T
E
L
K
O
M
NIKA
T
elec
o
mm
un
ica
t
io
n,
Co
m
pu
t
ing
,
E
lect
ro
nics
a
nd
Co
ntr
o
l
Vo
l.
19
,
No
.
6
,
Dec
em
b
er
2
0
2
1
,
p
p
.
1
9
0
2
~
1
9
1
2
I
SS
N:
1
6
9
3
-
6
9
3
0
,
ac
cr
ed
ited
First Gr
ad
e
b
y
Ke
m
e
n
r
is
te
k
d
i
k
ti,
Dec
r
ee
No
: 2
1
/E/KPT
/2
0
1
8
DOI
: 1
0
.
1
2
9
2
8
/
T
E
L
KOM
NI
KA
.
v
1
9
i6
.
1
8
4
0
2
1902
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//jo
u
r
n
a
l.u
a
d
.
a
c.
id
/in
d
ex
.
p
h
p
/TELK
OM
N
I
K
A
So
lid w
a
ste clas
sifica
tion usin
g
pyra
m
id
scene
pars
ing
net
w
o
rk
seg
m
en
tatio
n and
co
m
bi
ned
featur
es
K
ha
dija
h,
Su
km
a
w
a
t
i N
ur
E
nd
a
h,
Ret
no
K
us
u
m
a
nin
g
ru
m
, R
i
s
m
iy
a
t
i
,
P
riy
o
Sid
i
k
Sa
s
o
ng
k
o
,
I
f
f
a
Z
a
ina
n Ni
s
a
De
p
a
rtme
n
t
o
f
In
f
o
rm
a
ti
c
s,
F
a
c
u
lt
y
o
f
S
c
ien
c
e
a
n
d
M
a
th
e
m
a
ti
c
s
,
Un
iv
e
rsitas
Dip
o
n
e
g
o
ro
,
S
e
m
a
ra
n
g
,
In
d
o
n
e
sia
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
No
v
3
,
2
0
20
R
ev
i
s
ed
Ma
y
2
1
,
2
0
2
1
A
cc
ep
ted
J
u
n
5
,
2
0
21
S
o
li
d
w
a
ste
p
ro
b
lem
b
e
c
o
m
e
a
se
r
io
u
s
issu
e
f
o
r
th
e
c
o
u
n
tri
e
s
a
ro
u
n
d
th
e
w
o
rld
sin
c
e
th
e
a
m
o
u
n
t
o
f
g
e
n
e
ra
ted
so
li
d
w
a
ste
in
c
re
a
s
e
a
n
n
u
a
ll
y
.
A
s
a
n
e
ff
o
rt
to
re
d
u
c
e
a
n
d
re
u
se
o
f
so
li
d
w
a
ste
,
a
c
las
si
f
ica
ti
o
n
o
f
so
l
id
w
a
ste
i
m
a
g
e
is
n
e
e
d
e
d
to
su
p
p
o
rt
a
u
t
o
m
a
ti
c
w
a
st
e
so
rti
n
g
.
In
t
h
e
im
a
g
e
c
las
sif
ic
a
ti
o
n
t
a
sk
,
i
m
a
g
e
se
g
m
e
n
tatio
n
a
n
d
f
e
a
tu
re
e
x
trac
ti
o
n
p
lay
i
m
p
o
rtan
t
ro
les
.
T
h
is
re
se
a
rc
h
a
p
p
li
e
s
re
c
e
n
t
d
e
e
p
lea
n
in
g
-
b
a
se
d
se
g
m
e
n
tatio
n
,
n
a
m
e
l
y
p
y
ra
m
id
sc
e
n
e
p
a
rsin
g
n
e
tw
o
rk
(
P
S
P
Ne
t
)
.
W
e
a
lso
u
s
e
v
a
rio
u
s
c
o
m
b
in
a
ti
o
n
o
f
im
a
g
e
f
e
a
tu
r
e
e
x
t
ra
c
ti
o
n
(c
o
lo
r
,
tex
tu
re
,
a
n
d
sh
a
p
e
)
to
se
a
rc
h
f
o
r
th
e
b
e
st
c
o
m
b
in
a
ti
o
n
o
f
f
e
a
tu
re
s.
A
s
a
c
o
m
p
a
riso
n
,
w
e
a
lso
p
e
rf
o
rm
e
x
p
e
ri
m
e
n
t
w
it
h
o
u
t
u
sin
g
se
g
m
e
n
tatio
n
t
o
se
e
th
e
e
ff
e
c
t
o
f
P
S
P
Ne
t.
T
h
e
n
,
su
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
(
S
V
M
)
is ap
p
li
e
d
in
th
e
e
n
d
a
s
c
la
ss
if
ic
a
ti
o
n
a
lg
o
rit
h
m
.
Ba
se
d
o
n
th
e
re
su
lt
o
f
e
x
p
e
ri
m
e
n
t,
it
c
a
n
b
e
c
o
n
c
l
u
d
e
d
t
h
a
t
g
e
n
e
ra
ll
y
a
p
p
l
y
in
g
se
g
m
e
n
tatio
n
p
ro
v
i
d
e
b
e
tt
e
r
so
u
rc
e
f
o
r
f
e
a
tu
re
e
x
trac
ti
o
n
,
e
sp
e
c
ially
in
c
o
l
o
r
a
n
d
sh
a
p
e
f
e
a
t
u
re
,
h
e
n
c
e
in
c
re
a
se
th
e
a
c
c
u
ra
c
y
o
f
c
las
si
f
ier
.
It
is
a
lso
o
b
se
rv
e
d
t
h
a
t
th
e
m
o
st
im
p
o
rtan
t
f
e
a
tu
re
i
n
th
is
p
r
o
b
lem
is
c
o
l
o
r
f
e
a
tu
re
.
Ho
w
e
v
e
r,
th
e
a
c
c
u
ra
c
y
o
f
c
las
sif
ier
in
c
re
a
se
if
a
d
d
it
io
n
a
l
f
e
a
tu
re
s
a
re
in
tro
d
u
c
e
d
.
T
h
e
h
ig
h
e
st
a
c
c
u
ra
c
y
o
f
7
6
.
4
9
%
is
a
c
h
iev
e
d
w
h
e
n
P
S
P
Ne
t
se
g
m
e
n
tatio
n
is
a
p
p
li
e
d
a
n
d
a
ll
c
o
m
b
in
a
ti
o
n
o
f
f
e
a
tu
re
s
a
re
u
se
d
.
K
ey
w
o
r
d
s
:
Featu
r
e
ex
tr
ac
tio
n
P
SP
Net
Seg
m
en
tatio
n
SVM
W
aste c
lass
i
f
icatio
n
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
:
Kh
ad
ij
ah
Dep
ar
t
m
en
t o
f
I
n
f
o
r
m
atic
s
Facu
lt
y
o
f
Sc
ien
ce
a
n
d
Ma
th
e
m
atic
s
Un
i
v
er
s
ita
s
Dip
o
n
eg
o
r
o
P
r
o
f
.
H.
So
ed
a
r
to
St.
,
SH.
T
e
m
b
a
lan
g
Se
m
ar
an
g
5
0
2
7
5
,
I
n
d
o
n
esia
E
m
ail:
k
h
ad
ij
ah
@
li
v
e.
u
n
d
ip
.
ac
.
id
1.
I
NT
RO
D
UCT
I
O
N
So
lid
w
aste
p
r
o
b
le
m
b
ec
o
m
e
a
s
er
io
u
s
is
s
u
e
f
o
r
th
e
co
u
n
tr
i
es
ar
o
u
n
d
th
e
w
o
r
ld
s
in
ce
t
h
e
a
m
o
u
n
t
o
f
g
en
er
ated
s
o
lid
w
a
s
te
in
cr
ea
s
e
an
n
u
all
y
.
I
n
2
0
1
6
th
e
to
tal
g
en
er
atio
n
o
f
s
o
lid
w
aste
b
y
th
e
w
o
r
ld
’
s
cities
w
a
s
u
p
to
2
.
0
1
b
illi
o
n
to
n
n
es.
I
t
w
as
eq
u
al
to
0
.
7
4
k
ilo
g
r
a
m
o
f
s
o
lid
w
aste
g
e
n
er
ated
b
y
a
p
er
s
o
n
in
a
d
a
y
.
T
h
i
s
n
u
m
b
er
is
esti
m
ated
to
in
cr
ea
s
e
b
y
7
0
% o
r
u
p
to
3
.
4
0
b
illi
o
n
to
n
n
e
s
o
f
s
o
lid
w
aste i
n
2
0
5
0
.
P
o
p
u
latio
n
g
r
o
w
t
h
an
d
u
r
b
an
izatio
n
ar
e
th
e
m
o
s
t
s
ig
i
n
i
f
ica
n
t
f
ac
to
r
s
th
at
tr
i
g
g
er
th
e
in
cr
ea
s
e
in
th
e
a
m
o
u
n
t
o
f
w
aste.
P
o
o
r
m
an
a
g
e
m
e
n
t
o
f
w
as
te
m
a
y
cr
ea
te
s
er
io
u
s
p
r
o
b
le
m
r
elate
d
to
h
ea
lt
h
,
s
a
f
et
y
,
an
d
e
n
v
ir
o
n
m
en
t.
T
h
er
ef
o
r
e,
p
r
o
p
e
r
w
a
s
te
m
a
n
a
g
e
m
en
t
s
tr
ateg
y
is
n
ee
d
ed
to
m
i
n
i
m
ize
s
u
c
h
n
e
g
a
tiv
e
i
m
p
ac
ts
[
1
]
.
On
e
o
f
th
e
e
f
f
o
r
t
to
r
ed
u
ce
th
e
n
u
m
b
er
o
f
s
o
lid
w
aste
i
s
b
y
i
m
p
r
o
v
i
n
g
th
e
w
aste
r
e
u
s
ab
il
it
y
.
W
ast
e
s
o
r
tin
g
p
la
y
s
s
i
g
n
if
ican
t
r
o
le
to
s
u
p
p
o
r
t
th
e
w
as
te
r
eu
s
ab
ilit
y
[
2
]
.
Sin
ce
t
h
e
n
u
m
b
er
o
f
w
a
s
te
is
g
r
ea
t
a
n
d
th
e
a
w
ar
n
ess
o
f
p
eo
p
le
in
w
a
s
te
s
o
r
tin
g
is
s
ti
ll
lo
w
,
an
a
u
to
m
at
ic
w
aste
s
o
r
ti
n
g
i
s
n
ee
d
ed
.
T
h
e
s
t
ar
ti
n
g
p
o
in
t
to
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
S
o
lid
w
a
s
te
cla
s
s
ifica
tio
n
u
s
in
g
p
yra
mid
s
ce
n
e
p
a
r
s
in
g
n
etw
o
r
k
s
eg
men
ta
tio
n
a
n
d
…
(
K
h
a
d
ija
h
)
1903
p
r
o
d
u
ce
an
au
to
m
a
tic
w
aste
s
o
r
tin
g
is
b
y
b
u
ild
i
n
g
a
clas
s
if
i
ca
tio
n
m
o
d
el
th
at
m
a
y
r
ec
o
g
n
i
ze
th
e
t
y
p
e
o
f
w
a
s
te
i
m
a
g
e.
So
m
e
p
r
ev
io
u
s
s
t
u
d
ies
h
av
e
b
ee
n
in
v
es
tig
a
ted
th
e
u
s
e
o
f
m
ac
h
in
e
lear
n
i
n
g
al
g
o
r
ith
m
to
class
i
f
y
o
r
r
ec
o
g
n
ize
w
aste
i
m
ag
e.
M
u
s
ta
f
f
a
et
a
l.
[
3
]
an
d
T
o
r
r
es
-
g
r
ac
ia
et
a
l.
[
4
]
class
if
ied
w
as
te
i
m
a
g
e
in
to
th
r
ee
class
es
u
s
i
n
g
co
n
v
e
n
tio
n
al
m
ac
h
i
n
e
le
ar
n
in
g
alg
o
r
it
h
m
a
n
d
w
er
e
ab
le
to
ac
h
iev
e
g
o
o
d
ac
cu
r
ac
ies,
b
u
t
th
eir
ex
p
er
i
m
en
t
u
s
ed
o
n
l
y
2
0
s
a
m
p
les
in
ea
ch
class
.
T
h
er
ef
o
r
e,
th
e
g
en
er
ali
za
tio
n
o
f
th
e
r
esu
lti
n
g
m
o
d
el
f
o
r
class
if
y
i
n
g
th
e
v
ar
iet
y
o
f
r
ea
l
w
a
s
te
i
m
a
g
e
co
u
ld
n
o
t
b
e
ass
u
r
ed
.
A
d
ed
ej
i
a
n
d
W
an
g
[
5
]
an
d
C
o
s
ta
et
a
l
.
[
6
]
class
if
ied
w
ast
e
i
m
a
g
e
u
s
in
g
d
ee
p
lear
n
in
g
m
o
d
el
an
d
m
o
r
e
n
u
m
b
er
o
f
s
a
m
p
les,
b
u
t
t
h
e
y
u
s
ed
th
e
ca
p
t
u
r
e
o
f
w
a
s
te
i
m
ag
e
d
ir
ec
tl
y
w
ith
o
u
t
an
y
s
eg
m
e
n
ta
tio
n
.
On
th
e
o
t
h
er
h
an
d
,
s
eg
m
en
tatio
n
i
s
o
n
e
o
f
th
e
m
o
s
t
i
m
p
o
r
tan
t
p
ar
t
in
th
e
i
m
a
g
e
p
r
ep
r
o
ce
s
s
in
g
.
P
o
o
r
s
eg
m
en
tatio
n
r
e
s
u
l
t
m
a
y
d
e
g
r
ad
e
th
e
p
er
f
o
r
m
a
n
ce
o
f
th
e
s
u
b
s
eq
u
en
t
p
r
o
ce
s
s
es,
s
u
c
h
as f
ea
t
u
r
e
ex
tr
ac
tio
n
an
d
clas
s
i
f
icatio
n
[
7
]
.
Seg
m
en
tatio
n
i
s
t
h
e
p
r
o
ce
s
s
o
f
p
ar
titi
o
n
i
n
g
a
n
i
m
a
g
e
in
to
d
if
f
er
e
n
t
s
e
v
er
al
d
is
j
o
in
t
s
u
b
s
et
[
8
]
,
f
o
r
ex
a
m
p
le
p
ar
titi
o
n
i
n
g
an
i
m
ag
e
in
to
b
ac
k
g
r
o
u
n
d
an
d
f
o
r
eg
r
o
u
n
d
.
Seg
m
e
n
tatio
n
ca
n
also
b
e
u
s
ed
to
ex
tr
ac
t
r
eg
io
n
o
f
in
ter
est
o
f
an
i
m
ag
e
[
7
]
.
State
o
f
th
e
ar
t
o
f
s
eg
m
en
t
atio
n
m
et
h
o
d
s
ar
e
k
i
n
d
o
f
d
ee
p
lear
n
in
g
al
g
o
r
ith
m
w
it
h
s
p
ec
ial
ar
c
h
itect
u
r
e,
s
u
c
h
as
en
co
d
er
-
d
ec
o
d
er
w
h
ic
h
h
a
v
e
b
etter
p
er
f
o
r
m
a
n
ce
t
h
an
t
h
e
co
n
v
e
n
tio
n
al
o
n
e
(
s
u
c
h
as
th
r
e
s
h
o
ld
i
n
g
m
et
h
o
d
)
[
9
]
.
P
y
r
a
m
id
s
ce
n
e
p
ar
s
i
n
g
n
et
w
r
o
k
(
P
SP
Net
)
is
a
k
i
n
d
o
f
d
ee
p
lear
n
in
g
n
et
w
o
r
k
th
at
ca
n
b
e
u
s
ed
f
o
r
s
e
m
a
n
tic
i
m
ag
e
s
e
g
m
e
n
tat
io
n
.
P
SP
Net
s
u
cc
es
s
f
u
ll
y
o
u
tp
er
f
o
r
m
e
s
o
t
h
er
d
ee
p
lear
n
in
g
b
ased
s
eg
m
e
n
tatio
n
in
s
o
m
e
lar
g
e
b
en
ch
m
ar
k
d
ataset,
s
u
c
h
a
s
f
u
l
l
y
co
n
v
o
l
u
tio
n
al
n
e
t
w
o
r
k
(
F
C
N
)
,
Dee
p
L
ab
,
d
ee
p
p
ar
s
in
g
n
e
t
w
o
r
k
(
DP
N
)
an
d
L
ap
lacia
n
p
y
r
a
m
id
r
ec
o
n
s
tr
u
c
t
io
n
an
d
r
efin
e
m
e
n
t
(
L
R
R
)
.
P
SP
Net
is
ab
le
to
ac
h
iev
e
b
etter
s
eg
m
e
n
tat
io
n
r
es
u
lt
b
ec
au
s
e
it
co
n
s
id
er
s
g
lo
b
al
co
n
tex
t
o
f
t
h
e
i
m
ag
e
a
n
d
u
s
es
p
y
r
a
m
id
p
o
o
lin
g
m
o
d
u
le
to
o
b
tain
d
if
f
er
e
n
t r
eg
io
n
b
ase
d
co
n
tex
t o
f
a
n
i
m
a
g
e
[
1
0
]
.
I
n
ad
d
itio
n
,
f
ea
tu
r
e
e
x
tr
ac
tio
n
o
f
th
e
i
m
ag
e
m
u
s
t
b
e
d
eter
m
i
n
ed
p
r
o
p
e
r
ly
to
ac
h
ie
v
e
g
o
o
d
class
i
f
icatio
n
r
esu
lt
[
1
1
]
,
[
1
2
]
.
Featu
r
e
ex
tr
ac
tio
n
is
a
i
m
ed
to
e
x
tr
ac
t
r
ele
v
a
n
t
s
u
b
s
et
o
f
f
ea
t
u
r
es f
r
o
m
a
n
i
m
ag
e
an
d
to
r
ed
u
ce
th
e
lar
g
e
d
i
m
e
n
s
io
n
o
f
i
m
a
g
e
t
o
th
e
lo
w
er
d
i
m
en
s
io
n
al
s
et
o
f
i
m
a
g
e
f
ea
tu
r
e
s
[
1
3
]
.
C
o
lo
r
,
tex
tu
r
e,
an
d
s
h
ap
e
ar
e
th
e
m
o
s
t v
is
u
al
f
ea
tu
r
es
e
x
tr
ac
ted
f
r
o
m
a
n
i
m
a
g
e.
C
o
lo
r
m
o
m
en
ts
is
o
n
e
o
f
t
h
e
s
i
m
p
lest
co
lo
r
f
ea
tu
r
e
co
m
p
ar
ed
to
th
e
o
th
er
,
s
u
c
h
as
co
lo
r
h
i
s
to
g
r
a
m
,
co
lo
r
co
h
er
en
ce
v
ec
to
r
,
an
d
co
lo
r
co
r
r
elo
g
r
am
[
1
4
]
.
C
o
lo
r
m
o
m
en
t
is
al
s
o
p
r
o
v
en
to
b
e
ef
f
ec
tiv
e
a
n
d
ef
f
i
cien
t f
o
r
ex
tr
ac
ti
n
g
co
lo
r
f
ea
t
u
r
es
o
f
an
i
m
a
g
e
[
1
5
]
.
I
n
ad
d
itio
n
tex
t
u
r
e
f
ea
tu
r
e
i
s
also
i
m
p
o
r
tan
t
to
ex
tr
ac
t
t
h
e
r
elatio
n
s
h
ip
f
r
o
m
n
e
ig
h
b
o
r
in
g
p
ix
el.
Gr
a
y
le
v
el
co
-
o
cc
u
r
en
c
e
m
atr
i
x
(
GL
C
M)
i
s
o
n
e
o
f
t
h
e
p
o
p
u
lar
te
x
t
u
r
e
-
b
as
ed
f
ea
tu
r
e
e
x
tr
ac
tio
n
t
h
at
h
as
b
ee
n
s
u
cc
es
s
f
u
l
l
y
ap
p
lied
i
n
m
a
n
y
co
m
p
u
ter
v
i
s
io
n
p
r
o
b
lem
[
1
6
]
-
[
1
8
]
.
T
h
e
o
th
er
im
p
o
r
ta
n
t
i
m
a
g
e
f
ea
t
u
r
e
to
d
esc
r
ib
e
th
e
o
b
j
ec
t
o
f
an
i
m
ag
e
is
s
h
ap
e
f
ea
tu
r
e.
So
m
e
m
o
r
p
o
h
lo
g
ical
f
ea
tu
r
es,
s
u
c
h
as
ar
ea
,
p
er
im
eter
,
m
aj
o
r
an
d
m
i
n
o
r
ax
is
,
ce
n
tr
o
id
-
x
,
ce
n
tr
o
id
-
y
,
r
o
u
n
d
n
e
s
s
,
r
ec
tan
g
u
l
ar
it
y
,
ec
ce
n
tr
icit
y
a
n
d
elo
n
g
atio
n
,
ca
n
b
e
u
s
ed
as
s
h
ap
e
d
escr
ip
to
r
[
1
9
]
.
I
n
ad
d
i
tio
n
,
Hu
m
o
m
e
n
t
i
s
also
i
m
p
o
r
tan
t
to
ex
tr
ac
t
s
h
ap
e
f
ea
tu
r
es.
Hu
m
o
m
en
t
is
r
eg
i
o
n
-
b
ased
m
et
h
o
d
th
at
u
s
es
s
ec
o
n
d
an
d
th
ir
d
o
r
d
er
ce
n
tr
al
m
o
m
e
n
ts
a
n
d
co
n
s
tr
u
c
ts
s
e
v
en
in
v
ar
ia
n
t
m
o
m
e
n
ts
w
h
o
s
e
v
alu
e
s
ar
e
n
o
t
a
f
f
ec
ted
w
h
e
n
th
e
i
m
a
g
e
i
s
tr
an
s
lated
,
r
o
tated
,
o
r
s
ca
led
[
2
0
]
.
I
n
th
is
r
esear
ch
,
w
e
p
r
o
p
o
s
e
t
h
e
ap
p
licatio
n
o
f
P
SP
Net
as
s
eg
m
e
n
tat
io
n
to
p
r
o
v
id
e
g
o
o
d
s
o
u
r
ce
f
o
r
f
ea
t
u
r
e
ex
tr
ac
tio
n
.
W
e
also
u
s
e
v
ar
io
u
s
co
m
b
i
n
atio
n
o
f
i
m
a
g
e
f
ea
t
u
r
e
ex
tr
ac
tio
n
(
co
lo
r
,
te
x
tu
r
e
,
an
d
s
h
ap
e)
to
s
ea
r
ch
f
o
t
h
e
b
est
co
m
b
i
n
atio
n
o
f
f
ea
tu
r
e
s
f
o
r
s
o
lid
w
aste
i
m
a
g
e
clas
s
if
icatio
n
.
As
a
co
m
p
ar
i
s
o
n
,
w
e
al
s
o
p
er
f
o
r
m
ex
p
er
i
m
en
t
w
it
h
o
u
t
u
s
i
n
g
s
e
g
m
en
ta
tio
n
to
s
ee
t
h
e
ef
f
ec
t
o
f
P
SP
Net.
T
h
en
,
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
i
n
e
(
SVM)
is
ap
p
lied
at
th
e
en
d
a
s
a
cla
s
s
i
f
ier
.
S
VM
i
s
a
b
i
n
ar
y
class
i
f
ica
tio
n
al
g
o
r
it
h
m
p
r
o
p
o
s
ed
b
y
C
o
r
tes
an
d
Vap
n
ik
[
2
1
]
w
h
ic
h
w
o
r
k
s
b
y
f
in
d
in
g
t
h
e
o
p
ti
m
al
h
y
p
er
p
la
n
e
to
m
ax
i
m
ize
t
h
e
s
ep
ar
atio
n
b
et
w
ee
n
b
in
ar
y
clas
s
d
ata
.
SVM
h
a
s
b
ee
n
s
u
cc
e
s
s
f
u
ll
y
ap
p
lied
in
v
ar
io
u
s
clas
s
i
f
icatio
n
p
r
o
b
le
m
an
d
p
r
o
v
e
n
t
o
b
etter
th
an
o
t
h
er
p
o
p
u
lar
class
if
icatio
n
alg
o
r
ith
m
,
s
u
c
h
as
ar
tif
icial
n
e
u
r
al
n
et
w
o
r
k
(
A
NN)
[
2
2
]
,
[
2
3
]
,
Naïv
e
B
ay
e
s
class
i
f
ier
d
an
r
an
d
o
m
f
o
r
est
[
2
4
]
.
2.
RE
S
E
ARCH
M
E
T
H
O
DO
L
O
G
Y
Fig
u
r
e
1
s
h
o
w
s
th
e
s
ta
g
es
o
f
p
r
o
ce
s
s
in
th
is
r
esear
ch
.
First,
th
e
i
m
ag
e
d
ataset
is
s
e
g
m
en
t
ed
b
y
u
s
i
n
g
P
SP
Net
s
eg
m
e
n
tat
io
n
,
t
h
en
th
e
p
r
o
ce
s
s
is
co
n
tin
u
ed
b
y
f
ea
t
u
r
e
ex
tr
ac
tio
n
,
clas
s
i
f
icatio
n
,
an
d
ev
al
u
atio
n
.
A
s
a
co
m
p
ar
is
o
n
,
to
ex
a
m
i
n
e
t
h
e
e
f
f
ec
t o
f
P
SP
Net,
w
e
also
p
er
f
o
r
m
e
x
p
er
i
m
e
n
t
w
it
h
o
u
t
u
s
in
g
s
eg
m
e
n
tat
io
n
,
h
e
n
ce
th
e
p
r
o
ce
s
s
o
f
s
e
g
m
en
ta
tio
n
i
n
Fig
u
r
e
1
is
s
k
ip
p
ed
.
P
S
P
N
e
t
S
e
g
m
e
n
t
a
t
i
o
n
C
l
a
s
s
i
f
i
c
a
t
i
o
n
C
o
l
o
r
T
e
x
t
u
r
e
S
h
a
p
e
F
e
a
t
u
r
e
E
x
t
r
a
c
t
i
o
n
E
v
a
l
u
a
t
i
o
n
I
m
a
g
e
D
a
t
a
s
e
t
Fig
u
r
e
1
.
R
esear
ch
m
et
h
o
d
o
lo
g
y
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
19
,
No
.
6
,
Dec
em
b
er
2
0
2
1
:
1
9
0
2
-
1912
1904
2
.
1
.
Da
t
a
s
et
P
u
b
lic
tr
ash
i
m
a
g
e
d
ataset
f
r
o
m
T
r
ash
n
e
t
ar
e
u
s
ed
in
th
is
r
esear
ch
as
s
o
u
r
ce
d
ata
f
o
r
co
n
d
u
cti
n
g
ex
p
er
i
m
e
n
ts
.
T
r
ash
n
et
d
ataset
w
a
s
co
llected
b
y
Yan
g
an
d
T
h
u
n
g
[
2
5
]
.
T
h
is
d
ataset
co
n
t
ain
2
,
5
2
7
tr
ash
i
m
ag
e
s
o
f
2
2
4
×
2
2
4
p
ix
els
w
h
ich
is
g
r
o
u
p
ed
in
to
s
ix
class
es:
g
las
s
(
501
)
,
p
a
p
er
(
5
9
4
)
,
ca
r
d
b
o
ar
d
(
403
)
,
p
last
ic
(
482
)
,
m
etal
(
410
)
,
an
d
tr
ash
(
137
)
.
A
s
a
m
p
le
i
m
a
g
e
f
r
o
m
ea
c
h
cla
s
s
ca
n
b
e
s
ee
n
in
Fig
u
r
e
2
[
2
5
]
.
g
las
s
p
ap
er
ca
r
d
b
o
a
r
d
p
last
ic
m
etal
tr
ash
Fig
u
r
e
2
.
Sa
m
p
le
i
m
ag
e
f
r
o
m
ea
ch
class
o
f
T
r
ash
n
et
d
ataset
[
2
5
]
2
.
2
.
P
SPNe
t
f
o
r
im
a
g
e
s
eg
menta
t
io
n
P
SP
Net
is
p
e
r
f
o
r
m
ed
to
g
en
er
a
te
s
eg
m
en
ted
b
in
ar
y
i
m
ag
e,
t
h
en
th
e
b
o
u
n
d
i
n
g
b
o
x
o
f
s
eg
m
e
n
ted
i
m
a
g
e
ar
e
ca
lcu
lated
an
d
im
a
g
e
is
cr
o
p
p
ed
s
o
th
at
o
n
ly
t
h
e
m
ai
n
o
b
j
ec
t
r
em
ai
n
.
P
SP
Net
is
a
k
in
d
o
f
d
ee
p
lear
n
in
g
n
et
w
o
r
k
f
o
r
s
e
m
a
n
tic
i
m
ag
e
s
eg
m
e
n
tat
io
n
.
P
SP
Net
o
u
tp
er
f
o
r
m
ed
F
C
N
b
ased
s
eg
m
e
n
tati
o
n
b
ec
au
s
e
P
SP
Net
co
n
s
id
er
g
lo
b
al
co
n
tex
t
o
f
th
e
i
m
ag
e
a
n
d
u
s
es
p
y
r
a
m
id
p
o
o
lin
g
m
o
d
u
le
to
o
b
tain
d
if
f
er
e
n
t
r
eg
io
n
ba
s
ed
co
n
tex
t
o
f
an
i
m
ag
e.
T
h
e
ar
ch
itect
u
r
e
o
f
P
SP
Net
is
s
h
o
w
n
in
F
ig
u
r
e
3
[
1
0
]
.
(
a)
(
b
)
(
c)
(
d
)
Fig
u
r
e
3
.
T
h
e
A
r
ch
itect
u
r
e
o
f
P
SP
Net
[
1
0
]
: (
a)
i
n
p
u
t i
m
ag
e,
(
b
)
f
ea
tu
r
e
m
ap
,
(
c)
p
y
r
a
m
id
p
o
o
lin
g
m
o
d
u
le,
an
d
(
d
)
f
in
al
p
r
ed
ictio
n
First,
an
in
p
u
t
i
m
ag
e
(
Fi
g
u
r
e
3
(
a
)
)
is
f
ed
in
to
co
n
v
o
lu
tio
n
al
n
eu
r
al
n
e
t
w
o
r
k
s
(
C
N
N
)
w
i
th
d
ilated
n
et
w
o
r
k
s
tr
ateg
y
w
h
ic
h
g
en
ar
ate
f
ea
t
u
r
e
m
ap
(
F
ig
u
r
e
3
(
b
)
)
w
i
th
s
ize
1
/8
o
f
o
r
ig
in
al
in
p
u
t
i
m
a
g
e.
T
h
en
,
t
h
e
f
e
at
u
r
e
m
ap
is
f
o
r
w
ar
d
ed
to
th
e
p
y
r
a
m
id
p
o
o
lin
g
m
o
d
u
le
(
Fi
g
u
r
e
(
c
)
)
w
h
ic
h
g
en
er
ate
t
h
e
co
n
ca
ten
ated
f
ea
t
u
r
e
m
ap
in
t
h
e
e
n
d
o
f
t
h
e
m
o
d
u
le
.
I
n
th
e
la
s
t
s
tep
(
Fi
g
u
r
e
(
d
)
)
co
n
v
o
lu
tio
n
la
y
er
is
ap
p
lied
o
n
th
e
co
n
ca
te
n
ated
f
ea
t
u
r
e
m
ap
to
g
e
n
er
ate
th
e
f
in
al
p
r
ed
ictio
n
o
f
ea
c
h
p
ix
el
i
n
a
n
i
m
ag
e.
T
h
er
e
ar
e
f
o
u
r
o
p
er
atio
n
s
in
t
h
e
p
y
r
a
m
id
p
o
o
lin
g
m
o
d
u
le
a
s
[
1
0
]
,
[
2
6
]
:
Su
b
r
eg
io
n
a
v
er
ag
e
p
o
o
lin
g
E
ac
h
f
ea
t
u
r
e
m
ap
is
p
o
o
led
o
v
er
d
if
f
er
e
n
t
s
u
b
-
r
eg
io
n
to
o
b
tain
d
if
f
er
en
t
co
n
te
x
t
r
ep
r
s
en
t
atio
n
in
ea
ch
s
u
b
-
r
eg
io
n
.
I
n
t
h
e
f
ir
s
t
le
v
el
(
r
ed
)
,
th
e
g
lo
b
al
av
er
ag
e
p
o
o
lin
g
is
p
er
f
o
r
m
ed
in
ea
ch
f
ea
t
u
r
e
m
ap
.
T
h
e
r
esu
lt
is
a
s
in
g
le
b
in
o
u
tp
u
t f
o
r
ea
ch
f
at
u
r
e
m
ap
.
I
n
th
e
s
ec
o
n
d
lev
el
(
o
r
an
g
e)
,
th
ir
d
lev
el
(
b
lu
e)
,
an
d
f
o
u
r
th
lev
el
(
g
r
ee
n
)
,
ea
ch
f
ea
tu
r
e
m
ap
i
s
d
i
v
id
ed
in
to
2
×
2
,
3
×
3
an
d
6
×
6
s
u
b
-
r
e
g
io
n
,
re
s
p
ec
ti
v
el
y
,
t
h
en
e
ac
h
s
u
b
-
r
eg
io
n
is
p
o
o
led
b
y
av
er
ag
e
p
o
o
lin
g
.
C
o
n
v
o
lu
tio
n
T
h
e
1
x
1
co
n
v
o
lu
tio
n
is
p
er
f
o
r
m
ed
at
ea
ch
lev
el
to
r
ed
u
ce
s
th
e
s
ize
o
f
f
ea
t
u
r
e
m
ap
at
ea
ch
lev
el
in
to
1
/N
o
f
th
e
o
r
ig
i
n
al
o
n
e
(
b
lack
)
w
h
er
e
N
is
th
e
lev
e
l size
o
f
p
y
r
a
m
id
.
Up
s
a
m
p
li
ng
Up
s
a
m
p
li
n
g
i
s
p
er
f
o
r
m
ed
b
y
u
s
in
g
b
ili
n
ea
r
in
ter
p
o
latio
n
to
m
ak
e
ea
c
h
f
ea
t
u
r
e
m
ap
h
a
v
e
t
h
e
eq
u
al
s
ize
as
th
e
o
r
ig
i
n
al
o
n
e
(
b
lack
)
.
C
o
n
ca
te
n
atio
n
T
h
e
o
r
ig
in
al
f
ea
tu
r
e
m
ap
(
b
lack
)
an
d
all
u
p
s
a
m
p
led
f
ea
t
u
r
e
m
ap
f
r
o
m
th
e
f
ir
s
t
to
f
o
u
r
th
le
v
el
ar
e
co
n
ca
ten
ated
an
d
t
h
e
r
esu
lt is
f
o
r
w
ar
d
ed
to
co
n
v
o
lu
tio
n
al
la
y
er
f
o
r
p
r
ed
ictio
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
S
o
lid
w
a
s
te
cla
s
s
ifica
tio
n
u
s
in
g
p
yra
mid
s
ce
n
e
p
a
r
s
in
g
n
etw
o
r
k
s
eg
men
ta
tio
n
a
n
d
…
(
K
h
a
d
ija
h
)
1905
T
h
e
p
r
o
ce
s
s
o
f
s
eg
m
e
n
tatio
n
u
s
i
n
g
P
SP
Net
co
n
s
is
ts
o
f
tr
ai
n
in
g
a
n
d
test
in
g
.
Data
s
e
t
is
d
iv
id
ed
in
to
7
0
%
o
f
tr
ain
in
g
d
ata,
1
5
%
o
f
v
alid
atio
n
d
ata
,
an
d
1
5
%
o
f
test
in
g
d
ata.
T
r
ain
in
g
is
p
er
f
o
r
m
ed
u
s
in
g
s
o
m
e
co
m
b
i
n
atio
n
o
f
h
y
p
er
p
ar
a
m
et
er
:
lear
n
i
n
g
r
ate
(
0
.
0
0
1
,
0
.
0
0
0
1
,
an
d
0
.
0
0
0
0
1
)
an
d
b
atch
(
5
a
n
d
1
0
)
in
5
0
ep
o
ch
.
Af
ter
tr
ai
n
i
n
g
u
s
in
g
s
u
c
h
co
m
b
in
at
io
n
o
f
h
y
p
er
p
ar
a
m
eter
s
,
s
ix
m
o
d
els
o
f
i
m
a
g
e
s
e
g
m
e
n
tatio
n
ar
e
o
b
tain
ed
,
th
en
te
s
ti
n
g
d
ata
is
u
s
ed
to
ev
alu
ate
an
d
s
elec
t
th
e
b
est
m
o
d
el.
Dice
co
ef
f
icen
t
(
)
is
u
s
ed
to
ev
alu
ate
th
e
r
esu
lt
s
o
f
s
e
g
m
en
ta
tio
n
as
s
h
o
w
n
i
n
(
1
)
w
h
er
e
an
d
is
th
e
i
m
ag
e
r
eg
io
n
s
b
ein
g
co
m
p
ar
ed
[
2
7
]
.
=
2
|
∩
|
|
|
+
|
|
(
1
)
2
.
3
.
F
e
a
t
ure
ex
t
ra
ct
io
n
Featu
r
e
ex
tr
ac
tio
n
is
a
i
m
ed
to
ex
tr
ac
t
r
elev
a
n
t
s
u
b
s
et
o
f
f
ea
tu
r
es
f
r
o
m
an
i
m
a
g
e
[
1
3
]
.
T
h
is
r
esear
ch
u
s
e
s
th
r
ee
k
i
n
d
s
o
f
i
m
a
g
e
f
ea
t
u
r
es,
n
a
m
el
y
co
lo
r
f
ea
t
u
r
es
e
x
tr
ac
ted
b
y
u
s
in
g
co
lo
r
m
o
m
e
n
ts
,
tex
t
u
r
e
f
ea
t
u
r
es
ex
tr
ac
ted
b
y
u
s
i
n
g
g
r
a
y
le
v
el
c
o
-
o
cc
u
r
en
ce
m
at
r
ix
(
G
L
C
M
)
an
d
s
h
ap
e
f
ea
t
u
r
es.
E
x
p
er
i
m
en
t
s
ar
e
r
u
n
u
s
in
g
o
n
e
o
r
co
m
b
in
atio
n
o
f
s
u
ch
f
ea
t
u
r
es
to
o
b
tain
th
e
b
est
class
i
f
icati
o
n
r
esu
lt.
T
ab
le
1
s
h
o
w
s
th
e
co
m
p
ar
i
s
o
n
o
f
s
o
u
r
ce
i
m
a
g
e
f
o
r
ea
ch
f
ea
tu
r
e
ex
tr
ac
tio
n
m
et
h
o
d
.
W
h
en
u
s
i
n
g
P
SP
Net
s
eg
m
e
n
tatio
n
,
o
r
ig
in
al
r
ed
,
g
r
ee
n
,
an
d
b
lu
e
(
R
GB
)
i
m
ag
e
i
s
s
e
g
m
e
n
ted
r
e
s
u
lt
in
g
t
h
e
s
eg
m
e
n
ted
b
in
ar
y
i
m
a
g
e.
T
h
en
,
f
o
r
ex
tr
ac
tio
n
o
f
co
lo
r
an
d
tex
tu
r
e
f
ea
t
u
r
es,
th
e
i
m
a
g
e
is
cr
o
p
p
ed
ar
o
u
n
d
th
e
b
o
u
n
d
i
n
g
b
o
x
u
s
i
n
g
Op
en
C
V
lib
r
ar
y
,
f
i
n
d
co
n
to
u
r
.
W
h
en
t
h
e
s
eg
m
e
n
tatio
n
i
s
s
k
ip
p
ed
,
b
ef
o
r
e
s
h
ap
e
f
ea
t
u
r
es
ar
e
ex
tr
ac
te
d
,
ea
ch
i
m
ag
e
i
s
co
n
v
er
ted
in
to
b
in
ar
y
i
m
a
g
e
b
y
u
s
i
n
g
in
v
er
s
e
b
in
ar
y
th
r
e
s
h
o
ld
in
g
(
v
al
u
e
o
f
th
r
e
s
h
o
ld
=
1
2
8
)
.
T
ab
le
1
.
T
h
e
co
m
p
ar
is
o
n
o
f
i
n
p
u
t i
m
a
g
e
f
o
r
f
ea
tu
r
e
e
x
tr
ac
tio
n
b
et
w
ee
n
P
SP
Net
s
eg
m
e
n
tati
o
n
an
d
w
it
h
o
u
t
s
eg
m
e
n
tatio
n
F
e
a
t
u
r
e
U
si
n
g
P
S
P
N
e
t
se
g
me
n
t
a
t
i
o
n
W
i
t
h
o
u
t
se
g
me
n
t
a
i
o
n
C
o
l
o
r
C
r
o
p
p
e
d
b
o
u
n
d
i
n
g
b
o
x
o
f
se
g
me
n
t
e
d
R
G
B
i
mag
e
O
r
i
g
i
n
a
l
R
G
B
i
mag
e
T
e
x
t
u
r
e
C
r
o
p
p
e
d
b
o
u
n
d
i
n
g
b
o
x
o
f
se
g
me
n
t
e
d
g
r
a
y
sca
l
e
i
mag
e
T
r
a
n
sf
o
r
mat
i
o
n
f
r
o
m o
r
i
g
i
n
a
l
R
G
B
t
o
g
r
a
y
sca
l
e
i
mag
e
S
h
a
p
e
S
e
g
me
n
t
e
d
b
i
n
a
r
y
i
mag
e
T
r
a
n
sf
o
r
mat
i
o
n
f
r
o
m o
r
i
g
i
n
a
l
R
G
B
t
o
b
i
n
a
r
y
i
mag
e
u
si
n
g
i
n
v
e
r
se
b
i
n
a
r
y
t
r
e
sh
o
l
d
i
n
g
2
.
3
.
1
.
Co
lo
rm
o
m
ent
s
C
o
lo
r
f
ea
tu
r
e
is
v
is
u
al
f
ea
t
u
r
e
th
at
ca
n
b
e
u
s
ed
to
d
is
cr
i
m
i
n
at
e
o
r
r
ec
o
g
n
ize
v
is
u
al
i
n
f
o
r
m
at
io
n
.
I
f
th
e
co
lo
r
d
is
tr
ib
u
tio
n
o
f
an
i
m
a
g
e
is
i
n
ter
p
r
eted
as
a
p
r
o
b
ab
ilit
y
d
is
tr
ib
u
tio
n
,
th
e
n
co
lo
r
m
o
m
en
ts
ca
n
b
e
u
s
ed
to
ch
ar
ac
ter
ize
th
e
co
lo
r
d
is
tr
ib
u
tio
n
[
2
8
]
.
T
h
r
ee
co
lo
r
m
o
m
en
ts
(
m
ea
n
,
s
ta
n
d
ar
d
ev
iatio
n
,
an
d
s
k
e
w
n
ess
)
ar
e
ex
tr
ac
ted
f
o
r
ev
er
y
i
m
a
g
e
ch
a
n
n
el,
th
er
e
f
o
r
e
th
er
e
ar
e
9
n
u
m
er
ical
v
alu
e
s
e
x
tr
ac
ted
f
o
r
a
n
i
m
ag
e
in
R
GB
co
lo
r
s
p
ac
e.
Me
an
is
th
e
av
er
a
g
e
o
f
p
i
x
el
v
al
u
es
a
s
s
h
o
w
n
i
n
(
2
)
,
s
tan
d
ar
d
d
ev
iat
io
n
is
th
e
v
a
r
iatio
n
o
f
p
ix
e
l
v
alu
e
s
as
s
h
o
w
n
i
n
(
3
)
,
an
d
s
k
ew
n
es
s
is
th
e
d
eg
r
ee
o
f
as
y
m
m
etr
y
in
t
h
e
co
lo
r
d
is
tr
ib
u
tio
n
in
an
i
m
a
g
e
ch
a
n
n
el
as sh
o
w
n
i
n
(
4
)
.
is
to
tal
n
u
m
b
er
o
f
p
ix
el
in
ea
c
h
ch
a
n
n
e
l a
n
d
is
th
e
-
th
p
i
x
el
v
al
u
e
i
n
ch
a
n
n
e
l
[
1
5
]
.
=
1
∑
=
1
(
2
)
=
√
(
1
∑
(
−
)
2
=
1
)
(
3
)
=
√
(
1
∑
(
−
)
3
=
1
)
3
(
4
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
19
,
No
.
6
,
Dec
em
b
er
2
0
2
1
:
1
9
0
2
-
1912
1906
2
.
3
.
2
.
G
ra
y
lev
el
co
-
o
cc
urence
m
a
t
rix
(
G
L
CM
)
GL
C
M
is
a
m
et
h
o
d
f
o
r
ex
tr
ac
tin
g
te
x
t
u
r
e
f
ea
tu
r
e
s
o
f
an
i
m
a
g
e
.
First,
co
-
o
cc
u
r
en
ce
m
atr
i
x
P
is
cr
ea
ted
.
P
is
a
s
q
u
ar
e
m
a
tr
ix
w
h
o
s
e
s
ize
is
eq
u
al
to
th
e
n
u
m
b
er
o
f
g
r
a
y
in
te
n
s
it
y
v
alu
e
o
f
an
i
m
a
g
e.
E
ac
h
ele
m
e
n
t
i
n
th
e
m
atr
i
x
is
th
e
n
u
m
b
er
o
f
o
cc
u
r
en
ce
(
f
r
eq
u
e
n
c
y
)
o
f
t
w
o
n
eig
b
o
r
in
g
p
ix
el
in
s
p
ec
i
f
ic
o
r
ien
tatio
n
w
h
er
e
th
e
g
r
a
y
in
ten
s
it
y
v
alu
e
o
f
t
h
e
f
ir
s
t
p
ix
el
is
eq
u
al
to
an
d
g
r
ay
in
te
n
s
it
y
v
al
u
e
o
f
th
e
s
ec
o
n
d
p
ix
el
is
eq
u
al
to
[
2
9
]
.
Neig
h
b
o
r
in
g
p
ix
e
l
ca
n
b
e
s
ele
cted
b
ased
o
n
s
p
ec
if
ied
s
p
atial
o
r
ien
tatio
n
.
Fo
r
ex
a
m
p
le
w
h
e
n
th
e
o
r
ien
tat
io
n
is
0
0
,
th
en
t
h
e
n
e
ig
h
b
o
r
o
f
a
p
ix
e
l
is
a
p
ix
el
t
h
at
i
s
o
n
t
h
e
r
ig
h
t sid
e.
T
h
e
r
esu
lti
n
g
G
L
C
M
m
at
r
ix
ca
n
b
e
o
b
tain
ed
b
y
m
a
k
i
n
g
P
as
s
y
m
m
etr
ical
m
a
tr
ix
(
ad
d
in
g
m
atr
i
x
P
w
it
h
its
tr
a
n
s
p
o
s
e)
an
d
t
h
en
n
o
r
m
aliz
in
g
t
h
e
v
al
u
e
o
f
ea
c
h
ele
m
e
n
t
i
n
to
[
0
,
1
]
.
So
m
e
m
et
r
ics
ca
n
b
e
ca
lc
u
lated
b
ased
o
n
t
h
e
r
es
u
lti
n
g
G
L
C
M
m
a
tr
ix
,
th
e
y
ar
e
co
n
tr
as
t,
an
g
u
lar
s
ec
o
n
d
m
o
m
e
n
t
(
ASM
)
,
en
er
g
y
,
h
o
m
o
g
e
n
it
y
,
co
r
r
elatio
n
,
an
d
d
is
s
i
m
ilar
it
y
.
T
h
e
d
etail
f
o
r
m
u
la
f
o
r
ea
ch
m
etr
ic
ca
n
b
e
r
e
f
e
r
r
ed
at
[
3
0
]
.
I
n
t
h
is
r
e
s
ea
r
ch
,
w
e
co
n
s
tr
u
ct
G
L
C
M
m
atr
i
x
i
n
v
ar
io
u
s
s
p
atial
o
r
ien
tatio
n
(0
0
,
4
5
0
,
9
0
0
,
an
d
1
3
5
0
).
2
.
3
.
3
.
Sh
a
pe
Sh
ap
e
is
also
p
r
o
m
in
et
f
ea
tu
r
e
to
d
is
cir
m
in
ate
an
i
m
ag
e
t
o
an
o
th
er
.
T
h
is
r
esear
ch
ex
tr
ac
t
s
s
h
ap
e
d
escr
ip
to
r
s
o
f
an
im
a
g
e
f
r
o
m
m
o
r
p
h
o
lo
g
ical
f
ea
t
u
r
es a
n
d
Hu
I
n
v
ar
ia
n
t M
o
m
e
n
t.
So
m
e
m
o
r
p
o
h
lo
g
ical
f
ea
tu
r
es
ex
tr
ac
ted
ar
e
ar
ea
,
p
er
im
ete
r
,
m
aj
o
r
an
d
m
i
n
o
r
ax
i
s
,
c
en
tr
o
id
-
x
,
ce
n
tr
o
id
-
y
,
r
o
u
n
d
n
ess
,
r
ec
tan
g
u
lar
it
y
,
ec
ce
n
tr
icit
y
,
elo
n
g
at
io
n
,
d
is
p
er
s
io
n
I
,
d
is
p
er
s
io
n
I
R
,
co
n
v
ex
it
y
,
a
n
d
s
o
lid
it
y
[
1
9
]
.
T
h
e
illu
s
tr
atio
n
o
f
s
u
c
h
m
o
r
p
h
o
lo
g
ical
s
h
ap
e
f
ea
t
u
r
es
ca
n
b
e
s
ee
n
i
n
Fi
g
u
r
e
4
.
I
n
ad
d
itio
n
,
H
u
m
o
m
e
n
t
is
al
s
o
p
er
f
o
r
m
ed
to
e
x
tr
ac
t
s
h
ap
e
d
escr
ip
to
r
o
f
an
i
m
ag
e.
Hu
m
o
m
e
n
t
is
r
eg
io
n
-
b
ased
m
et
h
o
d
th
at
u
s
e
s
s
ec
o
n
d
an
d
th
ir
d
o
r
d
er
ce
n
tr
al
m
o
m
e
n
ts
a
n
d
co
n
s
tr
u
c
ts
7
in
v
ar
ia
n
t
m
o
m
e
n
ts
.
T
h
e
v
alu
e
o
f
i
n
v
ar
ian
t
m
o
m
en
t
f
ea
t
u
r
es
ar
e
n
o
t
af
f
ec
ted
w
h
e
n
th
e
i
m
a
g
e
is
tr
a
n
s
la
ted
,
r
o
tated
,
o
r
s
ca
led
.
T
h
e
d
etail
o
h
Hu
m
o
m
e
n
t c
an
b
e
r
ef
er
r
ed
in
[
2
0
]
.
ar
ea
p
er
im
e
ter
m
aj
o
r
a
nd
m
i
n
o
r
ax
is
ce
n
tr
o
id
(
x
,
y
)
r
o
u
n
d
n
e
s
s
r
ec
tan
g
u
lar
it
y
ec
ce
n
tr
icit
y
elo
n
g
atio
n
d
is
p
er
s
io
n
I
(
A
lter
n
ati
v
e
1)
d
is
p
er
s
io
n
IR
(
A
lter
n
ati
v
e
2
)
co
n
v
e
x
it
y
s
o
lid
it
y
Fig
u
r
e
4
.
T
h
e
illu
s
tr
atio
n
o
f
s
h
ap
e
d
escr
ip
to
r
s
f
r
o
m
m
o
r
p
h
o
l
o
g
ical
f
ea
t
u
r
es
2
.
4
.
Cla
s
s
if
ica
t
io
n
C
las
s
i
f
icatio
n
co
n
s
i
s
t
s
o
f
tr
ain
in
g
an
d
test
i
n
g
.
T
r
ain
in
g
is
u
s
ed
to
b
u
ild
th
e
class
i
f
ier
m
o
d
el,
w
h
ile
test
i
n
g
is
u
s
ed
to
ev
al
u
te
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
m
o
d
el
as
ill
u
s
tr
ated
i
n
Fig
u
r
e
5
.
Fir
s
t
t
h
e
d
ataset
is
s
p
litt
ed
b
y
u
s
i
n
g
1
0
-
f
o
ld
c
r
o
s
s
v
a
lid
atio
n
.
T
h
is
r
esear
ch
ap
p
lies
SVM
as
class
i
f
icatio
n
alg
o
r
it
h
m
.
D
a
t
a
S
p
l
i
t
t
i
n
g
T
r
a
i
n
i
n
g
a
l
g
o
r
i
t
h
m
T
e
s
t
i
n
g
a
l
g
o
r
i
t
h
m
E
v
a
l
u
a
t
i
o
n
T
r
a
i
n
i
n
g
d
a
t
a
T
e
s
t
i
n
g
d
a
t
a
T
r
a
i
n
e
d
m
o
d
e
l
R
e
s
u
l
t
s
o
f
p
r
e
d
i
c
t
i
o
n
&
e
v
a
l
u
a
t
i
o
n
Fig
u
r
e
5
.
T
r
an
in
g
an
d
test
i
n
g
i
n
clas
s
if
icatio
n
SVM
tr
ai
n
in
g
al
g
o
r
ith
m
w
o
r
k
s
b
y
f
in
d
i
n
g
th
e
o
p
ti
m
al
h
y
p
er
p
la
n
e
t
h
at
m
ax
i
m
ize
th
e
s
ep
ar
atio
n
b
et
w
ee
n
b
in
ar
y
clas
s
d
ata.
T
h
e
clo
s
est tr
ai
n
in
g
d
ata
to
t
h
e
o
p
ti
m
al
h
y
p
er
p
lan
e
t
h
at
d
e
f
in
ed
th
e
o
p
ti
m
al
m
ar
g
i
n
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
S
o
lid
w
a
s
te
cla
s
s
ifica
tio
n
u
s
in
g
p
yra
mid
s
ce
n
e
p
a
r
s
in
g
n
etw
o
r
k
s
eg
men
ta
tio
n
a
n
d
…
(
K
h
a
d
ija
h
)
1907
ar
e
ca
lled
s
u
p
p
o
r
t
v
ec
to
r
s
[
2
1
]
.
W
h
en
th
e
d
ata
ar
e
n
o
n
-
li
n
ea
r
l
y
s
ep
ar
ab
le,
n
o
n
-
li
n
ea
r
m
ap
p
in
g
(
)
is
ap
p
lied
to
tr
an
s
f
o
r
m
th
e
o
r
i
g
i
n
al
d
ata
in
to
h
ig
h
er
d
i
m
en
s
io
n
[
3
1
]
.
L
et
th
e
(
,
)
=
1
w
h
er
e
∈
is
i
n
p
u
t
tr
ain
i
n
g
d
ata,
is
tar
g
eted
d
ata
a
n
d
is
t
h
e
n
u
m
b
er
o
f
tr
ain
i
n
g
d
ata,
S
VM
f
i
n
d
t
h
e
s
o
lu
tio
n
b
y
s
o
l
v
i
n
g
th
e
f
o
llo
w
i
n
g
o
p
tim
izatio
n
p
r
o
b
lem
as
s
h
o
w
in
(
5
)
w
h
er
e
is
w
ei
g
h
t
v
e
cto
r
an
d
is
er
r
o
r
p
en
alt
y
.
Su
ch
o
p
ti
m
izatio
n
p
r
o
b
lem
ca
n
b
e
s
o
l
v
ed
u
s
in
g
L
ag
r
an
g
ia
n
f
o
r
m
u
lat
io
n
.
T
h
e
tr
ain
i
n
g
d
ata
is
n
o
r
m
a
lized
in
to
[
0
,
1
]
b
ef
o
r
e
th
e
y
ar
e
in
p
u
tted
to
th
e
SVM
a
n
d
is
s
et
i
n
to
-
1
o
r
1
[
3
2
]
.
min
,
,
1
2
+
∑
=
1
(
5
)
s
u
b
j
ec
t to
(
(
)
+
)
≥
1
−
,
≥
0
,
=
1
,
…
,
I
n
o
r
d
er
to
r
ed
u
ce
th
e
co
m
p
u
t
at
io
n
al
co
s
t
w
h
e
n
w
o
r
k
i
n
g
w
ith
n
o
n
li
n
ea
r
d
ata,
k
er
n
el
tr
ick
s
ca
n
b
e
u
s
ed
to
s
u
b
s
tit
u
t
e
t
h
e
d
o
t
p
r
o
d
u
ct
b
et
w
ee
n
tr
an
s
f
o
r
m
d
ata
tu
p
le
s
as
(
6
)
.
So
m
e
p
o
p
u
lar
k
er
n
el
f
u
n
ct
io
n
ca
n
b
e
u
s
ed
,
s
u
c
h
as p
o
l
y
n
o
m
ial
a
n
d
r
ad
ial
b
asis
f
u
n
ctio
n
(
R
B
F)
as
s
h
o
wn
in
(
7
)
an
d
(
8
)
,
r
esp
ec
tiv
el
y
[
3
3
]
.
(
,
)
=
(
)
.
(
)
(
6
)
(
,
)
=
e
xp
(
−
|
|
−
|
|
2
)
(
7
)
(
,
)
=
(
(
)
+
)
(
8
)
On
ce
t
h
e
o
p
ti
m
izatio
n
p
r
o
b
lem
s
o
l
v
ed
,
th
e
o
p
ti
m
al
h
y
p
er
p
l
an
e
an
d
t
h
e
s
u
p
p
o
r
t
v
ec
to
r
s
ar
e
o
b
tain
ed
.
T
h
en
,
th
e
o
u
tp
u
t
(
)
o
f
a
n
e
w
te
s
t
s
a
m
p
le
ca
n
b
e
d
eter
m
i
n
ed
b
y
u
s
i
n
g
(
9
)
w
h
er
e
ar
e
s
u
p
p
o
r
t v
ec
to
r
,
is
clas
s
lab
el
o
f
-
t
h
s
u
p
p
o
r
t
v
e
cto
r
,
is
t
h
e
n
u
m
b
er
o
f
s
u
p
p
o
r
t
v
er
cto
r
s
,
is
L
a
g
r
an
g
e
m
u
ltip
l
ier
s
,
an
d
is
b
ias
[
3
2
]
.
T
h
is
r
esear
ch
ap
p
li
es
th
e
o
n
e
-
v
er
s
u
s
-
r
e
s
t
s
tr
ateg
y
to
h
an
d
le
t
h
e
m
u
lticla
s
s
c
l
a
s
s
i
f
icatio
n
p
r
o
b
lem
,
b
ec
au
s
e
T
r
ash
n
et
d
ataset
co
n
s
is
t o
f
6
class
e
s
.
(
)
=
s
gn
(
∑
(
,
)
=
1
+
)
(
9
)
2
.
5
.
E
v
a
lua
t
i
o
n
E
v
alu
a
tio
n
i
s
p
er
f
o
r
m
ed
to
e
v
alu
ate
t
h
e
r
es
u
lti
n
g
cla
s
s
i
f
icati
o
n
m
o
d
el.
I
n
th
is
r
esear
c
h
,
e
v
alu
atio
n
o
f
class
i
f
icatio
n
m
o
d
el
is
m
ea
s
u
r
ed
in
ter
m
o
f
ac
c
u
r
ac
y
.
A
cc
u
r
a
c
y
s
h
o
w
s
t
h
e
r
atio
b
et
w
ee
n
th
e
co
r
r
ec
tly
p
r
ed
icted
d
ata
an
d
th
e
to
tal
n
u
m
b
er
o
f
d
ata
[
3
4
]
.
3.
RE
SU
L
T
S
A
ND
AN
AL
Y
SI
S
T
h
is
r
esear
ch
is
p
er
f
o
r
m
ed
in
t
w
o
m
ain
s
ce
n
ar
io
.
T
h
e
f
ir
s
t
s
ce
n
ar
io
is
p
er
f
o
r
m
ed
b
y
u
s
in
g
P
SP
Net
s
eg
m
e
n
tatio
n
,
w
h
ile
t
h
e
s
ec
o
n
d
s
ce
n
ar
io
s
k
ip
th
e
p
r
o
ce
s
s
o
f
s
e
g
m
e
n
tatio
n
.
I
n
ea
ch
s
ce
n
ar
io
,
s
i
n
g
le
o
r
co
m
b
i
n
atio
n
f
ea
t
u
r
e
ex
tr
ac
tio
n
o
f
co
lo
r
(
co
lo
r
m
o
m
en
ts
)
,
tex
tu
r
e
(
G
L
C
M)
a
n
d
s
h
ap
e
(
m
o
r
p
h
o
lo
g
ical
f
ea
t
u
r
es
an
d
Hu
I
n
v
ar
ian
t M
o
m
e
n
ts
)
ar
e
ex
p
er
i
m
en
ted
to
s
ea
r
ch
f
o
r
th
e
b
est i
m
a
g
e
f
ea
tu
r
e
s
th
at
w
e
ll d
escr
ib
e
th
e
tr
as
h
i
m
a
g
e
i
n
o
r
d
er
to
r
ea
ch
th
e
b
es
t
class
i
f
icat
io
n
r
es
u
lt
s
.
G
L
C
M
f
ea
t
u
r
e
ex
tar
ctio
n
m
eth
o
d
is
p
er
f
o
r
m
ed
in
v
ar
io
u
s
s
p
atial
o
r
ien
tatio
n
:
0
0
,
4
5
0
,
9
0
0
,
an
d
1
3
5
0
.
T
h
en
in
t
h
e
cla
s
s
i
f
ic
atio
n
,
SVM
tr
ain
in
g
al
g
o
r
ith
m
is
p
er
f
o
r
m
ed
u
s
in
g
s
o
m
e
co
m
b
i
n
atio
n
o
f
p
ar
a
m
ete
r
s
,
n
a
m
el
y
k
er
n
el
f
u
n
c
tio
n
(
R
B
F
an
d
p
o
l
y
n
o
m
ial)
an
d
er
r
o
r
p
en
alt
y
(
1
o
r
1
0
0
)
.
T
h
er
ef
o
r
e,
f
o
r
ea
ch
f
ea
tu
r
e
ex
tr
ac
tio
n
in
a
s
ce
n
ar
io
,
class
i
f
i
ca
tio
n
w
i
th
SVM
is
p
er
f
o
r
m
e
d
f
o
u
r
ti
m
es
u
s
in
g
d
if
f
er
e
n
t c
o
m
b
i
n
atio
n
o
f
k
er
n
e
l f
u
n
ctio
n
an
d
er
r
o
r
p
en
alt
y
.
I
n
th
e
la
s
t sect
io
n
t
h
e
r
esu
l
ts
o
f
th
e
f
ir
s
t sce
n
ar
io
an
d
th
e
s
ec
o
n
d
s
ce
n
ar
io
ar
e
co
m
p
ar
ed
.
3.
1
.
T
he
f
irst
s
ce
na
rio
I
n
th
is
s
ce
n
ar
io
s
e
g
m
e
n
tatio
n
is
p
er
f
o
r
m
ed
i
n
t
h
e
f
ir
s
t
s
ta
g
e
b
y
u
s
in
g
P
SP
Net.
I
n
o
r
d
er
to
o
b
tain
th
e
b
est m
o
d
el
o
f
P
SP
Net,
th
is
r
eseac
h
tr
y
s
o
m
e
co
m
b
i
n
atio
n
o
f
h
y
p
er
p
ar
a
m
eter
:
lear
n
i
n
g
r
ate
(
0
.
0
0
1
,
0
.
0
0
0
1
,
an
d
0
.
0
0
0
0
1
)
an
d
b
atch
(
5
an
d
1
0
)
.
A
n
ex
p
er
i
m
e
n
t
f
o
r
ea
ch
co
m
b
in
at
io
n
o
f
h
y
p
er
p
ar
a
m
eter
is
p
er
f
o
r
m
ed
i
n
5
0
ep
o
ch
.
B
ased
o
n
Fig
u
r
e
6
(
a)
,
it
is
s
h
o
w
n
th
at
t
h
e
lear
n
i
n
g
r
ate
o
f
0
.
0
0
0
1
g
iv
es
th
e
b
est
r
esu
lts
th
an
o
t
h
er
v
alu
e
s
.
I
t
ca
n
b
e
ex
p
lain
ed
th
at
w
h
en
t
h
e
lear
n
i
n
g
r
ate
is
to
o
s
m
all,
th
e
p
r
o
g
r
ess
o
f
n
e
t
wo
r
k
lear
n
in
g
is
v
er
y
s
lo
w
,
t
h
e
n
t
h
e
r
esu
lt
is
lo
w
er
.
C
o
n
v
er
s
el
y
,
w
h
e
n
t
h
e
lear
n
in
g
r
ate
is
to
o
h
i
g
h
t
h
e
p
r
o
g
r
ess
o
f
n
e
t
w
o
r
k
lear
n
in
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
19
,
No
.
6
,
Dec
em
b
er
2
0
2
1
:
1
9
0
2
-
1912
1908
m
a
y
d
i
v
er
g
e,
t
h
e
n
th
e
n
et
w
o
r
k
is
f
ailed
to
ac
h
ie
v
e
t
h
e
b
est
r
esu
lt.
Fi
g
u
r
e
6
(
b
)
s
h
o
w
s
th
a
t
th
e
b
atch
v
al
u
e
o
f
5
is
ab
le
to
r
ea
ch
b
etter
p
er
f
o
r
m
an
ce
th
a
n
t
h
e
b
atch
v
al
u
e
o
f
1
0
.
I
t
ca
n
b
e
ex
p
lain
ed
t
h
at
i
n
t
h
i
s
ca
s
e
t
h
e
s
to
c
h
asti
c
n
atu
r
e
o
f
u
s
in
g
lo
w
er
n
u
m
b
er
o
f
m
i
n
i
b
atc
h
m
a
y
lead
to
f
i
n
d
th
e
o
p
ti
m
u
m
s
o
l
u
tio
n
.
T
h
er
ef
o
r
e,
th
e
s
eg
m
e
n
tatio
n
in
th
e
r
est o
f
ex
p
er
i
m
e
n
t a
r
e
p
er
f
o
r
m
ed
b
y
u
s
i
n
g
th
e
b
est
s
eg
m
en
tatio
n
m
o
d
el
tr
ain
ed
b
y
th
o
s
e
co
m
b
i
n
atio
n
o
f
p
ar
am
eter
.
T
h
e
r
esu
lt
s
o
f
s
e
g
m
en
tatio
n
u
s
i
n
g
P
SP
Net
f
o
r
s
am
p
le
i
m
a
g
e
s
in
F
ig
u
r
e
2
ca
n
b
e
s
ee
n
i
n
Fi
g
u
r
e
7
.
(
a)
(
b
)
Fig
u
r
e
6
.
E
x
p
er
i
m
en
t
r
esu
l
ts
o
f
P
SP
Net
s
eg
m
en
tatio
n
o
n
te
s
t
in
g
d
ata
s
et:
(
a
)
t
h
e
a
v
e
r
a
g
e
r
e
s
u
l
t
o
f
a
l
l
v
a
r
i
a
t
i
o
n
o
f
b
a
t
c
h
v
a
l
u
e
s
i
n
e
a
c
h
l
e
a
r
n
i
n
g
r
a
t
e
a
n
d
(
b
)
t
h
e
a
v
e
r
a
g
e
r
e
s
u
l
t
s
o
f
a
l
l
v
a
r
i
a
t
i
o
n
o
f
l
e
a
r
n
i
n
g
r
a
t
e
v
a
l
u
e
s
i
n
e
a
c
h
b
a
t
c
h
g
las
s
p
ap
er
ca
r
d
b
o
a
r
d
p
last
ic
m
etal
tr
ash
Fig
u
r
e
7
.
T
h
e
s
am
p
le
r
es
u
lt
s
o
f
s
e
g
m
e
n
tat
io
n
u
s
i
n
g
P
SP
Net
T
h
e
r
esu
lt
o
f
ex
p
er
i
m
en
t
i
n
th
is
s
ce
n
ar
io
f
o
r
v
ar
io
u
s
f
ea
t
u
r
es
w
it
h
th
e
b
est
ac
cu
r
ac
y
i
n
ea
ch
SVM
k
er
n
el
(
R
B
F
a
n
d
p
o
ly
n
o
m
ial)
ca
n
b
e
s
ee
n
in
Fig
u
r
e
8
.
I
t
is
s
h
o
w
n
th
at
R
B
F
k
er
n
e
l
is
b
ett
er
th
an
p
o
l
y
n
o
m
ial
k
er
n
el
in
m
o
s
t
o
f
ex
p
er
i
m
e
n
t
s
,
b
u
t
w
h
e
n
m
o
r
e
co
m
b
in
atio
n
o
f
f
ea
tu
r
e
s
ar
e
u
s
e
d
,
th
e
p
o
l
y
n
o
m
ial
k
er
n
el
ar
e
b
etter
th
an
R
B
F
k
er
n
e
l.
T
h
e
h
ig
h
e
s
t
ac
cu
r
ac
y
o
f
7
6
.
4
9
%
in
th
is
s
ce
n
ar
io
is
ac
h
ie
v
ed
b
y
p
o
ly
n
o
m
ial
k
er
n
el
w
it
h
=
1
w
h
e
n
u
s
i
n
g
co
m
b
i
n
atio
n
f
e
atu
r
es
o
f
co
lo
r
,
GL
C
M
1
3
5
a
n
d
s
h
ap
e.
W
h
ile
t
h
e
h
i
g
h
est
a
cc
u
r
ac
y
o
f
R
B
F
k
er
n
el
i
n
th
i
s
s
ce
n
ar
io
is
7
4
.
5
5
%
w
h
e
n
u
s
i
n
g
=
100
an
d
th
e
s
a
m
e
c
o
m
b
i
n
atio
n
f
ea
t
u
r
es
o
f
co
lo
r
,
GL
C
M
1
3
5
an
d
s
h
ap
e.
T
h
er
ef
o
r
e,
it
ca
n
b
e
co
n
clu
d
ed
th
at
w
h
en
s
eg
m
e
n
tatio
n
i
s
u
s
ed
,
th
e
p
er
f
o
r
m
a
n
ce
o
f
class
i
f
icatio
n
in
cr
ea
s
e
as t
h
e
m
o
r
e
co
m
b
in
a
t
io
n
o
f
f
ea
t
u
r
es a
r
e
u
s
ed
.
T
h
e
u
s
e
o
f
m
o
r
e
co
m
b
in
a
tio
n
o
f
f
ea
tu
r
es g
iv
e
t
h
e
m
o
r
e
r
ep
r
esen
tativ
e
f
ea
t
u
r
e
s
ets
o
f
an
i
m
a
g
e,
th
er
e
f
o
r
e
th
e
ac
cu
r
ac
y
o
f
clas
s
if
icatio
n
i
n
cr
ea
s
e.
Ho
w
ev
er
,
t
h
e
m
o
s
t
i
m
p
o
r
tan
t
f
ea
tu
r
e
i
s
th
e
co
lo
r
f
ea
tu
r
e.
W
h
en
t
h
e
co
lo
r
f
ea
tu
r
e
is
r
e
m
o
v
ed
,
th
e
ac
c
u
r
ac
y
o
f
cl
ass
i
f
ier
d
ec
r
ea
s
e.
3
.
2
.
T
he
s
ec
o
nd
s
ce
na
rio
T
h
e
s
ec
o
n
d
s
ce
n
ar
io
is
p
er
f
o
r
m
ed
w
it
h
o
u
t
s
e
g
m
en
tatio
n
in
t
h
e
p
r
ep
r
o
ce
s
s
in
g
.
T
h
e
r
es
u
lt
o
f
ex
p
er
i
m
e
n
t
in
t
h
is
s
ce
n
ar
io
f
o
r
v
ar
io
u
s
f
ea
tu
r
es
w
i
th
th
e
b
e
s
t
ac
c
u
r
ac
y
i
n
ea
ch
S
VM
k
er
n
el
(
R
B
F
an
d
p
o
ly
n
o
m
ial)
ca
n
b
e
s
ee
n
i
n
Fig
u
r
e
9
.
I
t
is
s
h
o
w
n
th
at
R
B
F
k
er
n
el
is
b
etter
t
h
a
n
p
o
l
y
n
o
m
ial
k
er
n
el
i
n
m
o
s
t
o
f
ex
p
er
i
m
en
ts
.
H
o
w
e
v
er
,
th
e
h
ig
h
es
t
ac
cu
r
ac
y
o
f
7
4
.
8
3
%
in
th
i
s
s
ce
n
ar
io
is
ac
h
ie
v
ed
b
y
p
o
l
y
n
o
m
ial
k
er
n
el
w
it
h
=
100
w
h
e
n
u
s
i
n
g
co
m
b
i
n
atio
n
f
ea
t
u
r
es
o
f
co
lo
r
an
d
GL
C
M
9
0
.
W
h
ile
th
e
b
est
r
esu
lt
o
f
R
B
F
k
er
n
el
i
n
th
is
s
ce
n
ar
io
is
7
4
.
5
5
%
w
h
e
n
u
s
i
n
g
=
100
an
d
co
m
b
in
atio
n
f
ea
t
u
r
es
o
f
co
lo
r
an
d
GL
C
M
1
3
5
.
T
h
er
ef
o
r
e,
it
ca
n
b
e
co
n
c
lu
d
ed
th
a
t
w
h
e
n
s
eg
m
e
n
tatio
n
is
n
o
t
u
s
ed
,
th
e
b
est
co
m
b
in
at
io
n
o
f
f
ea
t
u
r
e
th
at
w
e
ll
d
escr
ib
e
th
e
tr
ash
i
m
ag
e
i
s
co
m
b
i
n
atio
n
o
f
co
lo
r
an
d
G
L
C
M.
W
h
e
n
,
t
h
e
s
h
ap
e
f
ea
t
u
r
es
ar
e
ad
d
ed
,
th
e
p
er
f
o
r
m
a
n
ce
o
f
cla
s
s
i
f
icatio
n
d
ec
r
ea
s
e.
T
o
ex
tr
ac
t
s
h
ap
e
f
ea
tu
r
es
i
n
t
h
is
s
ce
n
ar
io
,
a
co
n
v
e
n
tio
n
al
th
r
e
s
h
o
ld
i
n
g
o
p
e
r
atio
n
is
ap
p
lied
to
tr
an
s
f
o
r
m
a
R
GB
i
m
a
g
e
i
n
to
b
in
ar
y
i
m
a
g
e,
th
e
f
o
r
e
th
e
r
esu
lt
i
n
g
b
i
n
ar
y
i
m
ag
e
i
s
n
o
t
g
o
o
d
en
o
u
g
h
as so
u
r
ce
f
o
r
ex
tr
ac
ti
n
g
s
h
ap
e
f
ea
tu
r
es.
3
.
3
.
Co
m
pa
ri
s
o
n o
f
t
he
f
irst
s
ce
na
rio
a
nd
t
he
s
ec
o
nd
s
ce
n
a
rio
Fig
u
r
e
1
0
s
h
o
w
s
t
h
e
co
m
p
ar
is
o
n
b
et
w
ee
n
th
e
f
ir
s
t
s
ce
n
ar
io
an
d
th
e
s
ec
o
n
d
s
ce
n
ar
io
.
B
ased
o
n
Fig
u
r
e
1
0
,
it
is
s
h
o
w
n
th
at
th
e
f
ir
s
t
s
ce
n
ar
io
(
u
s
i
n
g
P
SP
Net
s
eg
m
e
n
tatio
n
)
i
s
b
etter
th
a
n
t
h
e
s
ec
o
n
d
s
ce
n
ar
io
(
w
it
h
o
u
t
s
eg
m
e
n
tatio
n
)
in
m
o
s
t
o
f
ex
p
er
i
m
en
t.
T
h
e
s
ec
o
n
d
s
ce
n
ar
io
o
u
tp
er
f
o
r
m
s
th
e
f
ir
s
t
s
ce
n
ar
io
o
n
l
y
i
n
5
f
r
o
m
1
9
ex
p
er
im
e
n
t
s
.
T
h
er
ef
o
r
e,
it
ca
n
b
e
co
n
clu
d
ed
th
at
g
en
er
all
y
ap
p
l
y
i
n
g
P
SP
Net
s
eg
m
en
tatio
n
p
r
o
v
id
e
b
etter
s
o
u
r
ce
f
o
r
f
ea
tu
r
e
ex
tr
ac
tio
n
,
esp
ec
iall
y
in
co
lo
r
an
d
s
h
ap
e
f
ea
t
u
r
e,
h
en
ce
in
cr
ea
s
e
th
e
p
er
f
o
r
m
an
ce
o
f
class
i
f
icatio
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
S
o
lid
w
a
s
te
cla
s
s
ifica
tio
n
u
s
in
g
p
yra
mid
s
ce
n
e
p
a
r
s
in
g
n
etw
o
r
k
s
eg
men
ta
tio
n
a
n
d
…
(
K
h
a
d
ija
h
)
1909
Fig
u
r
e
8
.
R
esu
lt o
f
e
x
p
er
i
m
e
n
t
s
in
t
h
e
f
ir
s
t sce
n
ar
io
Fig
u
r
e
9
.
R
esu
lt o
f
e
x
p
er
i
m
e
n
t
s
in
t
h
e
s
ec
o
n
d
s
ce
n
ar
io
Fig
u
r
e
1
0
.
C
o
m
p
ar
is
o
n
b
et
w
e
en
th
e
f
ir
s
t
s
ce
n
ar
io
(
u
s
in
g
s
e
g
m
en
tatio
n
)
a
n
d
th
e
s
ec
o
n
d
s
ce
n
ar
io
(
w
it
h
o
u
t
s
eg
m
e
n
tatio
n
)
I
t is also
o
b
s
er
v
ed
th
at
t
h
e
m
o
s
t i
m
p
o
r
tan
t f
ea
t
u
r
e
in
th
i
s
p
r
o
b
le
m
is
co
lo
r
f
ea
t
u
r
e.
W
h
en
u
s
in
g
s
i
n
g
l
e
f
ea
t
u
r
e,
co
lo
r
f
ea
tu
r
e
p
r
o
v
id
e
th
e
h
i
g
h
est
r
es
u
lt
co
m
p
ar
ed
to
GL
C
M
(
tex
tu
r
e)
an
d
s
h
ap
e
f
ea
t
u
r
e,
b
o
th
in
th
e
f
ir
s
t a
n
d
th
e
s
ec
o
n
d
s
ce
n
ar
io
.
Ho
w
e
v
er
,
th
e
ac
cu
r
ac
y
in
cr
ea
s
e
if
ad
d
itio
n
al
f
ea
tu
r
es a
r
e
in
t
r
o
d
u
ce
d
.
I
n
th
e
f
ir
s
t
s
ce
n
ar
io
b
etter
r
esu
lts
ar
e
ac
h
i
ev
ed
w
h
en
u
s
i
n
g
all
co
m
b
in
ati
o
n
o
f
f
ea
t
u
r
es,
w
h
ile
i
n
th
e
s
ec
o
n
d
s
ce
n
ar
io
b
etter
r
esu
lt
s
ar
e
ac
h
iev
ed
w
h
en
u
s
i
n
g
o
n
l
y
co
lo
r
an
d
tex
t
u
r
e
f
ea
tu
r
es.
T
h
er
ef
o
r
e,
it
ca
n
b
e
co
n
clu
d
ed
th
a
t
w
h
e
n
s
eg
m
e
n
tatio
n
is
ap
p
lied
b
y
u
s
i
n
g
P
SP
Net,
th
e
s
e
g
m
e
n
ted
b
in
ar
y
i
m
ag
e
p
r
o
v
id
e
b
etter
s
o
u
r
c
e
f
o
r
s
h
ap
e
f
ea
t
u
r
e
ex
tr
ac
tio
n
.
C
o
n
v
er
s
el
y
,
w
h
e
n
th
e
b
in
ar
y
i
m
a
g
e
is
o
n
l
y
o
b
tain
ed
b
y
u
s
i
n
g
i
n
v
er
s
e
b
in
ar
y
t
h
r
e
s
h
o
ld
in
g
,
th
e
r
esu
lt
is
n
o
t
g
o
o
d
en
o
u
g
h
f
o
r
s
h
ap
e
f
ea
tu
r
e
e
x
tr
ac
tio
n
.
He
n
ce
,
th
e
ac
cu
r
ac
y
o
f
cla
s
s
i
f
icatio
n
d
ec
r
ea
s
e
w
h
e
n
s
h
ap
e
f
ea
t
u
r
e
is
ad
d
ed
in
t
h
e
s
ec
o
n
d
s
ce
n
ar
io
.
Fro
m
all
co
m
b
in
a
ti
o
n
o
f
p
ar
a
m
e
ter
s
co
n
d
u
cted
i
n
t
h
is
r
esear
ch
,
t
h
e
h
ig
h
e
s
t
ac
cu
r
ac
y
o
f
7
6
.
4
9
%
is
ac
h
ie
v
ed
w
h
en
u
s
i
n
g
P
SP
N
et
s
eg
m
en
tatio
n
an
d
all
co
m
b
in
atio
n
o
f
f
ea
t
u
r
es
(
co
lo
r
,
tex
tu
r
e,
an
d
s
h
ap
e)
.
T
h
e
r
esu
lts
o
f
th
is
r
esear
ch
s
h
o
w
th
a
t
th
e
co
m
b
in
atio
n
o
f
f
ea
t
u
r
es
ar
e
ab
le
to
in
cr
ea
s
e
th
e
p
e
r
f
o
r
m
an
ce
o
f
th
e
r
es
u
lti
n
g
m
o
d
el
th
a
n
w
h
e
n
u
s
in
g
t
h
e
in
d
i
v
id
u
al
f
e
atu
r
e,
b
u
t
th
e
y
ar
e
s
till
n
o
t
e
n
o
u
g
h
to
u
n
iq
u
el
y
ch
ar
ac
ter
ize
ea
ch
clas
s
o
f
s
o
li
d
w
a
s
te
i
m
a
g
e.
T
h
e
m
o
r
e
r
ep
r
esen
tat
iv
e
ad
d
itio
n
al
f
ea
t
u
r
es
ar
e
s
till
r
eq
u
ir
ed
to
i
m
p
r
o
v
e
th
e
p
er
f
o
r
m
a
n
ce
o
f
class
i
f
ier
.
T
h
e
tu
n
in
g
o
f
p
ar
am
eter
o
f
class
i
f
icatio
n
alg
o
r
it
h
m
al
s
o
n
ee
d
to
b
e
ex
p
lo
r
ed
to
o
b
tain
b
etter
class
i
f
icatio
n
r
esu
l
ts
.
4.
CO
NCLU
SI
O
N
I
n
th
is
r
esear
ch
w
e
ap
p
l
y
P
SP
Net
as
s
eg
m
en
ta
tio
n
an
d
co
m
b
in
atio
n
o
f
i
m
ag
e
f
ea
t
u
r
e
ex
tr
ac
tio
n
(
co
lo
r
,
tex
t
u
r
e
,
an
d
s
h
ap
e)
to
class
i
f
y
t
h
e
s
o
lid
w
aste
i
m
a
g
e.
As
a
co
m
p
ar
is
o
n
,
to
s
ee
th
e
ef
f
ec
t
o
f
P
SP
Net
s
eg
m
e
n
tatio
n
,
w
e
a
ls
o
p
er
f
o
r
m
ex
p
er
i
m
e
n
t
w
it
h
o
u
t
u
s
in
g
s
eg
m
e
n
tat
io
n
.
B
ased
o
n
t
h
e
r
es
u
lt
o
f
e
x
p
er
i
m
e
n
t,
i
t
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
19
,
No
.
6
,
Dec
em
b
er
2
0
2
1
:
1
9
0
2
-
1912
1910
ca
n
b
e
co
n
cl
u
d
ed
th
a
t
g
en
er
all
y
ap
p
l
y
i
n
g
s
e
g
m
en
tatio
n
p
r
o
v
id
e
b
etter
s
o
u
r
ce
f
o
r
f
ea
t
u
r
e
ex
t
r
ac
tio
n
,
esp
ec
iall
y
in
co
lo
r
an
d
s
h
ap
e
f
ea
tu
r
e,
h
en
ce
in
cr
ea
s
e
t
h
e
ac
cu
r
ac
y
o
f
class
i
f
icatio
n
.
I
t
is
also
o
b
s
er
v
ed
th
at
th
e
m
o
s
t
i
m
p
o
r
tan
t
f
ea
t
u
r
e
in
th
i
s
p
r
o
b
l
e
m
is
co
lo
r
f
ea
tu
r
e,
b
o
th
w
h
e
n
th
e
s
eg
m
en
ta
tio
n
is
ap
p
lied
o
r
n
o
t.
Ho
w
e
v
er
,
th
e
ac
cu
r
ac
y
o
f
clas
s
i
f
ier
in
cr
ea
s
e
if
ad
d
itio
n
al
f
ea
tu
r
e
s
ar
e
in
tr
o
d
u
ce
d
.
W
h
en
s
eg
m
e
n
atio
n
is
n
o
t
u
s
ed
,
b
etter
r
esu
lt
is
ac
h
ie
v
ed
w
h
en
u
s
i
n
g
o
n
l
y
co
lo
r
an
d
tex
tu
r
e
f
ea
tu
r
e
s
,
w
h
ile
w
h
e
n
s
e
g
m
en
ta
tio
n
i
s
ap
p
lied
th
e
h
i
g
h
e
s
t
ac
cu
r
ac
y
o
f
7
6
.
4
9
% is
ac
h
ie
v
e
d
w
h
e
n
u
s
i
n
g
al
l c
o
m
b
in
a
tio
n
o
f
f
ea
t
u
r
es.
ACK
NO
WL
E
D
G
E
M
E
NT
S
T
h
is
r
esear
ch
i
s
s
u
p
p
o
r
ted
b
y
t
h
e
Fac
u
lt
y
o
f
Scie
n
ce
a
n
d
Ma
t
h
e
m
a
tics
,
U
n
i
v
er
s
ita
s
Dip
o
n
e
g
o
r
o
u
n
d
er
th
e
Gr
an
t o
f
P
r
i
m
ar
y
R
e
s
ea
r
ch
w
it
h
t
h
e
co
n
tr
ac
t n
u
m
b
er
2
0
0
7
/UN7
.
5
.
8
/
P
P
/2
0
2
0
.
RE
F
E
R
E
NC
E
S
[1
]
S
.
Ka
z
a
,
L
.
Ya
o
,
P
.
B.
T
a
ta
,
a
n
d
F
.
V
.
W
o
e
rd
e
n
,
W
h
a
t
A
W
a
ste
2
.
0
A
Glo
b
a
l
S
n
a
p
sh
o
t
o
f
S
o
li
d
W
a
ste
M
a
n
a
g
e
me
n
t
to
2
0
5
0
,
W
a
sh
in
g
to
n
:
W
o
rl
d
Ba
n
k
P
u
b
li
c
a
ti
o
n
s
,
2
0
1
8
.
[2
]
F.
A
.
Ra
h
m
a
n
,
“
Re
d
u
c
e
,
Re
u
se
,
R
e
c
y
c
l
e
:
A
lt
e
rn
a
ti
v
e
s
f
o
r
W
a
ste
M
a
n
a
g
e
m
e
n
t,
”
NM,
U
n
it
e
d
S
tate
:
L
a
s
Cru
c
e
s,
NM
:
NM
S
tate
Un
iv
e
rsit
y
,
Co
o
p
e
ra
ti
v
e
Ex
ten
sio
n
S
e
rv
ice
,
2
0
1
4
,
p
p
.
1
-
4.
[
O
n
li
n
e
]
.
A
v
a
il
a
b
le:
h
tt
p
s:/
/ac
e
s.n
m
su
.
e
d
u
/p
u
b
s/_
g
/G
3
1
4
.
p
d
f
[3
]
M.
R.
M
u
sta
ff
a
et
a
l.,
“
A
u
to
m
a
ted
re
c
y
c
lab
le
w
a
ste
c
las
si
f
ica
t
io
n
u
si
n
g
m
u
lt
ip
le
sh
a
p
e
-
b
a
se
d
p
ro
p
e
rti
e
s
a
n
d
q
u
a
d
ra
ti
c
d
isc
r
im
in
a
n
t,
”
In
t
.
J
.
I
n
n
o
v
a
ti
v
e
T
e
c
h
n
o
l.
a
n
d
Exp
l
o
rin
g
E
n
g
.
(
IJ
IT
EE
),
v
o
l.
8
,
p
p
.
2
7
0
-
2
7
4
,
Ju
n
.
2
0
1
9
,
[
On
li
n
e
]
.
A
v
a
il
a
b
le:
h
tt
p
s:
//
ww
w
.
ij
it
e
e
.
o
rg
/w
p
-
c
o
n
ten
t/
u
p
l
o
a
d
s/
p
a
p
e
rs/v
8
i8
s/H1
0
4
5
0
6
8
8
S
1
9
.
p
d
f
[4
]
A
.
T
.
G
a
rc
ía
,
O.
R
.
A
ra
g
ó
n
,
O.
L
.
G
a
n
d
a
ra
,
F
.
S
.
G
a
rc
ía,
a
n
d
L
.
E.
G
.
Jim
é
n
e
z
,
“
In
telli
g
e
n
t
wa
ste
se
p
a
ra
to
r,
”
Co
mp
u
t
a
c
io
n
y
S
istem
a
s
,
v
o
l.
1
9
,
n
o
.
3
,
p
p
.
4
8
7
-
5
0
0
,
2
0
1
5
,
d
o
i:
1
0
.
1
3
0
5
3
/Cy
S
-
19
-
3
-
2
2
5
4
.
[5
]
O.
A
d
e
d
e
ji
a
n
d
Z.
W
a
n
g
,
“
In
te
ll
ig
e
n
t
w
a
ste
c
las
si
f
ica
ti
o
n
u
sin
g
d
e
e
p
lea
rn
i
n
g
c
o
n
v
o
lu
t
io
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
,
”
P
r
o
c
e
d
ia M
a
n
u
f
.
,
v
o
l.
3
5
,
p
p
.
6
0
7
-
6
1
2
,
2
0
1
9
,
d
o
i:
1
0
.
1
0
1
6
/
j.
p
r
o
m
fg
.
2
0
1
9
.
0
5
.
0
8
6
.
[6
]
B.
S
.
Co
sta
a
t
a
l.
,
“
A
rti
f
i
c
ial
in
t
e
ll
ig
e
n
c
e
in
a
u
t
o
m
a
t
e
d
so
rti
n
g
in
tras
h
re
c
y
c
li
n
g
,
”
in
Co
n
fer
e
n
c
e
:
XV
E
n
c
o
n
tro
Na
c
io
n
a
l
d
e
In
teli
g
ê
n
c
ia
Arti
fi
c
i
a
l
e
Co
mp
u
ta
c
i
o
n
a
l
,
2
0
1
8
,
p
p
.
1
9
8
-
2
0
5
.
[7
]
S
.
M
a
tt
a
,
“
Re
v
ie
w
:
v
a
rio
u
s
im
a
g
e
se
g
m
e
n
tatio
n
tec
h
n
i
q
u
e
s,”
I
n
t.
J
.
Co
mp
u
t.
S
c
i.
a
n
d
I
n
f.
T
e
c
h
n
o
l.
,
v
o
l.
5
,
n
o
.
6
,
p
p
.
7
5
3
6
-
7
5
3
9
,
2
0
1
4
.
[
On
li
n
e
]
.
A
v
a
il
a
b
le:
h
tt
p
:/
/cite
se
e
rx
.
ist.
p
su
.
e
d
u
/v
iew
d
o
c
/d
o
w
n
lo
a
d
;j
se
ss
io
n
id
=
8
2
1
0
7
5
0
4
F
0
2
9
CD9
7
E6
1
1
CD
0
2
6
F
C5
8
3
4
F
?
d
o
i=
1
0
.
1
.
1
.
6
5
7
.
6
6
6
1
&
re
p
=
re
p
1
&
ty
p
e
=
p
d
f
[8
]
D.
L
ib
o
u
g
a
,
L
.
Gw
e
t,
M
.
Ote
st
e
a
n
u
,
I.
O.
L
ib
o
u
g
a
,
L
.
Bit
j
o
k
a
,
a
n
d
G
.
D.
P
o
p
a
,
“
A
re
v
ie
w
o
n
im
a
g
e
se
g
m
e
n
tatio
n
tec
h
n
iq
u
e
s
a
n
d
p
e
rf
o
rm
a
n
c
e
m
e
a
su
re
s,”
In
t.
J
.
Co
mp
u
t.
a
n
d
In
f
.
En
g
.
,
v
o
l.
1
2
,
n
o
.
1
2
,
p
p
.
1
1
0
7
-
1
1
1
7
,
2
0
1
8
,
d
o
i:
1
0
.
5
2
8
1
/ze
n
o
d
o
.
2
5
7
9
9
7
6
.
[9
]
S
.
M
in
a
e
e
,
Y.
Y.
B
o
y
k
o
v
,
F
.
P
o
ri
k
li
,
A
.
J.
P
laz
a
,
N.
Ke
h
tarn
a
v
a
z
,
a
n
d
D.
T
e
rz
o
p
o
u
lo
s,
“
Im
a
g
e
se
g
m
e
n
tatio
n
u
si
n
g
d
e
e
p
lea
rn
in
g
:
a
su
rv
e
y
,
”
i
n
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
P
a
tt
e
rn
A
n
a
lys
is
a
n
d
M
a
c
h
i
n
e
In
telli
g
e
n
c
e
,
d
o
i:
1
0
.
1
1
0
9
/T
P
A
M
I.
2
0
2
1
.
3
0
5
9
9
6
8
.
[1
0
]
H.
Zh
a
o
,
J.
S
h
i,
X
.
Qi,
X
.
W
a
n
g
,
J
.
Jia
,
a
n
d
S
.
G
.
L
im
it
e
d
,
“
P
y
ra
m
id
sc
e
n
e
p
a
rsin
g
n
e
t
w
o
rk
,
”
2
0
1
7
IE
EE
Co
n
fer
e
n
c
e
o
n
C
o
mp
u
ter
Vi
sio
n
a
n
d
Pa
tt
e
rn
Rec
o
g
n
it
io
n
(
CVP
R)
,
2
0
1
7
,
p
p
.
6
2
3
0
-
6
2
3
9
,
d
o
i:
1
0
.
1
1
0
9
/C
VP
R.
2
0
1
7
.
6
6
0
.
[1
1
]
I.
G
u
y
o
n
,
M
.
Nik
ra
v
e
sh
,
S
.
G
u
n
n
,
a
n
d
L
.
A
.
Zad
e
h
,
Fea
t
u
re
E
x
t
ra
c
ti
o
n
Fo
u
n
d
a
t
io
n
s
a
n
d
A
p
p
li
c
a
ti
o
n
s
,
S
p
r
in
g
e
r,
2
0
0
6
,
d
o
i:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
5
4
0
-
3
5
4
8
8
-
8
.
[1
2
]
B.
Re
m
e
s
e
iro
a
n
d
V
.
Bo
l
o
n
-
c
a
n
e
d
o
,
“
A
re
v
ie
w
o
f
fe
a
tu
re
se
lec
ti
o
n
m
e
th
o
d
s
in
m
e
d
ica
l
a
p
p
li
c
a
ti
o
n
s,”
Co
mp
u
t.
in
Bi
o
l.
a
n
d
M
e
d
icin
e
,
v
o
l.
1
1
2
,
n
o
.
1
0
3
3
7
5
,
p
p
.
1
-
9
,
J
u
ly
2
0
1
9
,
d
o
i
:
1
0
.
1
0
1
6
/j
.
c
o
m
p
b
io
m
e
d
.
2
0
1
9
.
1
0
3
3
7
5
[1
3
]
G
.
Ku
m
a
r,
“
A
d
e
tailed
re
v
ie
w
o
f
f
e
a
tu
re
e
x
trac
ti
o
n
in
im
a
g
e
p
ro
c
e
ss
in
g
s
y
ste
m
s,”
in
2
0
1
4
F
o
u
rt
h
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
A
d
v
a
n
c
e
d
Co
mp
u
t
in
g
&
Co
mm
u
n
ica
ti
o
n
T
e
c
h
n
o
lo
g
ies
,
2
0
1
4
,
p
p
.
5
-
1
2
,
2
0
1
4
,
d
o
i:
1
0
.
1
1
0
9
/A
CCT
.
2
0
1
4
.
7
4
.
[1
4
]
D.
T
ian
,
“
A
re
v
ie
w
o
n
im
a
g
e
f
e
a
t
u
re
e
x
trac
ti
o
n
a
n
d
re
p
re
se
n
tatio
n
tec
h
n
iq
u
e
s,”
In
t
.
J
.
M
u
lt
ime
d
i
a
a
n
d
U
b
iq
u
o
it
o
u
s
En
g
.
,
v
o
l.
8
,
n
o
.
4
,
p
p
.
3
8
5
-
3
9
5
,
Ju
ly
2
0
1
3
.
[1
5
]
A
.
J.
Af
i
f
i
a
n
d
W
.
M
.
A
sh
o
u
r,
“
I
m
a
g
e
re
tri
e
v
a
l
b
a
se
d
o
n
c
o
n
te
n
t
u
si
n
g
c
o
lo
r
f
e
a
tu
re
,
”
In
t
e
rn
a
ti
o
n
a
l
S
c
h
o
la
rly
Res
e
a
rc
h
No
ti
c
e
s
,
v
o
l.
2
0
1
2
,
p
p
.
1
-
1
1
,
2
0
1
2
,
d
o
i:
1
0
.
5
4
0
2
/
2
0
1
2
/2
4
8
2
8
5
.
[1
6
]
Z.
L
a
n
a
n
d
Y.
L
iu
,
“
S
tu
d
y
o
n
m
u
lt
i
-
sc
a
le
w
in
d
o
w
d
e
ter
m
i
n
a
ti
o
n
f
o
r
GL
CM
tex
tu
re
d
e
sc
rip
ti
o
n
in
h
ig
h
-
re
so
lu
t
i
o
n
re
m
o
te
se
n
sin
g
im
a
g
e
g
e
o
-
a
n
a
l
y
sis
su
p
p
o
rte
d
b
y
G
IS
a
n
d
d
o
m
a
in
k
n
o
w
led
g
e
,
”
In
t
e
rn
a
ti
o
n
a
l
J
o
u
rn
a
l
Ge
o
-
In
f
o
rm
a
ti
o
n
,
v
o
l
.
7
,
n
o
.
5
,
p
p
.
1
-
2
4
,
2
0
1
8
,
d
o
i:
1
0
.
3
3
9
0
/i
jg
i7
0
5
0
1
7
5
.
[1
7
]
P
.
M
o
h
a
n
a
iah
,
P
.
S
a
th
y
a
n
a
ra
y
a
n
a
,
a
n
d
L
.
G
u
ru
k
u
m
a
r,
“
I
m
a
g
e
tex
tu
r
e
f
e
a
tu
re
e
x
tra
c
ti
o
n
u
sin
g
G
L
CM
a
p
p
ro
a
c
h
,
”
In
t
.
J
.
S
c
ien
ti
fi
c
a
n
d
Res
.
Pu
b
li
c
a
ti
o
n
s
,
v
o
l.
3
,
n
o
.
5
,
p
p
.
1
-
5
,
2
0
1
3
.
[
O
n
li
n
e
]
.
A
v
a
il
a
b
le:
h
tt
p
:/
/w
ww
.
ij
sr
p
.
o
rg
/res
e
a
rc
h
-
p
a
p
e
r
-
0
5
1
3
/i
jsrp
-
p
1
7
5
0
.
p
d
f
[1
8
]
A
.
Zo
ti
n
,
Y.
Ha
m
a
d
,
K.
S
im
o
n
o
v
,
a
n
d
M
.
K
u
ra
k
o
,
“
L
u
n
g
b
o
u
n
d
a
r
y
d
e
tec
ti
o
n
f
o
r
c
h
e
st
X
-
ra
y
i
m
a
g
e
s
c
las
si
f
ica
ti
o
n
b
a
se
d
o
n
G
L
CM
a
n
d
p
ro
b
a
b
il
isti
c
n
e
u
ra
l
n
e
tw
o
rk
s,”
Pro
c
e
d
ia
Co
mp
u
t
er
S
c
i
e
n
c
e
,
v
o
l.
1
5
9
,
p
p
.
1
4
3
9
-
1
4
4
8
,
2
0
1
9
,
d
o
i:
1
0
.
1
0
1
6
/j
.
p
ro
c
s.
2
0
1
9
.
0
9
.
3
1
4
.
[1
9
]
F
.
Y.
M
a
n
ik
,
Y.
He
rd
iy
e
n
i
,
a
n
d
E.
N.
He
rli
y
a
n
a
,
“
L
e
a
f
m
o
r
p
h
o
lo
g
ica
l
f
e
a
tu
re
e
x
trac
ti
o
n
o
f
d
ig
it
a
l
im
a
g
e
A
n
th
o
c
e
p
h
a
lu
s
Ca
d
a
m
b
a
,
”
T
EL
KOM
NIKA
T
e
lec
o
mm
u
n
ica
ti
o
n
Co
mp
u
ti
n
g
El
e
c
tro
n
ics
a
n
d
Co
n
tro
l
,
v
o
l.
1
4
,
n
o
.
2
,
p
p
.
6
3
0
-
6
3
7
,
2
0
1
6
,
d
o
i:
1
0
.
1
2
9
2
8
/T
EL
K
OMNIK
A
.
v
1
4
i
2
.
2
6
7
5
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
S
o
lid
w
a
s
te
cla
s
s
ifica
tio
n
u
s
in
g
p
yra
mid
s
ce
n
e
p
a
r
s
in
g
n
etw
o
r
k
s
eg
men
ta
tio
n
a
n
d
…
(
K
h
a
d
ija
h
)
1911
[2
0
]
Z.
W
u
,
S
.
Jia
n
g
,
X
.
Z
h
o
u
,
Y.
W
a
n
g
,
a
n
d
Y.
Z
u
o
,
“
A
p
p
li
c
a
ti
o
n
o
f
i
m
a
g
e
re
tri
e
v
a
l
b
a
se
d
o
n
c
o
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
s
a
n
d
Hu
in
v
a
rian
t
m
o
m
e
n
t
a
lg
o
rit
h
m
in
c
o
m
p
u
ter
tel
e
c
o
m
m
u
n
ica
ti
o
n
s,”
Co
mp
u
t.
C
o
mm
u
n
.
,
v
o
l.
1
5
0
,
n
o
.
De
c
e
m
b
e
r
2
0
1
9
,
p
p
.
7
2
9
-
7
3
8
,
2
0
2
0
,
d
o
i:
1
0
.
1
0
1
6
/j
.
c
o
m
c
o
m
.
2
0
1
9
.
1
1
.
0
5
3
.
[2
1
]
C.
Co
r
tes
a
n
d
V
.
N.
V
a
p
n
ik
,
“
S
u
p
p
o
rt
-
V
e
c
to
r
Ne
tw
o
rk
s,”
M
a
c
h
in
e
L
e
a
r
n
in
g
,
v
o
l.
2
0
,
n
o
.
3
,
p
p
.
2
7
3
-
2
9
7
,
1
9
9
5
,
d
o
i:
1
0
.
1
0
0
7
/BF
0
0
9
9
4
0
1
8
.
[2
2
]
M
.
R.
P
h
a
n
g
tri
a
stu
,
J.
Ha
re
f
a
,
a
n
d
D.
F
.
T
a
n
o
t
o
,
“
Co
m
p
a
riso
n
b
e
t
w
e
e
n
Ne
u
ra
l
n
e
tw
o
rk
a
n
d
su
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
in
o
p
t
ica
l
c
h
a
ra
c
ter
re
c
o
g
n
it
io
n
,
”
Pro
c
e
d
i
a
C
o
mp
u
t.
S
c
i.
,
v
o
l.
1
1
6
,
p
p
.
3
5
1
-
3
5
7
,
2
0
1
7
,
d
o
i:
1
0
.
1
0
1
6
/j
.
p
ro
c
s.
2
0
1
7
.
1
0
.
0
6
1
.
[2
3
]
H.
Bisg
in
e
t
a
l.
,
“
Co
m
p
a
rin
g
S
VM
a
n
d
A
NN
b
a
se
d
m
a
c
h
in
e
lea
r
n
in
g
m
e
th
o
d
s
f
o
r
sp
e
c
ies
id
e
n
ti
f
ica
ti
o
n
o
f
f
o
o
d
c
o
n
tam
in
a
ti
n
g
b
e
e
tl
e
s,”
S
c
ien
ti
fi
c
R
e
p
o
rts
,
v
o
l.
8
,
n
o
.
6
5
3
2
,
p
p
.
1
-
1
2
,
2
0
1
8
,
d
o
i:
1
0
.
1
0
3
8
/s
4
1
5
9
8
-
0
1
8
-
2
4
9
2
6
-
7
.
[2
4
]
N.
Kra
n
jˇci´
c
,
D.
M
e
d
a
k
,
R.
Ž
u
p
a
n
,
a
n
d
M
.
Re
z
o
,
“
M
a
c
h
in
e
le
a
rn
in
g
m
e
th
o
d
s
f
o
r
c
las
sif
ica
ti
o
n
o
f
th
e
g
re
e
n
in
f
ra
stru
c
tu
re
in
c
it
y
a
re
a
s,”
In
t.
J
.
G
eo
-
In
f
,
v
o
l.
8
,
n
o
.
4
6
3
,
p
p
.
1
-
1
4
,
2
0
1
9
,
d
o
i:
1
0
.
3
3
9
0
/
ij
g
i8
1
0
0
4
6
3
.
[2
5
]
M
.
Ya
n
g
a
n
d
G
.
T
h
u
n
g
,
“
Clas
si
f
ica
ti
o
n
o
f
T
ra
sh
f
o
r
Re
c
y
c
lab
il
it
y
S
tatu
s,”
S
tan
f
o
rd
,
2
0
1
6
.
[
On
l
i
n
e
]
.
A
v
a
il
a
b
le:
h
tt
p
:
//
c
s2
2
9
.
sta
n
f
o
rd
.
e
d
u
/
p
ro
j2
0
1
6
/rep
o
r
t/
T
h
u
n
g
Ya
n
g
-
Clas
sif
ica
ti
o
n
Of
T
r
a
sh
F
o
rRe
c
y
c
lab
il
it
y
S
tatu
s
-
re
p
o
rt.
p
d
f
[2
6
]
S.
-
H.
T
sa
n
g
,
“
Re
v
ie
w
:
P
S
P
Ne
t
-
W
in
n
e
r
in
IL
S
V
RC
2
0
1
6
(S
e
m
a
n
ti
c
S
e
g
m
e
n
tatio
n
/
S
c
e
n
e
P
a
rsin
g
)
,
”
T
o
w
a
rd
s
Da
ta
S
c
ien
c
e
,
1
5
De
c
e
m
b
e
r,
2
0
1
8
.
A
c
c
e
ss
e
d
:
2
8
-
Oc
t
-
2
0
2
0
.
[
On
li
n
e
]
.
Av
a
il
a
b
le:
h
tt
p
s:
//
to
w
a
rd
sd
a
tas
c
ien
c
e
.
c
o
m
/r
e
v
ie
w
-
p
sp
n
e
t
-
w
in
n
e
r
-
in
-
il
sv
rc
-
2
0
1
6
-
se
m
a
n
ti
c
-
se
g
m
e
n
tatio
n
-
sc
e
n
e
-
p
a
rsi
n
g
-
e
0
8
9
e
5
d
f
1
7
7
d
.
[2
7
]
M.
-
P
.
H
o
ss
e
in
i,
M
.
R.
N
.
-
Z
a
d
e
h
,
D
.
P
o
m
p
il
i
,
a
n
d
H
.
S
.
-
Zad
e
h
“
S
tatisti
c
a
l
v
a
li
d
a
ti
o
n
o
f
a
u
to
m
a
ti
c
m
e
th
o
d
s
f
o
r
h
ip
p
o
c
a
m
p
u
s
se
g
m
e
n
tatio
n
in
M
R
im
a
g
e
s
o
f
e
p
il
e
p
ti
c
p
a
ti
e
n
ts,”
in
In
t.
C
o
n
f
.
o
f
th
e
I
EE
E
E
n
g
.
in
M
e
d
icin
e
a
n
d
Bi
o
l.
S
o
c
.
,
A
u
g
.
2
0
1
4
,
d
o
i
:
1
0
.
1
1
0
9
/em
b
c
.
2
0
1
4
.
6
9
4
4
6
7
5
.
[2
8
]
S
.
M
.
S
in
g
h
a
n
d
K
.
He
m
a
c
h
a
n
d
ra
n
,
“
Co
n
ten
t
-
b
a
se
d
im
a
g
e
re
tri
e
v
a
l
u
sin
g
c
o
lo
r
m
o
m
e
n
t
a
n
d
g
a
b
o
r
tex
tu
re
f
e
a
tu
re
,
”
IJ
CS
I
In
t.
J
.
Co
m
p
u
t
.
S
c
i.
Iss
u
e
s
,
v
o
l.
9
,
n
o
.
5
,
p
p
.
2
9
9
-
3
0
9
,
2
0
1
2
.
[2
9
]
Ş
.
Öz
tü
rk
a
n
d
B
.
A
k
d
e
m
ir,
“
A
p
p
li
c
a
ti
o
n
o
f
fe
a
tu
re
e
x
trac
ti
o
n
a
n
d
c
las
sif
ic
a
ti
o
n
m
e
th
o
d
s
f
o
r
h
is
t
o
p
a
t
h
o
l
o
g
ica
l
im
a
g
e
us
in
g
GL
CM
,
L
BP
,
L
B
GL
CM
,
GL
R
L
M
a
n
d
S
F
TA
,
”
Pro
c
e
d
ia
Co
mp
u
t
.
S
c
i
.
,
v
o
l
.
1
3
2
,
p
p
.
4
0
-
4
6
,
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j.
p
r
o
c
s.2
0
1
8
.
0
5
.
0
5
7
.
[3
0
]
P
.
S
.
S
.
Ku
m
a
r
a
n
d
V
.
S
.
Dh
a
ru
n
,
“
Ex
trac
ti
o
n
o
f
tex
tu
re
f
e
a
tu
re
s
u
sin
g
GL
CM
a
n
d
sh
a
p
e
f
e
a
tu
re
s
u
sin
g
c
o
n
n
e
c
ted
re
g
io
n
s,”
In
t.
J
.
En
g
.
a
n
d
T
e
c
h
n
o
l.
(
IJ
ET
)
,
v
o
l.
8
,
n
o
.
6
,
p
p
.
2
9
2
6
-
2
9
3
0
,
2
0
1
7
,
d
o
i:
1
0
.
2
1
8
1
7
/i
jet/
2
0
1
6
/v
8
i6
/
1
6
0
8
0
6
2
5
4
.
[3
1
]
J.
Ha
n
a
n
d
M
.
Ka
m
b
e
r,
Da
ta
M
i
n
in
g
Co
n
c
e
p
ts
a
n
d
T
e
c
h
n
i
q
u
e
s
S
e
c
o
n
d
E
d
it
i
o
n
,
M
o
rg
a
n
Ka
u
f
m
a
n
n
P
u
b
li
s
h
e
r
,
2
0
0
6
.
[
On
li
n
e
]
.
A
v
a
il
a
b
le:
h
tt
p
s://
m
it
m
e
c
se
p
t.
f
il
e
s.
w
o
rd
p
re
ss
.
c
o
m
/2
0
1
7
/0
4
/d
a
ta
-
m
in
in
g
-
c
o
n
c
e
p
ts
-
a
n
d
-
te
c
h
n
iq
u
e
s
-
2
n
d
-
e
d
it
io
n
-
im
p
re
ss
a
o
.
p
d
f
[3
2
]
C
.
-
C
.
Ch
a
n
g
a
n
d
C.
-
J.
L
in
,
“
L
IBS
V
M
:
A
L
ib
ra
r
y
f
o
r
S
u
p
p
o
r
t
V
e
c
to
r
M
a
c
h
i
n
e
s
,
”
ACM
T
ra
n
s.
In
tell
.
S
y
st.
T
e
c
h
n
o
l.
,
v
o
l.
2
,
n
o
.
3
,
p
p
.
1
-
2
7
,
2
0
1
1
,
d
o
i:
1
0
.
1
1
4
5
/
1
9
6
1
1
8
9
.
1
9
6
1
1
9
9
.
[3
3
]
A
.
Ka
ra
tzo
g
lo
u
,
D.
M
e
y
e
r
,
a
n
d
K
.
Ho
r
n
ik
,
“
S
u
p
p
o
rt
V
e
c
to
r
M
a
c
h
i
n
e
s
in
R
,
”
J
o
u
rn
a
l
o
f
S
t
a
t
isti
c
a
l
S
o
ft
w
er
,
v
o
l
.
1
5
,
n
o
.
9
,
p
p
.
1
-
2
8
,
2
0
0
6
,
d
o
i:
1
0
.
1
8
6
3
7
/j
ss
.
v
0
1
5
.
i0
9
.
[3
4
]
M
.
S
o
k
o
l
o
v
a
a
n
d
G
.
L
a
p
a
l
m
e
,
“
A
s
y
ste
m
a
ti
c
a
n
a
ly
sis
o
f
p
e
rf
o
rm
a
n
c
e
m
e
a
su
re
s
f
o
r
c
las
si
f
ica
ti
o
n
tas
k
s,”
In
f
o
rm
a
ti
o
n
Pro
c
e
ss
in
g
a
n
d
M
a
n
a
g
e
me
n
t
,
v
o
l
.
4
5
,
n
o
.
4
,
p
p
.
4
2
7
-
4
3
7
,
2
0
1
9
,
d
o
i
:
1
0
.
1
0
1
6
/j
.
i
p
m
.
2
0
0
9
.
0
3
.
0
0
2
.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
K
h
a
d
ij
a
h
o
b
tain
e
d
t
h
e
B
a
c
h
e
lo
r
of
In
f
o
rm
a
ti
c
s
En
g
in
e
e
rin
g
(S
.
Ko
m
)
f
ro
m
th
e
Un
iv
e
rsitas
Dip
o
n
e
g
o
r
o
,
In
d
o
n
e
sia
in
2
0
1
1
a
n
d
th
e
M
a
ste
r
of
Co
m
p
u
ter
S
c
ien
c
e
(M
.
Cs)
f
ro
m
th
e
Un
iv
e
rsitas
G
a
d
jah
M
a
d
a
,
In
d
o
n
e
sia
in
2
0
1
4
.
S
h
e
h
a
s
b
e
e
n
a
lec
tu
r
e
r
w
it
h
th
e
De
p
a
rtm
e
n
t
o
f
In
f
o
rm
a
ti
c
s,
Un
iv
e
rsita
s
Dip
o
n
e
g
o
ro
sin
c
e
2
0
1
4
.
He
r
m
a
in
re
se
a
rc
h
in
tere
sts
a
re
a
rti
f
icia
l
in
telli
g
e
n
c
e
a
n
d
m
a
c
h
in
e
lea
rn
i
n
g
.
S
h
e
is
a
lso
a
m
e
m
b
e
r
o
f
in
tern
a
ti
o
n
a
l
p
r
o
f
e
sio
n
a
l
a
ss
o
c
iatio
n
,
s
u
c
h
a
s
IEE
E.
S
u
k
m
a
w
a
ti
Nu
r
En
d
a
h
o
b
tain
e
d
t
h
e
Ba
c
h
e
lo
r
o
f
S
c
ien
c
e
(S
.
S
i)
in
M
a
th
e
m
a
ti
c
s
f
ro
m
th
e
Un
iv
e
rsitas
Dip
o
n
e
g
o
r
o
,
I
n
d
o
n
e
si
a
in
2
0
0
1
a
n
d
th
e
M
a
ste
r
of
C
o
m
p
u
ter
S
c
ien
c
e
(M
.
Ko
m
)
f
ro
m
th
e
Un
iv
e
rsitas
I
n
d
o
n
e
sia
,
I
n
d
o
n
e
sia
i
n
2
0
0
9
.
S
h
e
is
c
u
rre
n
tl
y
w
o
rk
in
g
a
s
a
lec
tu
re
r
w
it
h
th
e
De
p
a
rtme
n
t
o
f
In
f
o
r
m
a
ti
c
s,
Un
iv
e
rsitas
Dip
o
n
e
g
o
ro
.
S
in
c
e
2
0
1
9
,
s
h
e
h
a
s
b
e
e
n
th
e
H
e
a
d
o
f
L
a
b
o
ra
to
riu
m
o
f
In
telli
g
e
n
t
S
y
ste
m
in
th
e
De
p
a
rtm
e
n
t
o
f
In
f
o
rm
a
ti
c
s,
Un
iv
e
rsitas
Dip
o
n
e
g
o
r
o
.
He
r
m
a
in
re
se
a
rc
h
in
tere
sts
a
re
a
rti
f
icia
l
i
n
telli
g
e
n
c
e
a
n
d
m
a
c
h
in
e
l
e
a
rn
i
n
g
.
S
h
e
is
a
lso
th
e
m
e
m
b
e
r
o
f
A
P
T
IKO
M
,
t
h
e
In
d
o
n
e
sia
n
p
ro
f
e
ss
io
n
a
l
o
rg
a
n
iza
ti
o
n
i
n
c
o
m
p
u
ter
sc
ien
c
e
.
Evaluation Warning : The document was created with Spire.PDF for Python.