
er
20
2
1
:
1
21
–
1
31
124
Homoge
n
e
ity
=
∑
p
(
i
,
j
)
1
+
(
i
−
j
)
2
n
i
,
j
=
1
(
10
)
[
22
]
.
F
ig
u
re
3
.
S
h
o
w
a
d
jac
e
n
c
y
o
f
p
ix
e
l
in
f
o
u
r
d
irec
ti
o
n
s
[
21
]
R
elate
d
w
o
r
k
s
Ma
n
y
r
esear
ch
er
s
co
m
p
leted
ea
r
ly
to
d
iag
n
o
s
e
b
r
ea
s
t
ca
n
ce
r
in
an
attem
p
t
to
h
elp
r
ad
io
lo
g
is
ts
to
d
etec
t
ab
n
o
r
m
al
tis
s
u
e
in
th
e
f
o
r
m
o
f
m
am
m
o
g
r
a
m
s
;
w
e
w
ill
r
ev
iew
th
e
m
o
s
t
s
ig
n
if
ican
t
s
tu
d
ies
in
th
is
ar
ea
b
elo
w
:
I
n
2017
[
2
3
]
Har
ef
a,
et
a
l
,
ap
p
lied
to
p
r
e
-
p
r
o
ce
s
s
f
o
r
im
p
r
o
v
in
g
th
e
im
ag
e
q
u
ality
,
an
d
th
en
th
e
s
eg
m
en
ta
t
i
o
n
s
tag
e
h
as
b
ee
n
ap
p
lied
d
ep
en
d
in
g
o
n
th
e
d
atab
ase
to
o
b
tain
th
e
R
OI
.
Featu
r
es
ar
e
ex
tr
ac
ted
b
y
u
s
in
g
GL
C
M
at
0
o
,
4
5
o
,
9
0
o
,
an
d
1
3
5
o
w
ith
a
1
2
8
x
1
2
8
b
lo
ck
s
ize.
I
n
th
e
p
r
o
ce
d
u
r
e
o
f
th
e
class
if
icatio
n
,
th
is
s
tu
d
y
attem
p
te
d
at
co
m
p
ar
in
g
th
e
KNN
an
d
SVM
class
if
ier
s
f
o
r
ac
h
iev
in
g
a
h
ig
h
er
lev
el
o
f
ac
cu
r
ac
y
.
T
h
e
r
esu
lt
s
h
o
w
s
th
at
SVM
o
u
tp
er
f
o
r
m
s
KNN
in
b
r
ea
s
t c
an
ce
r
ab
n
o
r
m
alities
class
if
icatio
n
w
ith
9
3
.
8
8
% a
cc
u
r
ac
y
.
I
n
2
0
1
8
[
2
4
]
Sh
eb
a,
et
a
l
,
in
th
e
p
r
o
p
o
s
ed
m
eth
o
d
o
lo
g
y
,
f
o
r
th
e
p
r
e
-
p
r
o
ce
s
s
in
g
m
ed
ian
f
ilter
w
a
s
u
tili
ze
d
f
o
r
th
e
n
o
is
e
f
ilter
in
g
,
g
lo
b
al
th
r
esh
o
ld
in
g
f
o
r
r
em
o
v
in
g
th
e
s
m
all
ar
tif
ac
ts
.
B
B
is
u
tili
ze
d
f
o
r
r
em
o
v
i
n
g
p
ec
to
r
al
m
u
s
cles,
an
d
ad
ap
tiv
e
f
u
z
zy
lo
g
ic
b
ased
b
i
-
h
is
to
g
r
am
eq
u
aliza
tio
n
to
en
h
an
ce
th
e
q
u
ality
o
f
th
e
m
am
m
o
g
r
a
m
s
f
o
r
b
etter
p
er
ce
p
tio
n
.
T
h
e
R
OI
is
au
to
m
atica
lly
s
elec
t
ed
an
d
s
eg
m
en
tatio
n
f
r
o
m
m
am
m
o
g
r
a
m
s
im
ag
e
w
ith
th
e
u
s
e
o
f
m
o
r
p
h
o
lo
g
ical
o
p
er
atio
n
s
an
d
g
lo
b
al
th
r
esh
o
ld
in
g
.
Sh
ap
e,
GL
C
M
,
an
d
tex
tu
r
e
f
ea
tu
r
e
s
h
av
e
b
ee
n
o
b
tain
ed
f
r
o
m
R
OI
,
an
d
th
en
o
p
tim
u
m
f
ea
tu
r
es
h
av
e
b
ee
n
ch
o
s
en
w
ith
th
e
u
s
e
o
f
C
lass
if
ier
an
d
R
eg
r
ess
io
n
T
r
ee
(
C
A
R
T
)
.
Fin
ally
,
th
e
class
if
icatio
n
s
tep
h
as
b
ee
n
ca
r
r
ied
o
u
t
w
ith
th
e
Feed
-
f
o
r
w
ar
d
A
N
N
u
tili
zin
g
th
e
b
ac
k
p
r
o
p
ag
ati
o
n
.
T
h
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
ac
h
iev
ed
9
6
%
ac
cu
r
ac
y
.
I
n
2
0
1
9
[
25]
Mo
s
taf
a,
Sh
aim
a
a
,
et
a
l
.
T
h
e
r
esear
ch
er
s
u
tili
ze
d
f
ew
f
ea
tu
r
es
th
an
o
th
er
p
r
ev
io
u
s
r
esear
ch
th
at
u
s
ed
m
an
y
f
ea
tu
r
e
s
ets,
m
a
n
y
tech
n
iq
u
es
h
av
e
b
ee
n
u
s
ed
to
r
ed
u
ce
d
im
en
s
io
n
s
.
T
h
e
(
KNN)
an
d
(
A
NN)
class
if
ier
s
ar
e
u
s
ed
to
class
if
y
th
es
e
f
ew
f
ea
tu
r
es.
5
0
ca
s
es
o
f
th
e
'
B
A
HE
YA
Fo
u
n
d
atio
n
to
E
ar
ly
Dete
ctio
n
an
d
T
r
ea
tm
en
t
o
f
B
r
ea
s
t
C
an
ce
r
b
y
d
o
cto
r
s
an
d
r
ad
io
lo
g
is
ts
in
th
e
h
o
s
p
ital
h
a
ve
b
ee
n
u
tili
ze
d
f
o
r
th
e
p
r
o
p
o
s
ed
s
y
s
tem
.
T
h
e
im
ag
es
u
s
ed
ar
e
C
o
n
tr
ast
-
E
n
h
an
ce
d
Sp
ec
tr
al
Ma
m
m
o
g
r
am
s
(
C
E
SMs)
th
at
h
av
e
clea
r
er
an
d
m
o
r
e
co
n
tr
asti
n
g
im
ag
es
co
m
p
ar
ed
to
th
e
ty
p
ical
m
am
m
al
s
.
T
h
e
KNN
an
d
A
NN
class
if
ier
s
w
er
e
u
s
ed
an
d
th
e
o
u
tco
m
es
in
d
icate
to
ac
h
iev
e
ac
cu
r
ac
y
p
er
ce
n
t w
ith
9
2
p
er
ce
n
t w
ith
A
NN.
I
n
2
0
1
9
[
2
6
]
Salm
an
,
Na
s
s
ir
,
an
d
Sem
aa
I
b
r
ah
im
,
th
e
au
th
o
r
s
h
av
e
p
r
o
p
o
s
ed
a
s
y
s
tem
f
o
r
d
etec
t
p
o
ten
tial
ca
n
ce
r
tu
m
o
r
s
in
m
am
m
o
g
r
a
m
s
,
th
e
d
etec
tio
n
is
m
ad
e
th
r
o
u
g
h
au
to
m
atica
lly
d
iv
id
in
g
b
r
ea
s
t
im
ag
e
s
b
y
co
m
b
in
in
g
h
y
b
r
id
d
en
s
ity
s
licin
g
tech
n
iq
u
e
w
ith
th
e
ad
ap
tiv
e
k
-
m
ea
n
s
alg
o
r
it
h
m
,
also
b
y
d
iv
id
in
g
b
r
ea
s
t
im
ag
es
an
d
ex
tr
ac
tin
g
ar
ea
s
o
f
ca
n
ce
r
.
(
GL
C
M)
h
av
e
b
ee
n
u
s
ed
w
ith
p
r
o
p
o
s
ed
f
ea
tu
r
es
th
at
ar
e
g
r
ay
lev
el
d
en
s
ity
m
atr
ices
(
GL
DM
)
to
d
etec
t
ab
n
o
r
m
al
tis
s
u
e
u
s
in
g
ML
P
class
if
ier
s
.
E
x
p
er
im
en
tal
r
esu
lts
s
h
o
w
ed
a
s
ig
n
if
ican
t
im
p
r
o
v
em
e
n
t
in
b
r
ea
s
t
ca
n
ce
r
d
iag
n
o
s
is
ac
cu
r
ac
y
w
ith
m
o
r
e
th
an
,
9
1
.
1
7
%
.
I
n
2
0
1
9
[
2
7
]
R
.
Vij
ay
ar
aj
esw
ar
i
et
a
l
,
au
th
o
r
s
p
r
esen
t
a
C
A
D
s
y
s
tem
w
ith
th
e
f
ea
tu
r
es
o
b
tain
ed
w
ith
th
e
u
s
e
o
f
th
e
Ho
u
g
h
tr
an
s
f
o
r
m
atio
n
m
eth
o
d
,
it
is
a
2
-
D
tr
an
s
f
o
r
m
atio
n
.
W
h
ich
is
u
tili
ze
d
f
o
r
is
o
latin
g
th
e
f
ea
tu
r
e
o
f
a
s
p
ec
if
ic
s
h
ap
e
in
th
e
im
ag
e.
T
h
is
s
tu
d
y
d
is
cu
s
s
es
s
tr
ateg
ies
f
o
r
th
e
p
r
o
ce
s
s
o
f
class
if
icatio
n
an
d
f
ea
tu
r
e
ex
tr
ac
tio
n
.
Her
e,
it
is
u
tili
ze
d
f
o
r
th
e
d
etec
tio
n
o
f
th
e
m
am
m
o
g
r
a
m
im
ag
e
f
ea
tu
r
es
an
d
h
as
b
ee
n
class
if
ied
w
ith
th
e
u
s
e
o
f
th
e
SVMs.
T
h
e
r
esu
lts
h
av
e
s
h
o
w
n
th
at
th
e
s
u
g
g
ested
ap
p
r
o
ac
h
h
as
b
ee
n
s
u
cc
ess
f
u
l
in
class
if
y
in
g
th
e
ab
n
o
r
m
a
l
m
am
m
o
g
r
a
m
class
es
w
ith
an
ac
cu
r
ac
y
o
f
9
4
%.
2.
RE
S
E
ARCH
M
E
T
H
O
D
T
h
e
m
ain
g
o
al
o
f
th
e
p
r
esen
t
s
ea
r
ch
is
to
b
u
ild
a
class
if
ier
m
o
d
el
to
h
elp
s
p
h
y
s
icia
n
s
an
d
d
iag
n
o
s
ti
c
ex
p
er
ts
b
y
p
r
o
v
id
in
g
a
s
ec
o
n
d
d
iag
n
o
s
tic
v
iew
f
o
r
a
m
o
r
e
r
eliab
le
d
iag
n
o
s
tic
d
ec
is
io
n
.
Fig
u
r
e4
illu
s
tr
ates
th
e
d
iag
r
am
o
f
th
e
s
tag
es
o
f
th
e
p
r
o
p
o
s
ed
class
if
ier
m
o
d
el.
Evaluation Warning : The document was created with Spire.PDF for Python.
C
o
m
p
u
t.
Sci.
I
n
f
.
T
ec
h
n
o
l.
C
la
s
s
i
fica
tio
n
o
f m
a
mmo
g
r
a
m
s
b
a
s
ed
o
n
fea
t
u
r
es e
xtra
ctio
n
tech
n
iq
u
es
… (
E
n
a
s
Mo
h
a
mme
d
Hu
s
s
ein
S
a
ee
d
)
125
Fig
u
r
e
4
.
Sh
o
w
d
ia
g
r
a
m
o
f
th
e
s
u
g
g
e
s
ted
s
y
s
te
m
2
.
1
.
P
re
pro
ce
s
s
ing
Ma
m
m
o
g
r
ap
h
y
s
o
m
etim
es
co
n
tain
s
m
an
y
er
r
o
r
s
lik
e
n
o
is
e,
s
m
all
ar
tif
ac
ts
,
an
d
p
ec
to
r
al
m
u
s
cl
e
s
.
T
h
ese
ef
f
ec
ts
s
h
o
u
ld
b
e
r
em
o
v
ed
b
ec
au
s
e
th
ey
g
r
ea
tly
af
f
ec
t th
e
r
esu
lts
o
f
th
e
f
o
llo
w
in
g
s
tag
es,
s
u
ch
as
f
ea
tu
r
e
ex
tr
ac
tio
n
.
A
m
ed
ian
f
ilter
is
u
s
ed
f
o
r
n
o
is
e
an
d
lo
ca
l
co
n
tr
ast
r
em
o
v
al;
it
r
ep
r
esen
ts
a
f
ilter
o
f
th
e
s
p
atial
d
o
m
ain
,
w
h
ich
w
o
r
k
s
th
r
o
u
g
h
r
ep
lacin
g
th
e
ce
n
tr
al
p
ix
els
in
a
ce
r
tain
b
lo
ck
w
ith
b
lo
ck
p
ix
el
m
ed
ian
v
alu
es
.
T
h
e
s
m
all
ar
t
i
f
ac
ts
h
av
e
b
ee
n
elim
in
ated
th
r
o
u
g
h
th
e
co
n
v
er
s
io
n
o
f
th
e
im
ag
e
in
to
a
b
in
ar
y
f
o
r
m
at
w
ith
th
e
u
s
e
o
f
a
s
u
itab
le
th
r
esh
o
ld
an
d
af
ter
th
at,
th
e
ar
r
an
g
em
en
t
o
f
th
o
s
e
co
m
p
o
n
en
ts
th
r
o
u
g
h
th
e
ar
ea
f
o
r
th
e
is
o
latio
n
o
f
s
m
all
s
p
ac
es,
w
h
ich
in
clu
d
e
th
e
n
u
m
b
er
s
an
d
lab
els.
T
h
e
r
esu
lts
w
h
ich
h
av
e
b
ee
n
o
b
tain
ed
f
r
o
m
th
e
ap
p
licatio
n
o
f
th
is
s
ta
g
e
h
av
e
b
ee
n
illu
s
tr
ativ
e
in
F
ig
u
r
e
5
.
F
ig
u
r
e
5.
Sh
o
w
r
e
s
u
l
ts
t
h
e
p
r
e
-
p
r
o
ce
s
s
i
n
g
(
a)
in
p
u
t i
m
a
g
e
(
b
)
af
ter
ap
p
ly
m
ed
ian
f
ilter
(
c)
cu
t b
r
ea
s
t o
n
l
y
(
d
)
i
m
a
g
e
w
it
h
o
u
t
n
o
is
e
an
d
lab
el
2.2.
Seg
m
e
nta
t
io
n
Ma
n
y
s
eg
m
en
tatio
n
alg
o
r
ith
m
s
h
av
e
b
ee
n
u
s
ed
o
n
m
ed
ical
im
ag
es.
I
n
th
is
p
ap
er
,
th
is
s
tag
e
w
a
s
ap
p
lied
to
r
em
o
v
e
p
ec
to
r
al
m
u
s
cles
an
d
cu
t
th
e
lar
g
est
p
o
s
s
ib
le
s
q
u
ar
e
f
r
o
m
a
m
am
m
o
g
r
a
m
,
w
h
ich
r
ep
r
esen
ts
th
e
R
OI
.
Firstl
y
,
th
e
HB
B
R
G
alg
o
r
ith
m
w
as
u
s
ed
f
o
r
th
e
p
ec
to
r
al
m
u
s
cle
r
em
o
v
al
by
co
m
b
in
in
g
B
B
an
d
R
G.
W
h
er
e
B
B
alg
o
r
ith
m
s
w
er
e
ap
p
lied
ac
co
r
d
in
g
to
th
e
f
ac
t
th
at
p
ec
to
r
al
m
u
s
cles
ar
e
alm
o
s
t
tr
ian
g
u
lar
an
d
ar
e
ap
p
ea
r
in
g
in
th
e
b
r
ea
s
t
co
n
to
u
r
’
s
u
p
p
er
lef
t o
r
co
r
n
er
ac
co
r
d
in
g
to
w
h
eth
er
it is
th
e
r
ig
h
t
o
r
th
e
lef
t b
r
ea
s
t,
th
e
n
th
e
r
eg
io
n
'
s
g
r
o
w
in
g
alg
o
r
ith
m
ap
p
lied
b
y
s
elec
tin
g
a
s
ee
d
p
o
in
t
th
at
w
ill
b
e
au
to
m
atica
lly
ch
ar
ac
ter
ized
to
b
e
in
th
e
p
ec
to
r
al
m
u
s
cle
lim
its
.
I
n
ad
d
itio
n
to
th
at,
th
is
f
u
n
ctio
n
r
eq
u
ir
es
lo
ca
tin
g
th
e
d
is
tan
ce
o
f
th
e
m
ax
i
m
a
l
in
ten
s
ity
b
etw
ee
n
th
e
s
ee
d
p
o
in
t a
n
d
n
eig
h
b
o
r
p
ix
els,
f
in
ally
,
th
e
2
m
eth
o
d
s
h
av
e
b
ee
n
co
m
b
in
ed
to
o
b
tain
th
e
HB
B
R
G
m
ask
as sh
o
w
n
in
F
ig
u
r
e
6
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2722
-
3221
C
o
m
p
u
t.
Sci.
I
n
f
.
T
ec
h
n
o
l.
,
Vo
l.
2
,
No
.
3
,
No
v
em
b
er
20
2
1
:
1
21
–
1
31
126
Fig
u
r
e
6
.
R
esu
lts
o
f
t
h
e
HB
B
R
G
alg
o
r
it
h
m
(
a
)
.
p
r
ep
r
o
ce
s
s
in
g
i
m
a
g
e
(
b
)
.
cu
t b
r
ea
s
t o
n
l
y
(
c)
.
o
n
e'
s
v
al
u
e
f
o
r
b
r
ea
s
t
(
d
)
.
m
as
k
B
B
€
.
m
as
k
R
G
(
f
)
.
m
er
g
i
n
g
b
et
w
ee
n
B
B
,
an
d
R
G
m
as
k
(
HB
B
R
G
m
a
s
k
)
(
g
)
.
i
m
p
r
o
v
e
m
e
n
t
m
as
k
HB
B
R
G
al
g
o
r
ith
m
(
h
)
.
i
n
teg
r
ate
s
w
it
h
th
e
ze
r
o
m
atr
i
x
's
(
i
)
.
in
v
er
s
e
m
as
k
(
j
)
.
o
u
tp
u
t im
ag
e
I
n
th
e
s
ec
o
n
d
p
h
ase,
th
e
b
r
ea
s
t
im
ag
e
h
as
b
ee
n
s
eg
m
en
t
ed
to
o
b
tain
th
e
lar
g
est
p
o
s
s
ib
le
s
q
u
ar
e
ar
ea
th
at
ca
n
b
e
cu
t
f
r
o
m
th
e
im
ag
e,
as
th
is
ar
ea
is
s
q
u
ar
e,
T
h
is
p
r
o
ce
s
s
w
as
ap
p
lied
u
s
in
g
a
g
eo
m
etr
ical
m
eth
o
d
b
y
co
n
v
er
tin
g
th
e
im
ag
e
in
to
th
e
b
in
ar
y
an
d
f
in
d
in
g
th
e
m
ask
th
at
co
n
tain
s
o
n
ly
th
e
b
r
ea
s
t
w
ith
o
n
e'
s
v
alu
e
an
d
th
en
p
er
f
o
r
m
a
r
ev
er
s
e
s
ea
r
ch
p
r
o
ce
s
s
s
tar
ts
f
r
o
m
th
e
p
en
u
ltim
ate
p
ix
e
l in
th
e
lo
w
er
r
ig
h
t
co
r
n
er
an
d
co
m
p
ar
e
s
it
w
ith
its
th
r
ee
n
eig
h
b
o
r
s
r
ig
h
t,
b
o
tto
m
an
d
d
iag
o
n
al
to
f
in
d
th
e
least
v
alu
e
b
etw
ee
n
th
em
an
d
th
en
in
cr
ea
s
e
its
v
alu
e
in
o
n
e
m
ea
s
u
r
e,
th
is
p
r
o
ce
s
s
co
n
tin
u
es
o
n
all
th
e
m
ask
,
th
en
a
s
q
u
ar
e
is
d
r
aw
n
w
ith
th
e
co
o
r
d
in
ates
s
tar
tin
g
f
r
o
m
th
e
s
m
allest
p
ix
el
to
th
e
lar
g
est
v
alu
e.
Fin
ally
,
to
r
em
o
v
e
th
e
b
lack
b
ac
k
g
r
o
u
n
d
all
co
lu
m
n
s
an
d
r
o
w
s
w
ith
a
to
tal
s
u
m
eq
u
al
to
ze
r
o
ar
e
ex
clu
d
ed
.
A
lg
o
r
ith
m
1
d
escr
ib
es
th
e
p
r
o
ce
s
s
an
d
Fig
u
r
e
7
illu
s
tr
ate
s
r
esu
lts
.
Fig
u
r
e
7
.
Sh
o
w
r
e
s
u
l
ts
s
e
g
m
e
n
t
m
a
m
m
o
g
r
a
m
i
m
a
g
e
(
a)
m
a
m
m
o
g
r
a
m
i
m
a
g
e
(
b
)
b
in
ar
y
m
ask
(
c)
lar
g
est
s
q
u
ar
e
m
a
s
k
(
d
)
o
u
tp
u
t la
r
g
e
s
t
s
q
u
ar
e
(
e)
r
em
o
v
e
b
lac
k
b
ac
k
g
r
o
u
n
d
(
R
OI
)
A
l
g
o
r
i
t
h
m
(
1
)
:
C
u
t
t
h
e
l
a
r
g
e
st
p
o
ss
i
b
l
e
sq
u
a
r
e
t
o
f
i
n
d
t
h
e
R
OI
I
n
p
u
t
:
M
a
mm
o
g
r
a
m
i
m
a
g
e
w
i
t
h
o
u
t
p
e
c
t
o
r
a
l
mu
s
c
l
e
s
O
u
t
p
u
t
:
I
mag
e
w
i
t
h
t
h
e
l
a
r
g
e
st
p
o
ss
i
b
l
e
sq
u
a
r
e
B
e
g
i
n
S
t
e
p
1
:
I
M
=
r
e
a
d
i
mag
e
.
S
t
e
p
2
:
C
o
n
v
e
r
t
i
n
p
u
t
i
m
a
g
e
t
o
b
i
n
a
r
y
S
t
e
p
3
:
a
ss
i
g
n
z
e
r
o
a
r
r
a
y
s (M
a
sk
1
)
w
i
t
h
t
h
e
si
z
e
o
f
t
h
e
m
a
mm
o
g
r
a
m
i
mag
e
S
t
e
p
4
:
f
i
n
d
a
mas
k
w
i
t
h
o
n
e
's v
a
l
u
e
f
o
r
b
r
e
a
st
o
n
l
y
(
mask
2
)
f
r
o
m
se
g
me
n
t
a
t
i
o
n
o
f
t
h
e
mammo
g
r
a
m
i
mag
e
S
t
e
p
5
:
f
i
n
d
p
i
x
e
l
s p
(
r
,
c
)
b
e
f
o
r
e
t
h
e
l
a
st
f
o
r
(
M
a
s
k
2
)
,
w
h
i
c
h
i
s
i
n
t
h
e
l
o
w
e
r
r
i
g
h
t
c
o
r
n
e
r
,
t
h
e
n
a
p
p
l
y
T
h
e
f
o
l
l
o
w
i
n
g
o
p
e
r
a
t
i
o
n
s:
a
=
p
(
r
,
c
+
1
)
;
b
=
p
(
r
+
1
,
c
)
;
d
=
p
(
r
+
1
,
c
+
1
)
T
e
mp
=
mi
n
(
[
a
b
d
]
)
;
p
(
r
,
c
)
=
T
e
mp
+
1
.
S
t
e
p
6
:
F
i
n
d
t
h
e
p
i
x
e
l
t
h
a
t
c
o
n
t
a
i
n
s
t
h
e
max
i
m
u
m v
a
l
u
e
i
n
t
h
e
f
o
r
me
d
a
r
r
a
y
:
max
(
p
(
r
,
c
)
.
T
h
e
n
a
p
p
l
y
A
f
o
l
l
o
w
i
n
g
e
q
u
a
t
i
o
n
t
o
f
i
n
d
a
sq
u
a
r
e
w
i
t
h
w
h
i
t
e
v
a
l
u
e
s,
M
a
s
k
3
(
c
:
c
+
p
(
c
,
r
)
,
r
:
r
+
p
(
c
,
r
)
)
=
1
S
t
e
p
7
:
f
o
r
i
=
r
t
o
r
+
p
(
c
,
r
)
S
t
e
p
8
:
f
o
r
j
=
c
t
o
c
+
p
(
c
,
r
)
S
t
e
p
9
:
I
M
=
M
u
l
t
i
p
l
y
M
a
sk
3
(
i
,
j
)
×
I
M
(
i
,
j
)
S
t
e
p
1
0
:
e
n
d
f
o
r
S
t
e
p
1
1
:
e
n
d
f
o
r
S
t
e
p
1
2
:
c
o
u
n
t
i
n
g
t
h
e
su
m
o
f
p
i
x
e
l
s v
a
l
u
e
s fo
r
a
l
l
r
o
w
s a
n
d
c
o
l
u
m
n
a
n
d
t
h
e
n
f
i
n
d
t
h
e
mi
n
a
n
d
max
F
o
r
r
o
w
s a
n
d
c
o
l
u
mn
s c
o
n
t
a
i
n
i
n
g
a
su
m g
r
e
a
t
e
r
t
h
a
n
z
e
r
o
S
t
e
p
1
3
:
R
O
I
=
I
M
(
mi
n
r
o
w
:
max
r
o
w
,
mi
n
c
o
l
u
mn
:
m
a
x
c
o
l
u
m
n
)
En
d
W
h
er
e
r
is
th
e
r
o
w
,
c
is
t
h
e
co
l
u
m
n
s
,
an
d
T
em
p
is
t
h
e
te
m
p
o
r
ar
y
ta
n
k
.
Evaluation Warning : The document was created with Spire.PDF for Python.
C
o
m
p
u
t.
Sci.
I
n
f
.
T
ec
h
n
o
l.
C
la
s
s
i
fica
tio
n
o
f m
a
mmo
g
r
a
m
s
b
a
s
ed
o
n
fea
t
u
r
es e
xtra
ctio
n
tech
n
iq
u
es
… (
E
n
a
s
Mo
h
a
mme
d
Hu
s
s
ein
S
a
ee
d
)
127
2
.
3
.
F
ea
t
ures e
x
t
ra
ct
io
n
I
n
th
is
p
h
ase,
th
e
R
OI
is
ass
ig
n
ed
a
s
et
o
f
f
ea
tu
r
es
th
at
r
ep
r
esen
t
th
e
p
r
o
p
er
ties
o
f
th
e
tis
s
u
e.
T
h
ese
f
ea
tu
r
es
ca
n
b
e
a
s
et
o
f
r
ea
l
n
u
m
b
er
s
th
r
o
u
g
h
w
h
ich
th
e
n
o
r
m
al
tis
s
u
e
ca
n
b
e
d
is
tin
g
u
is
h
ed
f
r
o
m
th
e
ab
n
o
r
m
a
l
an
d
m
alig
n
an
t
f
r
o
m
b
en
ig
n
tis
s
u
e.
I
n
th
is
p
ap
er
,
af
ter
f
in
d
in
g
th
e
r
eg
io
n
o
f
in
ter
est
(
R
OI
)
f
o
r
ea
ch
im
ag
e
th
e
v
ec
to
r
f
ea
tu
r
es
co
n
s
is
t
o
f
7
3
f
ea
tu
r
es
th
at
w
er
e
ca
lcu
lated
b
ased
o
n
th
r
ee
tech
n
iq
u
es.
Firstl
y
ten
f
ea
tu
r
es
o
f
th
e
f
ir
s
t
-
o
r
d
er
f
ea
tu
r
es
ar
e
(
m
ea
n
an
d
SD
o
f
th
e
m
ea
n
,
m
ea
n
an
d
SD
o
f
SD,
m
ea
n
an
d
SD
o
f
v
ar
ian
ce
,
m
ea
n
an
d
SD
o
f
s
k
ew
n
ess
,
m
ea
n
an
d
s
tan
d
ar
d
d
ev
iatio
n
o
f
k
u
r
to
s
is
)
,
Seco
n
d
ly
th
e
f
if
ty
-
n
in
e
f
e
atu
r
e
o
f
th
e
L
B
P
m
eth
o
d
w
h
ich
is
lis
ted
f
r
o
m
elev
en
to
s
ix
ty
-
n
in
e
f
ea
tu
r
es.
Fin
ally
,
f
o
u
r
f
ea
tu
r
es
o
f
th
e
GL
C
M
ar
e
co
n
tr
ast,
co
r
r
elatio
n
,
en
e
r
g
y
,
an
d
h
o
m
o
g
en
eity
.
A
ll
o
f
th
o
s
e
f
ea
tu
r
es
w
er
e
ca
lcu
lated
w
ith
th
e
aim
o
f
cr
ea
tin
g
p
o
w
er
f
u
l
tex
tu
r
e
f
ea
tu
r
es.
T
h
e
s
tep
s
f
o
r
ex
tr
ac
tin
g
th
ese
f
ea
tu
r
es
ar
e
d
escr
ib
ed
in
A
lg
o
r
ith
m
2
.
A
l
g
o
r
i
t
h
m
(
2
)
:
Fe
a
t
u
r
e
E
x
t
r
a
c
t
i
o
n
f
o
r
t
h
e
R
OI
I
n
p
u
t
:
R
O
I
f
o
r
i
mag
e
s
O
u
t
p
u
t
:
A
r
r
a
y
o
f
f
e
a
t
u
r
e
s
B
e
g
i
n
S
t
e
p
1
:
R
e
a
d
R
O
I
i
mag
e
I
a
n
d
G
e
t
r
o
w
a
n
d
c
o
l
u
mn
o
f
t
h
e
i
mag
e
I
[
r
c
]
.
S
t
e
p
2
:
F
o
r
j
=
1
t
o
c
S
t
e
p
3
:
F
i
n
d
t
h
e
me
a
n
,
S
t
a
n
d
a
r
d
d
e
v
i
a
t
i
o
n
,
v
a
r
i
a
n
c
e
,
s
k
e
w
n
e
ss,
a
n
d
k
u
r
t
o
si
s.
S
t
e
p
4
:
e
n
d
f
o
r
S
t
e
p
5
:
f
e
a
t
u
r
e
s =
[
M
(
me
a
n
)
,
S
D
(
me
a
n
)
,
M
(
st
a
n
d
a
r
d
d
e
v
i
a
t
i
o
n
)
,
S
D
(
s
t
a
n
d
a
r
d
D
e
v
i
a
t
i
o
n
,
M
(
v
a
r
i
a
n
c
e
)
,
S
D
(
v
a
r
i
a
n
c
e
)
,
M
(
s
k
e
w
n
e
ss)
,
S
D
(
sk
e
w
n
e
ss)
,
M
(
k
u
r
t
o
si
s)
,
a
n
d
S
D
(
k
u
r
t
o
s
i
s)
]
.
S
t
e
p
6
:
F
i
n
d
f
e
a
t
u
r
e
s L
B
P
u
si
n
g
i
t
e
x
t
r
a
c
t
L
B
P
F
e
a
t
u
r
e
s (
I
)
S
t
e
p
7
:
C
r
e
a
t
e
a
G
L
C
M
f
r
o
m
I
mag
e
I
w
i
t
h
(
0
°
)
a
n
d
d
i
s
p
l
a
c
e
me
n
t
d
=
1
S
t
e
p
8
:
F
i
n
d
f
e
a
t
u
r
e
s GL
C
M
u
si
n
g
i
t
s
f
e
a
t
u
r
e
s =
g
r
a
y
c
o
p
r
o
p
s (G
L
C
M
)
.
S
t
e
p
9
:
F
i
n
d
t
h
e
C
o
n
t
r
a
s
t
o
f
R
O
I
s.
S
t
e
p
1
0
:
F
i
n
d
t
h
e
C
o
r
r
e
l
a
t
i
o
n
o
f
R
O
I
s
S
t
e
p
1
1
:
F
i
n
d
t
h
e
En
e
r
g
y
o
f
R
O
I
s.
S
t
e
p
1
2
:
F
i
n
d
t
h
e
H
o
mo
g
e
n
e
i
t
y
o
f
R
O
I
s.
S
t
e
p
1
3
:
S
a
v
e
f
e
a
t
u
r
e
s
En
d
2.4.
Cla
s
s
if
ica
t
io
n
On
ce
ex
tr
ac
ted
th
e
f
ea
tu
r
es,
ch
o
o
s
e
th
e
ap
p
r
o
p
r
iate
o
n
es to
en
ter
it
in
to
th
e
SVM
class
if
ier
.
SVM
h
as
b
ee
n
ap
p
lied
in
tw
o
lev
els
o
f
b
in
ar
y
class
if
icatio
n
,
th
e
f
ir
s
t
lev
el
r
ep
r
esen
tin
g
th
e
class
if
icatio
n
o
f
im
a
g
e
f
ea
tu
r
es
to
a
n
o
r
m
al
o
r
ab
n
o
r
m
al
im
ag
e,
th
en
if
th
e
r
esu
lts
o
f
th
e
f
ir
s
t
lev
el
ar
e
ab
n
o
r
m
al,
th
e
s
ec
o
n
d
lev
el
o
f
b
in
ar
y
class
if
icatio
n
is
ap
p
lied
,
w
h
ich
class
if
ies
f
ea
tu
r
es
o
f
b
en
ig
n
o
r
m
alig
n
an
t
im
ag
es.
SVM
ca
n
b
e
d
ef
in
ed
as
a
s
u
p
er
v
is
ed
ML
class
if
ier
,
w
h
er
e
a
r
ed
u
ce
d
f
ea
tu
r
e
v
ec
to
r
f
r
o
m
th
e
s
tep
o
f
th
e
f
ea
tu
r
es
s
elec
tio
n
h
as
b
ee
n
p
r
o
v
id
ed
as
in
p
u
t
d
ata
to
SVMs
class
if
ier
.
I
t
p
r
o
d
u
ce
s
s
u
p
p
o
r
t
v
ec
to
r
s
f
o
r
th
e
id
en
tif
i
ca
tio
n
o
f
b
o
u
n
d
ar
ies
b
etw
ee
n
b
o
th
class
es.
T
h
is
s
u
p
p
o
r
t
-
v
ec
to
r
is
u
tili
ze
d
f
o
r
th
e
d
eter
m
in
atio
n
o
f
th
e
h
y
p
er
p
lan
e
p
o
s
itio
n
w
h
er
e
it
h
as
b
ee
n
test
ed
w
ith
a
v
ar
iety
o
f
k
er
n
el
f
u
n
ctio
n
s
.
T
h
er
e
is
an
in
f
in
ite
am
o
u
n
t
o
f
s
ep
ar
atin
g
lin
es
w
h
ich
m
a
y
b
e
d
r
aw
n
,
th
e
o
b
j
ec
tiv
e
is
f
in
d
in
g
th
e
“
o
p
tim
al”
o
n
e,
w
h
ich
m
ea
n
s
,
o
n
e
w
h
ich
h
as
a
m
in
im
al
class
if
icati
o
n
er
r
o
r
o
n
th
e
p
r
ev
io
u
s
ly
u
n
s
ee
n
tu
p
les.
T
h
e
SVM
h
as
ap
p
r
o
ac
h
ed
th
is
is
s
u
e
b
y
s
ea
r
ch
in
g
f
o
r
m
ax
im
al
m
ar
g
i
n
a
l
h
y
p
er
-
p
lan
e.
T
h
e
o
p
tim
al
s
p
litt
in
g
o
f
a
h
y
p
er
p
lan
e
is
s
h
o
w
n
in
Fig
u
r
e
8
.
Fig
u
r
e
8
.
Op
ti
m
al
h
y
p
er
p
lan
e
f
o
r
s
u
p
p
o
r
t v
ec
to
r
m
ac
h
i
n
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2722
-
3221
C
o
m
p
u
t.
Sci.
I
n
f
.
T
ec
h
n
o
l.
,
Vo
l.
2
,
No
.
3
,
No
v
em
b
er
20
2
1
:
1
21
–
1
31
128
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
NS
T
h
e
ef
f
icien
cy
o
f
th
e
ap
p
r
o
ac
h
es
o
f
m
ac
h
in
e
lear
n
in
g
is
ev
alu
ated
b
ased
o
n
s
o
m
e
o
f
th
e
in
d
ic
es
o
f
th
e
p
er
f
o
r
m
an
ce
m
ea
s
u
r
es.
T
h
e
r
esu
lt,
F
alse
-
P
o
s
itiv
e
(
FP
)
,
m
ay
p
u
t
th
e
p
atien
t
in
a
f
r
ag
ile
p
o
s
itio
n
h
o
w
ev
er
,
u
s
i
n
g
co
m
p
lem
en
tar
y
ex
am
s
,
th
e
r
esu
lt
m
ay
b
e
ex
clu
d
ed
.
W
h
ile
in
a
ca
s
e
w
h
er
e
th
e
r
esu
lts
th
e
F
alse
-
N
eg
ativ
e
(
FN)
it
is
a
m
o
r
e
w
o
r
r
y
in
g
ca
s
e
if
an
in
d
iv
id
u
al
h
as
th
e
lesi
o
n
b
u
t
th
e
alg
o
r
ith
m
d
o
es
n
o
t
d
etec
t
[
2
8
]
.
A
co
n
f
u
s
i
o
n
m
atr
ix
f
o
r
th
e
p
r
ed
icted
an
d
ac
tu
al
class
es
is
ca
r
r
ied
o
u
t
co
m
p
r
is
in
g
f
alse
p
o
s
itiv
e
(
FP
)
,
tr
u
e
p
o
s
itiv
e
(
T
P)
,
f
alse
-
n
eg
ativ
e
(
FN)
,
an
d
tr
u
e
n
eg
ativ
e
(
T
N)
.
I
n
th
is
class
if
icatio
n
,
p
o
s
itiv
e/n
eg
ativ
e
in
d
icate
s
th
e
d
ec
is
io
n
w
h
ich
h
as
b
ee
n
m
ad
e
b
y
th
e
alg
o
r
ith
m
,
an
d
tr
u
e/f
alse
in
d
icate
s
th
e
w
ay
b
y
w
h
ich
th
e
d
ec
is
io
n
ag
r
ee
s
w
ith
th
e
ac
tu
al
clin
ical
s
tate.
W
h
er
e
in
ca
s
e
th
e
tw
o
class
es th
er
e
ar
e
o
n
ly
f
o
u
r
p
o
s
s
ib
le
o
u
tp
u
ts
r
ep
r
esen
ted
elem
en
ts
o
f
th
e
co
n
f
u
s
io
n
m
atr
ix
o
f
(
2
*
2
)
f
o
r
a
b
in
a
r
y
class
if
ier
s
ee
F
ig
u
r
e
9
[
29
]
,
[
30
]
.
F
ig
u
r
e
9
.
An
ill
u
s
tr
ati
v
e
ex
a
m
p
le
o
f
th
e
2
*
2
co
n
f
u
s
io
n
m
atr
i
x
[
29
]
T
h
er
e
ar
e
s
i
x
s
tati
s
tical
m
etr
i
cs
u
tili
ze
d
f
o
r
th
e
e
v
alu
a
tio
n
o
f
th
e
ef
f
icie
n
c
y
o
f
th
e
p
r
o
p
o
s
ed
s
y
s
te
m
b
ased
o
n
th
e
co
n
f
u
s
io
n
m
atr
ix
ar
e
ac
cu
r
ac
y
(
AC
C
)
,
er
r
o
r
r
ate
(
E
R
R
)
,
s
en
s
iti
v
it
y
(
S
N)
,
f
alse
-
p
o
s
itiv
e
r
ate
(
FP
R
)
,
s
p
ec
if
icit
y
(
SP
)
,
an
d
p
r
ec
is
io
n
(
P).
I
n
th
e
p
r
esen
t
r
esear
ch
,
th
e
p
r
o
p
o
s
ed
s
y
s
te
m
h
a
s
b
ee
n
ap
p
lied
to
all
im
a
g
e
s
in
th
e
MI
A
S
d
atab
ase,
w
h
er
e
7
0
%
o
f
th
e
i
m
ag
e
w
as
u
s
ed
f
o
r
th
e
tr
ain
i
n
g
p
h
a
s
e
an
d
3
0
%
test
i
n
g
p
h
a
s
e
o
f
r
an
d
o
m
in
s
tan
ts
o
f
i
m
ag
e
f
ea
t
u
r
es
f
r
o
m
t
h
e
d
ataset
w
it
h
1
0
0
iter
atio
n
s
.
T
h
e
r
esu
lts
s
h
o
w
th
at
SVM
f
o
r
th
e
f
ir
s
t
lev
el
h
as
ac
h
ie
v
ed
th
e
a
v
er
ag
e,
b
est,
an
d
w
o
r
s
e
ac
c
u
r
a
c
y
t
h
e
y
ar
e
89
.
171
%,
9
5
.4
54
%,
an
d
79
.
2
9
3
%
,
r
esp
ec
tiv
el
y
,
s
ee
T
ab
le
1
.
A
l
s
o
,
th
e
r
es
u
lt
s
s
h
o
w
t
h
at
SVM
f
o
r
th
e
s
ec
o
n
d
le
v
el
h
as
ac
h
ie
v
e
d
th
e
a
v
er
ag
e,
b
est,
an
d
w
o
r
s
e
ac
c
u
r
ac
y
t
h
e
y
ar
e
90
.
4
9
3
%,
9
7
.
6
0
%,
an
d
80
.
3
4
2
%,
r
esp
ec
tiv
el
y
,
as s
h
o
w
n
i
n
F
ig
u
r
e
1
0
.
T
ab
le
1
.
C
o
n
f
u
s
io
n
m
atr
i
x
r
es
u
lt o
f
S
VM
f
o
r
f
ir
s
t
-
lev
e
l
#
A
C
C
ER
R
SE
F
P
R
SP
P
T
h
e
a
v
e
r
a
g
e
r
e
su
l
t
s
8
9
.
1
7
1
1
0
.
8
2
8
9
9
.
4
5
3
3
.
3
5
9
6
6
.
6
4
8
3
.
5
1
9
T
h
e
b
e
st
r
e
su
l
t
s
9
5
.
4
5
4
4
.
5
4
5
1
0
0
1
1
.
5
4
3
8
8
.
4
5
6
9
3
.
3
7
7
T
h
e
w
o
r
st
r
e
su
l
t
s
7
9
.
2
8
1
2
0
.
7
1
7
8
6
.
3
0
1
4
7
.
0
5
8
5
2
.
9
4
1
7
3
.
8
5
6
An
al
y
s
i
s
o
f
th
e
p
er
f
o
r
m
a
n
ce
o
f
r
es
u
lt
s
o
f
th
e
ev
al
u
atio
n
m
e
t
r
ics
f
o
r
a
f
ir
s
t
-
le
v
el
clas
s
if
icat
io
n
w
h
ich
class
i
f
ies
t
h
e
i
m
ag
e
to
n
o
r
m
al
o
r
ab
n
o
r
m
al
class
e
s
is
s
h
o
w
n
i
n
F
i
g
u
r
e
1
0
,
w
h
er
e
it
ca
n
b
e
o
b
s
er
v
ed
th
e
b
est
an
d
w
o
r
s
t
v
al
u
e
o
b
tain
ed
f
o
r
th
i
s
class
i
f
ier
.
F
ig
u
re
1
0
.
Disp
lay
s th
e
p
e
rf
o
rm
a
n
c
e
a
n
a
ly
sis o
f
th
e
S
VM
f
o
r
f
irst
lev
e
l
f
ro
m
tab
le1
Evaluation Warning : The document was created with Spire.PDF for Python.
C
o
m
p
u
t.
Sci.
I
n
f
.
T
ec
h
n
o
l.
C
la
s
s
i
fica
tio
n
o
f m
a
mmo
g
r
a
m
s
b
a
s
ed
o
n
fea
t
u
r
es e
xtra
ctio
n
tech
n
iq
u
es
… (
E
n
a
s
Mo
h
a
mme
d
Hu
s
s
ein
S
a
ee
d
)
129
T
ab
le
2
s
h
o
w
s
C
o
n
f
u
s
io
n
m
a
tr
ix
r
esu
l
t
o
f
SVM
f
o
r
s
ec
o
n
d
-
lev
el
.
An
al
y
s
i
s
o
f
th
e
p
er
f
o
r
m
an
ce
o
f
r
esu
lt
s
o
f
th
e
e
v
alu
a
tio
n
m
etr
i
cs
f
o
r
a
s
ec
o
n
d
-
le
v
el
clas
s
i
f
ica
tio
n
w
h
ic
h
clas
s
i
f
ies t
h
e
ab
n
o
r
m
al
i
m
ag
e
s
o
n
l
y
to
b
en
ig
n
an
d
m
al
ig
n
a
n
t
class
e
s
i
s
s
h
o
w
n
in
Fi
g
u
r
e1
1
,
w
h
er
e
it
ca
n
b
e
o
b
s
er
v
ed
th
e
b
est
an
d
w
o
r
s
t
v
alu
e
o
b
tain
ed
f
o
r
th
is
cla
s
s
i
f
ier
.
T
ab
le
2.
C
o
n
f
u
s
io
n
m
atr
i
x
r
es
u
lt o
f
S
VM
f
o
r
s
ec
o
n
d
-
le
v
el
#
A
C
C
ER
R
SE
F
P
R
SP
P
T
h
e
a
v
e
r
a
g
e
r
e
su
l
t
s
8
9
.
1
7
1
1
0
.
8
2
8
9
9
.
4
5
3
3
.
3
5
9
6
6
.
6
4
8
3
.
5
1
9
T
h
e
b
e
st
r
e
su
l
t
s
9
5
.
4
5
4
4
.
5
4
5
1
0
0
1
1
.
5
4
3
8
8
.
4
5
6
9
3
.
3
7
7
T
h
e
w
o
r
st
r
e
su
l
t
s
7
9
.
2
8
1
2
0
.
7
1
7
8
6
.
3
0
1
4
7
.
0
5
8
5
2
.
9
4
1
7
3
.
8
5
6
F
ig
u
re
1
1
. D
is
p
la
y
s
t
h
e
r
esu
l
ts
an
d
p
er
f
o
r
m
an
ce
a
n
al
y
s
i
s
o
f
t
h
e
SVM
f
o
r
s
ec
o
n
d
lev
el
I
n
th
is
p
ar
t,
in
o
r
d
e
r
t
o
b
etter
e
v
alu
a
te
o
u
r
p
r
o
p
o
s
ed
s
y
s
te
m
,
o
u
r
p
r
o
p
o
s
ed
s
y
s
te
m
w
a
s
co
m
p
ar
ed
w
it
h
a
s
et
o
f
p
r
ev
io
u
s
w
o
r
k
s
b
y
co
m
p
ar
i
n
g
all
t
h
e
m
ai
n
s
ta
g
es
o
f
th
e
s
y
s
te
m
,
al
g
o
r
it
h
m
s
,
tech
n
iq
u
es,
an
d
t
h
e
le
v
el
f
o
r
w
h
ich
t
h
e
s
y
s
te
m
w
a
s
d
esig
n
ed
w
it
h
o
u
t
ac
r
es
an
d
th
e
r
esu
lt
s
o
b
tain
ed
.
T
h
ese
co
m
p
ar
is
o
n
s
ar
e
s
h
o
w
n
i
n
T
ab
le
3
.
T
ab
le
3.
Sh
o
w
s
t
h
e
co
m
p
ar
is
o
n
o
f
t
h
e
p
r
o
p
o
s
ed
s
y
s
te
m
w
i
th
th
e
r
elate
d
w
o
r
k
s
R
e
l
a
t
e
d
w
o
r
k
Pre
-
P
r
o
c
e
ssi
n
g
a
n
d
E
n
h
a
n
c
e
me
n
t
S
e
g
me
n
t
a
t
i
o
n
F
e
a
t
u
r
e
e
x
t
r
a
c
t
i
o
n
C
l
a
ssi
f
i
e
r
A
c
c
u
r
a
c
y
(
%)
(
S
h
e
b
a
,
a
n
d
R
a
j
,
2
0
1
8
)
[
2
2
]
me
d
i
a
n
f
i
l
t
e
r
g
l
o
b
a
l
t
h
r
e
s
h
o
l
d
i
n
g
B
B
,
g
l
o
b
a
l
t
h
r
e
sh
o
l
d
i
n
g
,
a
n
d
mo
r
p
h
o
l
o
g
i
c
a
l
S
h
a
p
e
,
G
L
R
L
M
,
a
n
d
G
L
C
M
.
F
e
e
d
f
o
r
w
a
r
d
n
e
u
r
a
l
n
e
t
w
o
r
k
s
(FFNN)
9
6
f
o
r
t
h
e
f
i
r
s
t
a
n
d
se
c
o
n
d
l
e
v
e
l
s
(
S
a
l
man
,
a
n
d
S
e
maa
,
2
0
1
9
)
[
2
4
]
me
d
i
a
n
f
i
l
t
e
r
a
n
d
h
i
s
t
o
g
r
a
m
e
q
u
a
l
i
z
a
t
i
o
n
S
e
g
me
n
t
a
t
i
o
n
b
a
se
d
d
a
t
a
se
t
,
K
-
me
a
n
,
a
n
d
RG
G
L
C
M
,
a
n
d
G
L
D
M
F
e
a
t
u
r
e
M
L
P
9
1
.
1
7
f
o
r
t
h
e
se
c
o
n
d
l
e
v
e
l
V
i
j
a
y
a
r
a
j
e
sw
a
r
i
e
t
l
,
2
0
1
9
)
[
2
5
]
R
e
mo
v
e
a
l
l
u
n
w
a
n
t
e
d
p
a
r
t
max
i
mi
z
a
t
i
o
n
e
st
i
m
a
t
i
o
n
H
o
u
g
h
t
r
a
n
sf
o
r
m
S
V
M
9
4
%
f
o
r
se
c
o
n
d
l
e
v
e
l
P
r
o
p
o
se
d
me
t
h
o
d
me
d
i
a
n
f
i
l
t
e
r
g
l
o
b
a
l
t
h
r
e
sh
o
l
d
i
n
g
H
B
B
R
G
a
l
g
o
r
i
t
h
m
a
n
d
c
u
t
t
i
n
g
l
a
r
g
e
st
p
o
ss
i
b
l
e
sq
u
a
r
e
F
i
r
st
-
o
r
d
e
r
,
L
B
P
,
a
n
d
G
L
C
M
.
S
V
M
9
5
.
4
5
4
F
o
r
f
i
r
st
l
e
v
e
l
,
9
7
.
2
6
0
f
o
r
se
c
o
n
d
l
e
v
e
l
4.
CO
NCLU
SI
O
N
Ma
m
m
o
g
r
ap
h
y
is
th
e
m
o
s
t
ef
f
ec
tiv
e
m
eth
o
d
th
at
is
u
s
ed
in
ea
r
ly
d
etec
tio
n
s
o
f
b
r
ea
s
t
ca
n
ce
r
,
T
h
e
m
a
i
n
o
b
j
ec
tiv
e
is
d
ev
elo
p
in
g
th
e
C
A
D
s
y
s
tem
f
o
r
d
iag
n
o
s
is
m
am
m
o
g
r
a
m
im
ag
es
to
ass
is
t
d
o
cto
r
s
an
d
d
iag
n
o
s
ti
c
ex
p
er
ts
b
y
p
r
o
v
id
in
g
a
s
ec
o
n
d
v
iew
p
o
in
t,
it
g
iv
es
m
o
r
e
co
n
f
id
en
ce
to
th
e
d
iag
n
o
s
tic
p
r
o
ce
s
s
.
I
n
th
is
p
ap
er
,
T
h
e
r
esu
lts
p
r
o
v
ed
th
at
th
e
m
ed
ian
f
ilter
is
an
id
ea
l
f
ilter
to
r
em
o
v
e
n
o
is
e
an
d
lo
ca
l
co
n
tr
ast
f
o
u
n
d
in
th
e
m
am
m
o
g
r
a
m
.
T
h
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
f
o
r
r
em
o
v
in
g
s
m
all
ar
tif
ac
ts
h
as
ac
h
iev
ed
1
0
0
%
r
esu
lts
f
o
r
th
is
p
u
r
p
o
s
e.
R
em
o
v
al
o
f
th
e
p
ec
to
r
al
m
u
s
cles r
ep
r
esen
ts
th
e
b
ig
g
est
o
b
s
tacle
in
th
e
tr
ea
tm
en
t
o
f
m
am
m
o
g
r
a
m
s
b
ec
au
s
e
th
e
y
clo
s
ely
r
esem
b
le
tu
m
o
r
s
an
d
th
e
r
est
o
f
th
e
b
r
ea
s
t
tis
s
u
e,
p
ar
ticu
lar
ly
in
th
e
ty
p
es
o
f
f
atty
an
d
ad
en
o
ca
r
cin
o
m
a
s
,
as
w
ell
as
th
eir
p
r
esen
ce
,
h
as
a
g
r
ea
t
im
p
ac
t
o
n
th
e
r
esu
lts
o
f
th
e
f
o
llo
w
in
g
s
tag
es
.
A
g
eo
m
etr
ic
s
eg
m
en
ta
t
i
o
n
m
eth
o
d
w
as
p
r
o
p
o
s
ed
to
cu
t
th
e
lar
g
est
p
o
s
s
ib
le
s
q
u
ar
e
r
ep
r
esen
tin
g
a
s
am
p
le
o
f
b
r
ea
s
t
tis
s
u
e,
it
ac
h
iev
i
n
g
1
0
0
%
s
u
cc
ess
w
ith
n
o
r
m
al
im
ag
es
an
d
m
o
r
e
th
an
9
8
%
w
ith
ab
n
o
r
m
al
im
ag
es,
w
h
er
e
th
e
tu
m
o
r
is
w
ith
in
th
e
cr
o
p
ar
ea
.
T
h
e
p
r
o
p
o
s
ed
tex
tu
r
e
f
ea
tu
r
es
ar
e
b
ased
o
n
th
r
ee
tech
n
o
lo
g
ies
ar
e
f
ir
s
t
-
o
r
d
er
,
L
B
P
,
an
d
GL
C
M,
th
es
e
f
ea
tu
r
es
g
iv
e
th
e
alg
o
r
ith
m
m
o
r
e
r
o
b
u
s
tn
ess
d
u
e
to
its
r
esis
tan
ce
to
m
an
y
im
ag
e
s
itu
atio
n
v
ar
iatio
n
s
th
at
lead
to
th
e
b
est
d
is
cr
im
in
atio
n
p
o
ten
tial
f
o
r
class
if
icatio
n
th
e
ty
p
e
o
f
im
ag
e.
Fin
ally
,
s
ev
er
al
f
u
tu
r
e
r
esear
ch
w
o
r
k
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2722
-
3221
C
o
m
p
u
t.
Sci.
I
n
f
.
T
ec
h
n
o
l.
,
Vo
l.
2
,
No
.
3
,
No
v
em
b
er
20
2
1
:
1
21
–
1
31
130
ca
n
b
e
d
o
n
e
f
o
r
th
e
p
r
o
d
u
ctio
n
o
f
o
u
r
p
ap
er
s
u
ch
as
m
o
d
el
d
ev
elo
p
m
en
t
f
r
o
m
th
e
d
iag
n
o
s
tic
m
o
d
el
to
th
e
d
iag
n
o
s
tic
an
d
p
r
ed
ictio
n
m
o
d
els,
an
d
test
s
o
f
n
ew
s
eg
m
en
tatio
n
m
eth
o
d
s
th
at
w
ill
p
r
o
v
id
e
b
etter
r
esu
lts
to
id
en
tif
y
th
e
d
am
ag
e
an
d
in
s
u
latio
n
f
r
o
m
th
e
r
est o
f
o
u
r
b
r
ea
s
t
tis
s
u
e,
p
ar
ticu
lar
ly
in
f
atty
an
d
g
lan
d
u
lar
p
h
o
to
s
,
d
u
r
in
g
th
is
s
tag
e
,
also
a
p
p
ly
th
e
p
r
o
p
o
s
ed
b
r
ea
s
t
d
iag
n
o
s
is
m
o
d
el
to
o
th
er
b
r
ea
s
t
im
ag
in
g
an
d
ex
am
in
at
i
o
n
m
eth
o
d
s
,
lik
e
MRI
an
d
C
T
.
RE
F
E
R
E
NC
E
S
[1
]
Bra
y
,
F
re
d
d
ie,
e
t
a
l
.
"
G
lo
b
a
l
c
a
n
c
e
r
sta
ti
sti
c
s 2
0
1
8
:
G
L
OBO
C
A
N e
stim
a
tes
o
f
in
c
id
e
n
c
e
a
n
d
m
o
rtalit
y
w
o
rld
w
id
e
f
o
r
3
6
c
a
n
c
e
rs i
n
1
8
5
c
o
u
n
tr
ies
.
"
CA:
a
c
a
n
c
e
r jo
u
rn
a
l
fo
r c
li
n
icia
n
s
,
v
o
l.
68
,
n
o
.
6
,
p
p
.
3
9
4
-
4
2
4
,
2
0
1
8
.
[2
]
Qiu
,
Yu
c
h
e
n
,
e
t
a
l
.
"
A
n
in
it
ial
in
v
e
stig
a
ti
o
n
o
n
d
e
v
e
lo
p
i
n
g
a
n
e
w
m
e
th
o
d
t
o
p
re
d
ict
sh
o
rt
-
term
b
re
a
st
c
a
n
c
e
r
ris
k
b
a
se
d
o
n
d
e
e
p
lea
rn
in
g
tec
h
n
o
lo
g
y
.
"
Pro
c
e
e
d
in
g
s,
M
e
d
ica
l
Ima
g
i
n
g
2
0
1
6
:
C
o
mp
u
ter
-
Ai
d
e
d
Dia
g
n
o
sis
;
v
o
l.
9
7
8
,
9
7
8
5
2
1
,
2
0
1
6
.
[3
]
Ja
lalian
,
Af
sa
n
e
h
,
e
t
a
l
.
"
Co
m
p
u
ter
-
a
id
e
d
d
e
tec
ti
o
n
/d
iag
n
o
sis
o
f
b
r
e
a
st
c
a
n
c
e
r
in
m
a
m
m
o
g
ra
p
h
y
a
n
d
u
lt
ra
so
u
n
d
:
a
re
v
ie
w
.
"
Cli
n
ica
l
ima
g
in
g
,
v
o
l.
37
,
n
o
.
3
,
p
p
.
4
2
0
-
4
2
6
.
,
2
0
1
3
.
[4
]
“
M
a
it
ra
,
In
d
ra
Ka
n
ta,
S
a
n
jay
Na
g
,
a
n
d
S
a
m
ir
Ku
m
a
r
Ba
n
d
y
o
p
a
d
h
y
a
y
.
Tec
h
n
iq
u
e
f
o
r
p
re
p
r
o
c
e
ss
in
g
o
f
a
d
ig
it
a
l
m
a
m
m
o
g
r
a
m
.
"
Co
mp
u
ter
me
th
o
d
s a
n
d
p
r
o
g
ra
ms
i
n
b
io
me
d
icin
e
,
v
o
l.
1
0
7
,
n
o
.
2
,
p
p
.
1
7
5
-
1
8
8
,
2
0
1
2
.
[5
]
M
a
k
a
n
d
a
r,
A
z
iz,
a
n
d
Bh
a
g
irath
i
Ha
lalli
.
"
P
re
-
p
ro
c
e
ss
in
g
o
f
m
a
m
m
o
g
ra
p
h
y
i
m
a
g
e
f
o
r
e
a
rl
y
d
e
tec
ti
o
n
o
f
b
re
a
st
c
a
n
c
e
r.
"
In
ter
n
a
ti
o
n
a
l
J
o
u
r
n
a
l
o
f
Co
mp
u
ter
A
p
p
li
c
a
ti
o
n
s
,
v
o
l.
1
4
4
,
n
o
.
3
,
p
p
.
11
-
15
,
2
0
1
6
.
[6
]
Be
rb
a
r,
M
o
h
a
m
e
d
A
.
"
H
y
b
rid
m
e
th
o
d
s
f
o
r
f
e
a
tu
re
e
x
trac
ti
o
n
f
o
r
b
r
e
a
st
m
a
ss
e
s
c
las
si
f
ica
ti
o
n
.
"
Eg
y
p
t
ia
n
in
f
o
rm
a
ti
c
s
jo
u
rn
a
l
,
v
o
l.
19
,
n
o
.
1
,
p
p
.
63
-
73
,
2
0
1
8
.
[7
]
T
e
m
b
e
y
,
M
u
g
d
h
a
,
“
Co
m
p
u
ter
-
a
id
e
d
d
iag
n
o
sis
f
o
r
m
a
m
m
o
g
ra
p
h
ic
m
icro
c
a
lci
f
ica
ti
o
n
c
lu
ste
rs,”
P
h
.
D.
d
isse
rtatio
n
,
De
p
t.
Co
m
p
.
S
c
i
.
a
n
d
En
g
.,
Un
iv
.
o
f
So
u
th
Flo
r
id
a
,
S
c
h
o
lar Co
m
m
o
n
s
,
200
3
.
[8
]
Ch
e
n
g
,
He
n
g
-
Da
,
e
t
a
l.
"
A
u
to
m
a
t
e
d
b
re
a
st
c
a
n
c
e
r
d
e
tec
ti
o
n
a
n
d
c
las
sif
i
c
a
ti
o
n
u
sin
g
u
lt
ra
so
u
n
d
im
a
g
e
s:
A
su
rv
e
y
.
"
Pa
tt
e
rn
re
c
o
g
n
it
i
o
n
,
v
o
l.
43
,
n
o
.
1
,
p
p
.
2
9
9
-
3
1
7
,
2
0
1
0
.
[9
]
E.
M
.
H.
S
a
e
e
d
a
n
d
H.
A
.
S
a
leh
,
"
P
e
c
to
ra
l
M
u
sc
les
Re
m
o
v
a
l
in
M
a
m
m
o
g
ra
m
I
m
a
g
e
b
y
H
y
b
rid
Bo
u
n
d
in
g
Bo
x
a
n
d
Re
g
io
n
G
ro
w
in
g
A
lg
o
rit
h
m
,
"
2
0
2
0
In
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
n
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
S
o
ft
w
a
re
En
g
in
e
e
rin
g
(
CS
AS
E)
,
Du
h
o
k
,
Ira
q
,
2
0
2
0
,
p
p
.
1
4
6
-
1
5
1
,
d
o
i:
1
0
.
1
1
0
9
/C
S
A
S
E4
8
9
2
0
.
2
0
2
0
.
9
1
4
2
0
5
5
.
[1
0
]
Du
th
,
P
.
S
u
d
h
a
rsh
a
n
,
V
im
a
l
V
isw
a
n
a
th
,
a
n
d
P
a
n
k
a
j
S
re
e
k
u
m
a
r.
"
Br
a
in
im
a
g
e
se
g
m
e
n
tatio
n
u
si
n
g
lev
e
l
se
t:
A
n
h
y
b
rid
a
p
p
ro
a
c
h
.
"
I
n
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
A
d
v
a
n
c
e
s in
Ap
p
li
e
d
S
c
ien
c
e
s
,
v
o
l.
6
,
n
o
.
3
,
p
p
.
2
6
7
-
2
7
6
,
2
0
1
7
.
[1
1
]
L
I,
S
h
e
n
g
lan
,
e
t
a
l.
"
P
e
rf
o
rm
a
n
c
e
e
v
a
lu
a
ti
o
n
o
f
a
CA
D
s
y
ste
m
f
o
r
d
e
tec
ti
n
g
m
a
ss
e
s
o
n
m
a
m
m
o
g
ra
m
s
b
y
u
sin
g
th
e
M
IA
S
d
a
tab
a
se
.
"
M
e
d
ica
l
Ima
g
in
g
a
n
d
In
f
o
rm
a
ti
o
n
S
c
ien
c
e
s
,
v
o
l.
18
,
n
o
.
3
,
p
p
.
p
1
4
4
-
p
1
5
3
,
2
0
0
1
.
[1
2
]
J
S
u
c
k
li
n
g
,
S
A
stle
y
,
D
Be
tal,
N
Ce
rn
e
a
z
,
D
R
D
a
n
c
e
,
S
-
L
Ko
k
,
J
P
a
rk
e
r,
I
Rick
e
tt
s,
J
S
a
v
a
g
e
,
E
S
ta
m
a
ta
k
is
a
n
d
P
T
a
y
lo
r,
N
Ka
rrse
m
e
ij
e
r,
A
Cl
a
rk
,
T
h
e
m
in
i
-
M
IAS
d
a
tab
a
se
o
f
m
a
m
m
o
g
r
a
m
s,
[
On
li
n
e
]
h
tt
p
:
//
p
e
i
p
a
.
e
ss
e
x
.
a
c
.
u
k
/i
n
f
o
/
m
ias
.
h
tm
l.
[1
3
]
Hu
ss
a
in
,
A
l
y
a
a
,
A
l
a
a
No
o
ri
M
a
z
h
e
r,
a
n
d
A
sra
a
Ra
z
a
k
.
"
Clas
si
f
ic
a
ti
o
n
o
f
Bre
a
st
T
issu
e
f
o
r
m
a
m
m
o
g
ra
m
s
i
m
a
g
e
s
u
sin
g
in
ten
sity
h
isto
g
ra
m
a
n
d
sta
ti
stica
l
m
e
th
o
d
s."
Ira
q
i
J
o
u
rn
a
l
o
f
S
c
ien
c
e
,
v
o
l.
5
3
,
n
o
.
5
,
p
p
.
1
0
9
2
-
1
0
9
6
,
2
0
1
2
.
[1
4
]
Hla
in
g
,
K.
Ny
e
m
N
y
e
m
.
"
F
irst
o
rd
e
r
sta
ti
stics
a
n
d
G
L
CM
b
a
se
d
f
e
a
tu
re
e
x
trac
ti
o
n
f
o
r
re
c
o
g
n
it
i
o
n
o
f
M
y
a
n
m
a
r
p
a
p
e
r
c
u
rre
n
c
y
.
"
Pro
c
e
e
d
in
g
s o
f
th
e
3
1
s
t
IIE
R
I
n
ter
n
a
t
io
n
a
l
C
o
n
fer
e
n
c
e
,
Ba
n
g
k
o
k
,
T
h
a
il
a
n
d
.
2
0
1
5
.
[1
5
]
Ha
m
o
u
d
a
,
S
.
K.
M
.
,
R.
H.
B.
El
-
Ezz
,
a
n
d
M
o
h
a
m
m
e
d
E.
W
a
h
e
d
.
"
En
h
a
n
c
e
m
e
n
t
a
c
c
u
ra
c
y
o
f
b
re
a
st
tu
m
o
r
d
iag
n
o
si
s
in
d
ig
it
a
l
m
a
m
m
o
g
ra
m
s."
J
o
u
rn
a
l
o
f
Bi
o
me
d
ica
l
S
c
ien
c
e
s
,
v
o
l.
6
.
n
o
.
4
,
p
p
.
1
-
8
,
2
0
1
7
.
[1
6
]
M
a
n
iar,
S
w
e
e
t
y
,
a
n
d
Ja
g
d
ish
S
.
S
h
a
h
.
"
Clas
sif
i
c
a
ti
o
n
o
f
c
o
n
te
n
t
b
a
se
d
m
e
d
ica
l
i
m
a
g
e
re
tri
e
v
a
l
u
sin
g
tex
tu
re
a
n
d
sh
a
p
e
f
e
a
tu
re
w
it
h
n
e
u
ra
l
n
e
tw
o
r
k
.
"
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
A
d
v
a
n
c
e
s
i
n
Ap
p
l
ied
S
c
ien
c
e
s
,
v
o
l.
6
,
n
o
.
4
,
p
p
.
3
6
8
-
374
,
2
0
1
7
.
[1
7
]
Da
sh
,
P
ra
jn
a
P
a
rim
it
a
,
Dip
ti
P
a
tra
,
a
n
d
S
u
d
h
a
n
su
Ku
m
a
r
M
ish
ra
.
"
L
o
c
a
l
b
i
n
a
ry
p
a
tt
e
rn
a
s
a
tex
tu
re
f
e
a
tu
re
d
e
sc
rip
t
o
r
in
o
b
jec
t
trac
k
in
g
a
lg
o
rit
h
m
.
"
In
telli
g
e
n
t
Co
m
p
u
t
in
g
,
Ne
two
rk
in
g
,
a
n
d
I
n
f
o
rm
a
ti
c
s.
S
p
rin
g
e
r,
Ne
w
De
lh
i,
v
o
l.
2
4
3
,
p
p
.
5
4
1
-
548
,
2
0
1
4
.
[1
8
]
Na
n
n
i,
L
o
ris,
A
les
sa
n
d
ra
L
u
m
in
i
,
a
n
d
S
h
e
ry
l
Br
a
h
n
a
m
.
"
L
o
c
a
l
b
in
a
ry
p
a
tt
e
rn
s
v
a
rian
ts
a
s
te
x
tu
re
d
e
sc
rip
to
rs
f
o
r
m
e
d
ica
l
ima
g
e
a
n
a
l
y
sis."
Arti
fi
c
ia
l
in
telli
g
e
n
c
e
in
me
d
icin
e
,
v
o
l.
49
,
n
o
.
2
,
p
p
.
1
1
7
-
1
2
5
,
2
0
1
0
.
[1
9
]
No
sa
k
a
,
R
y
u
su
k
e
,
Ya
su
h
iro
Oh
k
a
wa
,
a
n
d
Ka
z
u
h
ir
o
F
u
k
u
i.
"
F
e
a
tu
re
e
x
trac
ti
o
n
b
a
se
d
o
n
c
o
-
o
c
c
u
rre
n
c
e
o
f
a
d
jac
e
n
t
lo
c
a
l
b
in
a
ry
p
a
tt
e
rn
s."
Pro
c
e
e
d
in
g
s,
Pa
rt
II:
Ad
v
a
n
c
e
s
in
Ima
g
e
a
n
d
Vi
d
e
o
T
e
c
h
n
o
l
o
g
y
-
5
th
P
a
c
if
ic
Ri
m
S
y
mp
o
si
um
,
P
S
IV
T
2
0
1
1
,
G
w
a
n
g
ju
,
S
o
u
t
h
Ko
re
a
,
No
v
e
m
b
e
r
20
-
23
,
2
0
1
1
.
[2
0
]
De
S
iq
u
e
ira,
F
e
rn
a
n
d
o
Ro
b
e
rti
,
W
il
li
a
m
Ro
b
so
n
S
c
h
w
a
rtz,
a
n
d
H
e
li
o
P
e
d
r
in
i.
"
M
u
lt
i
-
sc
a
le g
ra
y
le
v
e
l
c
o
-
o
c
c
u
rre
n
c
e
m
a
tri
c
e
s
f
o
r
tex
tu
re
d
e
sc
rip
ti
o
n
.
"
Ne
u
ro
c
o
mp
u
ti
n
g
,
v
o
l
.
1
2
0
,
n
o
.
1
,
p
p
.
3
3
6
-
3
4
5
,
2
0
1
3
.
[2
1
]
A
l
M
u
taz
,
M
.
A
b
d
a
ll
a
,
S
a
f
a
a
i
Dre
ss
,
a
n
d
Na
z
a
r
Zak
i.
"
De
tec
ti
o
n
o
f
m
a
ss
e
s
in
d
ig
it
a
l
m
a
m
m
o
g
ra
m
u
sin
g
se
c
o
n
d
o
rd
e
r
sta
ti
stics
a
n
d
a
rti
f
icia
l
n
e
u
ra
l
n
e
tw
o
rk
.
"
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
Co
mp
u
ter
S
c
ien
c
e
&
In
fo
rm
a
ti
o
n
T
e
c
h
n
o
l
o
g
y
(
IJ
CS
IT
)
,
v
o
l.
3
,
n
o
.
3
,
p
p
.
1
7
6
-
1
8
6
,
Ju
n
e
2
0
1
1
.
[2
2
]
M
a
h
m
o
o
d
,
F
a
leh
H.
,
a
n
d
W
a
f
a
a
A
.
A
b
b
a
s.
"
T
e
x
tu
re
F
e
a
tu
re
s
A
n
a
ly
sis
u
sin
g
G
r
a
y
L
e
v
e
l
Co
-
o
c
c
u
rre
n
c
e
M
a
tri
x
f
o
r
A
b
n
o
rm
a
li
t
y
D
e
tec
ti
o
n
in
Ch
e
st
CT
I
m
a
g
e
s.
"
Ira
q
i
J
o
u
rn
a
l
o
f
S
c
ie
n
c
e
,
v
o
l.
57
,
n
o
.
1
A
,
p
p
.
2
7
9
-
2
8
8
,
2
0
1
6
.
[2
3
]
Ha
re
f
a
,
Je
k
li
n
,
A
le
x
a
n
d
e
r
A
le
x
a
n
d
e
r,
a
n
d
M
e
ll
isa
P
ra
ti
w
i.
"
Co
m
p
a
riso
n
c
las
sif
ier:
su
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
(S
VM
)
a
n
d
K
-
n
e
a
re
st
n
e
ig
h
b
o
r
(K
-
NN
)
i
n
d
ig
it
a
l
m
a
m
m
o
g
ra
m
i
m
a
g
e
s."
J
u
rn
a
l
In
fo
rm
a
t
ika
d
a
n
S
istem
In
f
o
rm
a
si
,
v
o
l.
2
,
n
o
.
2
,
p
p
.
35
-
40
,
2
0
1
6
.
Evaluation Warning : The document was created with Spire.PDF for Python.