I
nte
rna
t
io
na
l J
o
urna
l o
f
I
nfo
rm
a
t
ics a
nd
Co
m
m
un
ica
t
io
n T
ec
hn
o
lo
g
y
(
I
J
-
I
CT
)
Vo
l.
15
,
No
.
3
,
Sep
tem
b
er
20
26
,
p
p
.
1
3
7
6
~
1
3
8
4
I
SS
N:
2252
-
8
7
7
6
,
DOI
:
1
0
.
1
1
5
9
1
/iji
ct
.
v15
i
3
.
pp
1
3
7
6
-
1
3
8
4
1376
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ict.
ia
esco
r
e.
co
m
A multi
-
e
x
per
t
a
p
pro
a
ch t
o
conte
nt
-
ba
sed ima
g
e re
tr
iev
a
l using
feature
f
usio
n an
d late r
e
-
ra
nking
Ali A
bd
ula
ze
ez
M
o
ha
m
m
ed
B
a
qer
Q
a
zz
a
z
1
,
Yo
us
if
Sa
m
e
r
M
ud
ha
f
a
r
1,
2
1
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
F
a
c
u
l
t
y
o
f
E
d
u
c
a
t
i
o
n
,
U
n
i
v
e
r
s
i
t
y
o
f
K
u
f
a
,
N
a
j
a
f
,
I
r
a
q
2
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
Te
c
h
n
i
q
u
e
s
En
g
i
n
e
e
r
i
n
g
,
F
a
c
u
l
t
y
o
f
Te
c
h
n
i
c
a
l
E
n
g
i
n
e
e
r
i
n
g
,
T
h
e
I
sl
a
mi
c
U
n
i
v
e
r
si
t
y
,
N
a
j
a
f
,
I
r
a
q
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Sep
2
5
,
2
0
2
5
R
ev
is
ed
Ap
r
1
8
,
2
0
2
6
Acc
ep
ted
J
u
l 5
,
2
0
2
6
As
d
ig
it
a
l
d
a
ta
ra
p
i
d
l
y
g
ro
ws
,
c
o
n
ten
t
-
b
a
se
d
ima
g
e
re
tri
e
v
a
l
(
CBIR)
h
a
s
b
e
c
o
m
e
imp
o
rta
n
t
fo
r
o
p
ti
m
izi
n
g
c
o
ll
e
c
ti
o
n
s
o
f
v
is
u
a
l
d
a
ta.
Th
is
w
o
rk
p
ro
p
o
se
s
a
re
tri
e
v
a
l
fra
m
e
wo
rk
wh
ich
o
p
e
ra
tes
in
two
sta
g
e
s
a
n
d
imp
ro
v
e
s
a
c
c
u
ra
c
y
b
y
u
si
n
g
sy
ste
m
a
ti
c
fu
si
o
n
o
f
fe
a
tu
re
s.
In
t
h
e
first
sta
g
e
,
f
irst
-
sta
g
e
wid
e
-
sc
o
p
e
d
e
sc
rip
to
rs
c
a
ll
e
d
b
a
g
-
of
-
v
isu
a
l
-
wo
r
d
s
(Bo
VW)
,
sc
a
tt
e
rin
g
wa
v
e
let
tran
sfo
rm
(S
WT
)
,
d
isc
re
te
c
o
sin
e
tran
sfo
rm
(DCT),
a
n
d
p
ri
n
c
ip
a
l
c
o
m
p
o
n
e
n
t
a
n
a
ly
sis
(P
CA)
re
tri
e
v
e
in
it
ial
c
a
n
d
i
d
a
te
ima
g
e
s.
Th
e
se
c
o
n
d
sta
g
e
u
n
d
e
rtak
e
s
d
e
taile
d
re
-
o
rd
e
rin
g
o
f
c
a
n
d
i
d
a
te
ima
g
e
s
b
y
imp
lem
e
n
ti
n
g
th
e
lo
c
a
l
b
i
n
a
ry
p
a
tt
e
rn
(LBP
)
,
h
isto
g
ra
m
o
f
o
rien
ted
g
ra
d
ien
ts
(HO
G
),
a
n
d
sin
g
u
lar
v
a
l
u
e
d
e
c
o
m
p
o
siti
o
n
(
S
VD
)
d
e
sc
rip
t
o
rs
t
o
re
-
e
v
a
lu
a
te
sim
il
a
rit
y
sc
o
re
s.
Eac
h
i
n
d
i
v
i
d
u
a
l
d
e
sc
rip
to
r
re
tu
r
n
e
d
re
su
lt
s
f
o
r
m
e
a
n
a
v
e
ra
g
e
p
re
c
isio
n
fo
r
th
e
to
p
1
0
re
tri
e
v
e
d
ima
g
e
s
(m
AP,
to
p
-
1
0
)
o
f
b
e
twe
e
n
0
.
6
3
a
n
d
0
.
7
9
a
n
d
th
e
fu
se
d
fra
m
e
wo
rk
a
c
h
iev
e
d
0
.
8
8
,
wh
ich
is
e
v
id
e
n
c
e
o
f
t
h
e
v
iab
il
i
ty
o
f
c
o
m
p
lem
e
n
tary
fe
a
tu
re
in
teg
ra
ti
o
n
.
T
h
e
se
fin
d
in
g
s
s
u
p
p
o
rt
t
h
e
h
y
p
o
t
h
e
s
is
th
a
t
wh
il
e
m
u
lt
i
p
le d
e
sc
rip
to
rs p
e
rfo
rm
e
d
we
ll
a
n
d
d
e
li
v
e
re
d
h
ig
h
re
tri
e
v
a
l
a
c
c
u
ra
c
y
,
h
iera
rc
h
ica
l
fu
sio
n
o
f
m
u
lt
ip
le
h
a
n
d
c
ra
fted
d
e
sc
rip
to
rs
d
o
e
s
n
o
t
i
n
v
o
lv
e
t
h
e
c
o
m
p
u
tati
o
n
a
l
c
o
sts
a
ss
o
c
iate
d
wit
h
d
e
e
p
lea
rn
in
g
m
e
th
o
d
s.
K
ey
w
o
r
d
s
:
C
o
n
ten
t
-
b
ased
im
ag
e
r
etr
iev
al
Dis
cr
ete
co
s
in
e
tr
an
s
f
o
r
m
Prin
cip
al
co
m
p
o
n
en
t a
n
al
y
s
is
Scatter
in
g
wav
elet
tr
an
s
f
o
r
m
Sin
g
u
lar
v
alu
e
d
ec
o
m
p
o
s
itio
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
:
Yo
u
s
if
Sam
er
Mu
d
h
a
f
ar
Dep
ar
tm
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
,
Facu
lty
o
f
E
d
u
ca
tio
n
,
Un
i
v
er
s
ity
o
f
Ku
f
a
Al
-
Ku
f
a
Stre
et,
Naja
f
,
Ku
f
a,
I
r
aq
E
m
ail:
y
o
u
s
if
s
.
m
u
d
h
af
ar
@
u
o
k
u
f
a.
ed
u
.
iq
1.
I
NT
RO
D
UCT
I
O
N
As
we
en
ter
a
v
is
u
al
d
ata
-
ce
n
ter
ed
wo
r
ld
,
ef
f
icien
tly
ac
ce
s
s
in
g
im
ag
es
f
r
o
m
a
n
en
o
r
m
o
u
s
,
u
n
s
tr
u
ctu
r
ed
d
atab
ase
in
r
esp
o
n
s
e
to
a
u
s
er
q
u
er
y
h
as
b
ec
o
m
e
a
p
r
ess
in
g
p
r
o
b
lem
.
C
o
n
ten
t
-
b
ased
im
a
g
e
r
etr
ie
v
al
(
C
B
I
R
)
s
y
s
tem
s
co
n
f
r
o
n
t
th
is
p
r
o
b
lem
b
y
allo
win
g
th
e
u
s
er
to
s
ea
r
ch
im
ag
es
b
ased
o
n
v
is
u
al
co
n
ten
t
-
b
ased
f
ea
t
u
r
es
s
u
ch
a
s
co
lo
r
,
tex
tu
r
e,
a
n
d
s
h
ap
e
in
s
tead
o
f
r
ely
in
g
o
n
h
u
m
an
a
n
n
o
tated
m
eta
-
d
ata.
C
B
I
R
is
im
p
o
r
tan
t
in
ap
p
licat
io
n
s
lik
e
d
ig
ital
lib
r
ar
ie
s
,
m
ed
ical
im
ag
in
g
,
e
-
co
m
m
e
r
ce
,
a
n
d
law
en
f
o
r
ce
m
en
t,
wh
er
e
th
e
n
ee
d
f
o
r
lar
g
e
-
s
ca
l
e
an
d
ac
cu
r
ate
r
etr
iev
al
o
f
im
ag
es
co
n
tin
u
es
to
g
r
o
w
[
1
]
.
C
B
I
R
,
an
d
esp
ec
ially
C
B
I
R
th
at
u
s
es
d
ee
p
lea
r
n
in
g
m
eth
o
d
s
b
ased
o
n
co
n
v
o
l
u
tio
n
al
n
eu
r
al
n
etwo
r
k
s
(
C
NNs),
h
av
e
b
ec
o
m
e
a
v
er
s
atile
an
d
d
o
m
i
n
an
t
ap
p
r
o
ac
h
th
at
ca
n
lear
n
h
ier
ar
c
h
ica
l
f
ea
tu
r
e
r
ep
r
esen
tatio
n
s
f
r
o
m
v
er
y
lar
g
e
d
atasets
an
d
ca
n
b
e
co
n
s
id
er
ed
s
tate
-
of
-
th
e
-
a
r
t
[
2
]
.
Ho
wev
er
,
tr
ad
i
tio
n
al
C
NN
m
o
d
els
f
u
n
ctio
n
as
b
lack
b
o
x
es
a
n
d
lack
ex
p
lan
atio
n
o
f
th
ei
r
o
u
tp
u
t:
th
is
is
esp
ec
ially
co
n
ce
r
n
in
g
in
h
ig
h
-
s
tak
es
ap
p
licatio
n
s
,
e.
g
.
m
ed
ica
l
d
iag
n
o
s
tics
,
wh
er
e
an
e
x
p
lan
a
tio
n
is
war
r
an
ted
.
Fu
r
th
er
m
o
r
e,
C
NNs
r
eq
u
ir
e
m
ass
iv
e
am
o
u
n
ts
o
f
la
b
eled
im
ag
e
d
ata
a
n
d
c
o
m
p
u
tatio
n
a
l
p
o
wer
to
tr
ain
,
wh
ich
m
a
k
e
s
th
em
im
p
r
ac
tical
in
en
v
ir
o
n
m
en
ts
with
lim
ited
ac
ce
s
s
to
r
eso
u
r
ce
s
[
3
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
A
mu
lti
-
ex
p
ert a
p
p
r
o
a
c
h
to
c
o
n
ten
t
-
b
a
s
ed
im
a
g
e
r
etri
ev
a
l
… (
A
li A
b
d
u
l
a
z
ee
z
Mo
h
a
mme
d
B
a
q
er
Qa
z
z
a
z
)
1377
T
h
is
r
esear
ch
r
etu
r
n
s
to
class
ical
h
an
d
c
r
af
ted
f
ea
tu
r
e
-
b
ased
im
ag
e
r
etr
iev
al
s
y
s
tem
s
citin
g
d
esira
b
le
p
r
o
p
er
ties
o
f
ef
f
icien
c
y
a
n
d
e
x
p
lain
ab
ilit
y
o
v
er
b
lack
-
b
o
x
d
ee
p
lear
n
i
n
g
ap
p
r
o
ac
h
es
f
o
r
r
e
tr
iev
al.
T
h
e
au
t
h
o
r
s
s
u
g
g
est
th
at
ag
g
r
eg
atin
g
s
ev
er
al
co
m
p
lem
en
tar
y
f
ea
tu
r
e
ex
tr
ac
to
r
s
,
ev
en
in
a
s
in
g
le
r
etr
iev
al
ap
p
licatio
n
,
will
p
r
o
v
id
e
a
r
ich
er
,
m
o
r
e
co
m
p
r
eh
en
s
iv
e,
an
d
m
o
r
e
in
ter
p
r
et
ab
le
u
n
d
er
s
tan
d
i
n
g
o
f
im
ag
er
y
th
an
an
y
s
in
g
le
ex
tr
ac
ted
-
d
escr
ip
to
r
co
u
ld
[
4
]
.
E
ac
h
ex
p
er
t
f
ea
tu
r
e
ex
tr
a
cto
r
s
d
escr
ib
es
u
n
iq
u
e
f
ea
tu
r
e
ch
ar
ac
ter
is
tics
,
in
clu
d
in
g
tex
tu
r
e,
s
tr
u
ctu
r
e
,
an
d
f
r
eq
u
e
n
cy
r
ep
r
esen
tatio
n
,
with
in
a
s
im
p
le
m
u
lti
-
s
tag
e,
tr
ain
in
g
-
f
r
ee
u
s
er
im
p
r
o
v
em
e
n
t f
r
am
ew
o
r
k
[
5
]
.
T
h
e
p
r
o
p
o
s
ed
m
u
lti
-
ex
p
er
t
s
y
s
tem
em
p
lo
y
s
s
ev
er
al
k
ey
d
escr
ip
to
r
s
:
th
e
b
ag
-
of
-
v
i
s
u
al
-
wo
r
d
s
(
B
o
VW
)
m
o
d
el
ca
p
tu
r
es
l
o
c
al
k
ey
p
o
in
ts
th
r
o
u
g
h
cl
u
s
ter
ed
v
is
u
al
v
o
ca
b
u
lar
ies;
th
e
s
ca
tter
in
g
wav
elet
tr
an
s
f
o
r
m
(
SW
T
)
ex
tr
ac
ts
m
u
lti
-
s
ca
le
s
tr
u
ctu
r
e;
th
e
d
is
cr
ete
co
s
in
e
tr
a
n
s
f
o
r
m
(
DC
T
)
en
co
d
es
f
r
eq
u
en
cy
co
m
p
o
n
en
ts
;
an
d
p
r
in
cip
al
c
o
m
p
o
n
en
t
an
aly
s
is
(
PC
A
)
co
m
p
r
ess
es
co
lo
r
h
is
to
g
r
am
s
.
L
o
ca
l
b
in
ar
y
p
att
er
n
(
L
B
P)
an
d
s
in
g
u
lar
v
alu
e
d
ec
o
m
p
o
s
itio
n
(
SVD)
p
r
o
ce
d
u
r
es
co
n
d
u
ct
d
im
en
s
io
n
ality
r
ed
u
ctio
n
an
d
f
in
e
r
an
k
in
g
.
T
h
e
late
-
f
u
s
io
n
p
h
as
e
em
p
lo
y
s
a
weig
h
ted
v
o
te
to
ag
g
r
eg
ate
th
e
ex
p
er
ts
'
o
u
tp
u
ts
,
in
clu
d
in
g
a
f
in
e
-
g
r
ain
ed
r
e
-
r
an
k
in
g
to
ac
h
ie
v
e
a
f
in
al
ac
c
u
r
ate
r
etr
iev
al.
T
h
e
k
ey
co
n
tr
ib
u
tio
n
s
ar
e:
i)
a
h
y
b
r
id
C
B
I
R
f
r
am
ewo
r
k
f
u
s
in
g
d
iv
er
s
e
h
a
n
d
cr
af
ted
d
escr
ip
to
r
s
;
ii)
a
weig
h
ted
v
o
tin
g
a
n
d
r
e
-
r
an
k
in
g
m
e
ch
an
is
m
f
o
r
r
o
b
u
s
t
f
u
s
io
n
; iii)
an
ev
al
u
atio
n
d
em
o
n
s
tr
atin
g
co
m
p
etitiv
e
p
er
f
o
r
m
an
ce
o
n
a
lar
g
e
p
u
b
lic
d
ataset.
2.
RE
L
AT
E
D
WO
RK
Ma
n
y
s
tu
d
ies
h
av
e
s
h
o
wn
p
r
o
g
r
ess
to
war
d
s
im
p
r
o
v
in
g
C
B
I
R
u
s
in
g
v
ar
io
u
s
m
eth
o
d
s
,
p
ar
ticu
lar
ly
th
r
o
u
g
h
th
e
u
s
e
o
f
d
ee
p
lea
r
n
i
n
g
an
d
h
y
b
r
id
m
eth
o
d
s
.
I
n
r
ef
er
en
ce
to
s
tu
d
y
[
5
]
,
a
d
u
al
-
C
NN
f
r
am
ewo
r
k
u
s
ed
r
ed
,
g
r
ee
n
,
b
lu
e
(
R
GB
)
an
d
tex
tu
r
e
-
b
ased
r
ep
r
esen
tatio
n
s
f
r
o
m
th
e
v
ar
io
u
s
L
B
P
v
ar
ian
ts
,
n
am
ely
o
r
th
o
g
o
n
a
l
co
m
b
in
atio
n
lo
ca
l
b
i
n
ar
y
p
at
ter
n
(
OC
-
L
B
P),
ce
n
ter
-
s
y
m
m
etr
ic
lo
ca
l
b
in
ar
y
p
atter
n
(
C
SLBP
)
,
an
d
lo
ca
l
d
ir
ec
tio
n
al
p
atter
n
(
L
DP)
,
ac
h
iev
ed
av
e
r
ag
e
p
r
ec
is
io
n
s
co
r
es
o
f
9
4
.
5
%,
8
9
.
7
%
an
d
8
8
.
7
%
r
esp
ec
tiv
ely
o
n
th
e
C
o
r
el
-
1
K,
C
altec
h
-
2
5
6
an
d
O
x
f
o
r
d
1
0
2
Flo
wer
s
im
ag
e
d
at
asets
.
Usi
n
g
a
g
r
ea
ter
n
u
m
b
e
r
o
f
e
n
co
d
er
ty
p
es
s
ig
n
if
ican
tly
in
cr
ea
s
es c
o
m
p
u
t
atio
n
al
co
m
p
lex
ity
.
Stu
d
y
b
y
Ma
g
lian
i
et
a
l.
[
6
]
f
o
cu
s
ed
o
n
cr
ea
tin
g
an
en
c
r
y
p
t
io
n
b
ased
s
o
lu
tio
n
f
o
r
p
r
o
v
id
i
n
g
p
r
iv
ac
y
as
well
as
r
ec
eiv
in
g
im
ag
e
r
etu
r
n
s
th
r
o
u
g
h
a
p
r
iv
ac
y
p
r
eser
v
in
g
C
B
I
R
s
y
s
tem
wh
ich
allo
ws
u
s
er
s
to
r
etr
iev
e
im
ag
e
r
esu
lts
ag
ain
s
t
e
n
cr
y
p
ted
im
ag
es
u
s
in
g
a
g
g
r
eg
atio
n
o
f
lo
ca
l
im
ag
e
f
ea
tu
r
e
d
a
ta
an
d
u
s
e
o
f
a
n
asy
m
m
etr
ic
s
ca
lar
p
r
o
d
u
ct
p
r
eser
v
in
g
e
n
cr
y
p
tio
n
.
W
h
ile
th
ese
ad
d
r
ess
es
p
r
o
v
id
e
p
r
iv
ac
y
co
n
s
id
er
atio
n
s
,
it
g
en
er
ates
ad
d
itio
n
al
co
m
p
u
t
atio
n
al
o
v
er
h
ea
d
.
Stu
d
y
by
Ga
y
ath
r
i
[
7
]
p
r
o
p
o
s
ed
a
s
k
etc
h
b
ased
p
r
o
ce
s
s
C
B
I
R
u
s
in
g
ed
g
e
h
is
to
g
r
am
d
escr
ip
t
o
r
s
an
d
s
ca
le
-
in
v
ar
ian
t
f
ea
tu
r
e
tr
an
s
f
o
r
m
(
SIFT
)
r
ef
in
em
e
n
ts
to
o
b
tain
r
elate
d
im
ag
e
r
esu
lts
;
h
o
wev
er
,
n
o
q
u
an
titativ
e
ev
alu
atio
n
o
n
a
la
r
g
e
d
ataset
was
p
r
o
v
id
ed
to
s
h
o
w
ac
cu
r
ac
y
o
r
v
iab
ilit
y
o
f
th
is
m
eth
o
d
,
th
u
s
lim
itin
g
its
o
v
er
all
ca
p
ab
ilit
ies to
1
1
0
i
m
ag
es.
Dee
p
lear
n
in
g
-
b
ased
m
et
h
o
d
s
h
av
e
also
s
h
o
wn
s
tr
o
n
g
p
er
f
o
r
m
an
ce
.
Vis
u
al
g
eo
m
etr
y
g
r
o
u
p
1
6
-
lay
er
n
etwo
r
k
(
VGG1
6
)
an
d
r
esid
u
al
n
etwo
r
k
-
5
0
(
R
esNet
-
50
)
wer
e
b
o
th
p
air
e
d
with
a
s
im
ilar
ity
m
atch
in
g
m
eth
o
d
o
l
o
g
y
f
o
r
an
o
v
er
al
l
r
esu
lt
o
f
8
1
.
6
8
%
p
r
ec
is
io
n
a
n
d
9
0
.
1
8
%
p
r
ec
is
io
n
,
r
esp
ec
tiv
ely
[
8
]
.
T
h
o
u
g
h
s
ca
lab
ilit
y
r
em
ain
s
li
m
ited
.
A
n
ew
lo
s
s
f
u
n
ctio
n
ca
lled
class
an
ch
o
r
m
ar
g
in
(
C
AM
)
in
tr
o
d
u
ce
d
by
Gh
ita
an
d
I
o
n
escu
in
[
9
]
is
ab
le
to
e
n
h
an
ce
f
ea
tu
r
e
em
b
ed
d
in
g
am
o
n
g
n
o
n
-
h
o
m
o
g
en
eo
u
s
d
atasets
(
e.
g
.
,
C
I
FAR
-
1
0
0
an
d
I
m
ag
eNe
t
-
2
0
0
)
an
d
o
f
f
er
s
im
p
r
o
v
ed
r
esu
lts
wh
en
co
m
p
a
r
ed
to
c
o
n
tr
asti
v
e
an
d
cr
o
s
s
-
en
tr
o
p
y
lo
s
s
es,
b
u
t h
as a
n
in
cr
ea
s
ed
s
e
n
s
itiv
ity
to
an
ch
o
r
in
itializatio
n
.
Hy
b
r
id
ap
p
r
o
ac
h
es
attem
p
t
to
b
alan
ce
ac
cu
r
ac
y
an
d
ef
f
i
cien
cy
.
T
h
e
E
f
f
icien
tNet
-
h
is
to
g
r
am
o
f
o
r
ien
ted
g
r
a
d
ien
ts
(
HOG
)
m
o
d
el
in
[
1
0
]
co
m
b
in
es
d
ee
p
lear
n
in
g
f
ea
tu
r
es
with
h
an
d
-
cr
af
ted
f
ea
t
u
r
es;
ac
h
iev
in
g
m
AP
v
alu
es
o
f
0
.
8
9
,
0
.
8
5
,
an
d
0
.
8
3
o
n
t
h
e
C
o
r
el
-
1
K,
Ox
f
o
r
d
5
K,
a
n
d
Par
is
6
K
d
atasets
,
r
esp
ec
tiv
ely
.
T
h
er
e
ar
e
s
till
lim
itatio
n
s
in
s
ca
lin
g
th
e
h
y
b
r
id
s
o
lu
tio
n
to
h
ig
h
-
r
eso
lu
tio
n
im
ag
es.
I
m
p
lem
en
tin
g
a
v
is
io
n
tr
an
s
f
o
r
m
er
(
ViT
)
-
b
a
s
ed
C
B
I
R
s
y
s
tem
in
[
1
1
]
,
alo
n
g
with
n
atu
r
al
lan
g
u
ag
e
p
r
o
ce
s
s
in
g
(
NL
P
)
-
d
r
iv
en
s
ea
r
ch
an
d
u
s
er
f
ee
d
b
ac
k
,
a
ch
iev
ed
a
n
F1
-
s
co
r
e
o
f
0
.
7
8
5
1
o
n
C
o
r
el
-
1
K,
b
u
t
th
is
r
esu
lted
in
g
r
ea
ter
co
m
p
u
tatio
n
al
c
o
m
p
lex
ity
an
d
d
ep
en
d
en
ce
o
n
u
s
er
in
ter
ac
tio
n
.
Ov
er
all,
d
ee
p
lear
n
i
n
g
-
b
ased
C
B
I
R
m
eth
o
d
s
,
in
clu
d
i
n
g
C
NNs
an
d
ViT
s
,
ac
h
iev
e
h
i
g
h
r
et
r
iev
al
ac
cu
r
ac
y
b
u
t
r
e
q
u
ir
e
lar
g
e
la
b
eled
d
atasets
an
d
s
ig
n
if
ica
n
t
c
o
m
p
u
tatio
n
al
r
eso
u
r
ce
s
[
1
2
]
.
T
h
ese
m
eth
o
d
s
a
r
e
ass
o
ciate
d
with
h
ig
h
r
etr
iev
al
ac
cu
r
ac
y
,
b
u
t
at
a
co
s
t
o
f
r
eq
u
ir
in
g
lar
g
e
,
lab
elled
d
atasets
an
d
s
ig
n
if
ican
t
co
m
p
u
tin
g
p
o
wer
[
1
2
]
.
Han
d
-
cr
af
ted
f
ea
tu
r
e
C
B
I
R
m
eth
o
d
s
o
n
th
e
o
th
er
h
an
d
ty
p
ically
d
em
o
n
s
tr
ate
ex
p
lo
it
ef
f
icien
cy
an
d
in
ter
p
r
etab
ilit
y
h
o
wev
e
r
t
h
ey
o
f
ten
lack
s
em
an
tic
r
ich
n
ess
wh
en
co
n
s
id
er
ed
as
s
tan
d
alo
n
e
C
B
I
R
m
eth
o
d
s
.
T
h
ese
lim
itatio
n
s
m
o
tiv
ate
t
h
e
p
r
o
p
o
s
ed
m
u
lti
ex
p
er
t
C
B
I
R
f
r
am
ewo
r
k
wh
ich
in
teg
r
ates
co
m
p
lem
en
tar
y
h
an
d
c
r
af
ted
d
escr
ip
to
r
s
with
in
a
h
ier
ar
ch
ical
f
u
s
io
n
an
d
r
e
-
r
an
k
i
n
g
s
tr
u
ctu
r
e
to
b
alan
ce
ac
cu
r
ac
y
ef
f
icien
cy
a
n
d
in
te
r
p
r
etab
ilit
y
.
T
ab
le
1
s
u
m
m
ar
ize
s
th
e
k
ey
d
i
f
f
er
en
ce
s
am
o
n
g
e
x
is
tin
g
ap
p
r
o
ac
h
es
in
ter
m
s
o
f
d
atasets
,
m
eth
o
d
o
l
o
g
ies,
p
er
f
o
r
m
an
ce
m
etr
ics,
a
d
v
an
tag
es,
an
d
lim
itatio
n
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
15
,
No
.
3
,
Sep
tem
b
er
20
26
:
1
3
7
6
-
1
3
8
4
1378
T
ab
le
1
.
Su
m
m
a
r
y
o
f
th
e
r
elate
d
wo
r
k
s
#
D
a
t
a
s
e
t
(
s)
u
se
d
M
e
t
h
o
d
o
l
o
g
y
su
mm
a
r
y
B
e
st
r
e
s
u
l
t
s a
n
d
met
r
i
c
s
A
d
v
a
n
t
a
g
e
s
Li
mi
t
a
t
i
o
n
s
[
5
]
C
o
r
e
l
-
1
K
,
C
a
l
t
e
c
h
-
2
5
6
,
O
x
f
o
r
d
1
0
2
F
e
a
t
u
r
e
f
u
si
o
n
e
n
h
a
n
c
e
d
(
LD
P
)
i
ma
g
e
s
u
s
i
n
g
d
u
a
l
C
N
N
e
n
c
o
d
e
r
s.
A
v
g
.
P
r
e
c
i
si
o
n
:
9
4
.
5
%
(
o
n
C
o
r
e
l
-
1
K
)
R
o
b
u
st
n
e
ss
t
o
c
l
a
ss
i
mb
a
l
a
n
c
e
;
s
u
p
e
r
i
o
r
p
r
e
c
i
si
o
n
v
i
a
f
e
a
t
u
r
e
f
u
si
o
n
.
M
u
l
t
i
p
l
e
e
n
c
o
d
e
r
s
i
n
c
r
e
a
s
e
c
o
m
p
u
t
a
t
i
o
n
a
l
c
o
s
t
a
n
d
t
r
a
i
n
i
n
g
c
o
m
p
l
e
x
i
t
y
.
[
6
]
O
x
f
o
r
d
5
K
,
P
a
r
i
s6
K
D
e
e
p
C
N
N
f
e
a
t
u
r
e
s
w
i
t
h
D
D
R
a
n
d
l
o
c
V
LA
D
a
g
g
r
e
g
a
t
i
o
n
.
mA
P
:
0
.
7
1
8
8
(
o
n
P
a
r
i
s6
K
)
En
h
a
n
c
e
d
f
e
a
t
u
r
e
a
g
g
r
e
g
a
t
i
o
n
a
n
d
a
c
c
u
r
a
c
y
H
i
g
h
e
r
c
o
m
p
u
t
a
t
i
o
n
a
l
c
o
s
t
[
7
]
l
o
c
a
l
d
a
t
a
(
1
1
0
)
i
ma
g
e
s
S
k
e
t
c
h
-
b
a
s
e
d
u
si
n
g
Ed
g
e
H
i
st
o
g
r
a
m w
i
t
h
S
I
F
T.
N
o
me
t
r
i
c
s
p
r
o
v
i
d
e
d
.
I
n
t
e
r
f
a
c
e
:
r
o
b
u
st
t
o
n
o
i
se
.
La
c
k
o
f
r
i
g
o
r
o
u
s
e
v
a
l
u
a
t
i
o
n
;
v
e
r
y
sm
a
l
l
d
a
t
a
se
t
.
[
8
]
C
u
s
t
o
m
d
a
t
a
(
2
,
1
0
0
)
i
ma
g
e
s
V
G
G
1
6
/
R
e
sN
e
t
-
5
0
w
i
t
h
si
m
i
l
a
r
i
t
y
mat
c
h
i
n
g
.
P
r
e
c
i
s
i
o
n
:
9
0
.
1
8
%
(
R
e
sN
e
t
-
5
0
)
S
u
p
e
r
i
o
r
p
e
r
f
o
r
ma
n
c
e
;
e
f
f
i
c
i
e
n
t
r
e
t
r
i
e
v
a
l
.
I
mag
e
d
o
w
n
si
z
i
n
g
;
l
i
m
i
t
e
d
d
a
t
a
s
e
t
sc
o
p
e
.
[
9
]
C
I
F
A
R
-
1
0
0
,
F
o
o
d
-
1
0
1
,
S
V
H
N
,
I
mag
e
N
e
t
-
2
0
0
I
n
t
r
o
d
u
c
e
s C
A
M
l
o
ss
t
o
o
p
t
i
mi
z
e
e
mb
e
d
d
i
n
g
.
S
u
p
e
r
i
o
r
mA
P
c
o
m
p
a
r
e
d
t
o
c
o
n
t
r
a
s
t
i
v
e
a
n
d
c
r
o
ss
-
e
n
t
r
o
p
y
l
o
sses
.
El
i
m
i
n
a
t
e
s
c
o
s
t
l
y
p
a
i
r
mi
n
i
n
g
;
e
n
f
o
r
c
e
s
c
o
m
p
a
c
t
a
n
d
se
p
a
r
a
b
l
e
f
e
a
t
u
r
e
s;
e
f
f
i
c
i
e
n
t
.
S
e
n
s
i
t
i
v
e
t
o
a
n
c
h
o
r
i
n
i
t
i
a
l
i
z
a
t
i
o
n
;
mar
g
i
n
a
l
g
a
i
n
s
i
n
l
o
w
-
d
a
t
a
r
e
g
i
m
e
s.
[
1
0
]
C
o
r
e
l
-
1
K
,
O
x
f
o
r
d
5
K
,
P
a
r
i
s6
K
A
h
y
b
r
i
d
mo
d
e
l
c
o
m
b
i
n
i
n
g
H
O
G
a
n
d
Ef
f
i
c
i
e
n
t
N
e
t
c
o
-
a
t
t
e
n
t
i
o
n
m
e
c
h
a
n
i
sm
.
mA
P
:
0
.
8
9
(
o
n
C
o
r
e
l
-
1
K
)
A
d
a
p
t
i
v
e
f
e
a
t
u
r
e
f
u
si
o
n
;
c
o
m
p
u
t
a
t
i
o
n
a
l
e
f
f
i
c
i
e
n
c
y
.
R
e
l
i
a
n
c
e
o
n
H
O
G
f
o
r
l
o
c
a
l
f
e
a
t
u
r
e
s.
[
1
1
]
C
o
r
e
l
-
1K
V
i
T
w
i
t
h
N
LP
a
n
d
u
ser fe
e
d
b
a
c
k
mec
h
a
n
i
sms
.
F1
-
sc
o
r
e
:
0
.
7
8
5
1
S
e
ma
n
t
i
c
q
u
e
r
y
u
n
d
e
r
s
t
a
n
d
i
n
g
;
a
d
a
p
t
i
v
e
l
e
a
r
n
i
n
g
f
r
o
m
u
ser fe
e
d
b
a
c
k
.
H
i
g
h
c
o
m
p
u
t
a
t
i
o
n
a
l
o
v
e
r
h
e
a
d
;
d
e
p
e
n
d
e
n
c
y
o
n
f
e
e
d
b
a
c
k
q
u
a
l
i
t
y
.
3.
RE
S
E
ARCH
M
E
T
H
O
D
T
h
e
p
r
o
p
o
s
ed
C
B
I
R
f
r
am
ewo
r
k
is
estab
lis
h
ed
as
a
co
m
p
lete
m
u
lti
-
ex
p
er
t
s
y
s
tem
th
at
d
o
es
n
o
t
r
eq
u
ir
e
d
ata
-
h
ea
v
y
ar
ch
iv
al
d
ee
p
lear
n
in
g
tr
ain
in
g
.
I
n
th
is
p
r
o
ce
s
s
,
th
e
p
r
o
p
o
s
ed
f
r
am
e
wo
r
k
em
p
l
o
y
s
two
lev
els:
f
ea
tu
r
e
ex
tr
ac
tio
n
,
w
h
er
e
th
e
f
iv
e
co
m
p
lem
en
tar
y
v
is
u
al
d
escr
ip
to
r
s
ex
a
m
in
e
ea
ch
im
ag
e
u
n
d
e
r
d
if
f
er
en
t
asp
ec
ts
,
an
d
f
u
s
io
n
a
n
d
r
e
-
r
a
n
k
in
g
,
wh
er
e
th
e
o
u
tp
u
ts
f
o
r
th
e
f
iv
e
im
ag
e
d
escr
ip
t
o
r
s
ar
e
co
m
b
in
e
d
to
ar
r
iv
e
at
a
f
in
al
r
etr
iev
al
lis
t
th
at
is
h
ig
h
ly
p
r
ec
is
e,
ac
c
u
r
ate,
a
n
d
r
o
b
u
s
t.
T
h
is
h
ier
a
r
ch
ical
ar
c
h
itectu
r
e
lev
er
ag
es
th
e
s
tr
en
g
th
s
o
f
cl
ass
ical
m
eth
o
d
s
wh
ile
h
e
d
g
i
n
g
a
g
ain
s
t
th
e
wea
k
n
ess
es
b
y
u
s
e
o
f
a
s
im
p
le
d
ec
is
io
n
-
m
ak
in
g
p
r
o
ce
s
s
.
T
h
e
co
m
p
lete
m
o
d
el
ar
ch
itectu
r
e
i
s
d
ep
icted
in
Fig
u
r
e
1
.
Fig
u
r
e
1
.
T
h
e
b
lo
c
k
d
iag
r
am
o
f
th
e
p
r
o
p
o
s
ed
s
y
s
tem
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
A
mu
lti
-
ex
p
ert a
p
p
r
o
a
c
h
to
c
o
n
ten
t
-
b
a
s
ed
im
a
g
e
r
etri
ev
a
l
… (
A
li A
b
d
u
l
a
z
ee
z
Mo
h
a
mme
d
B
a
q
er
Qa
z
z
a
z
)
1379
3
.
1
.
F
e
a
t
ure
des
cr
ipto
rs o
v
e
rv
iew
T
h
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
co
m
b
in
es
m
u
ltip
le
class
ical
f
ea
tu
r
e
d
escr
ip
to
r
s
to
en
co
d
e
th
e
v
ar
io
u
s
v
is
u
al
b
eh
av
io
r
s
in
im
ag
es.
Fo
r
ex
a
m
p
le,
SIFT
u
s
es
g
r
ad
ien
t
-
b
ased
s
ig
n
atu
r
es
to
id
en
tify
s
alien
t
k
ey
p
o
in
ts
ac
r
o
s
s
m
u
ltip
le
s
ca
les
an
d
u
n
f
av
o
r
a
b
le
lig
h
tin
g
co
n
d
itio
n
s
w
h
ile
en
s
u
r
in
g
th
at
th
e
p
r
o
ce
s
s
r
e
m
ain
s
in
v
ar
ian
t
t
o
g
eo
m
etr
ic
c
h
an
g
es
a
n
d
illu
m
in
atio
n
[
1
3
]
.
B
o
VW
ap
p
r
o
ac
h
es
im
ag
es
as
co
m
p
ac
t
r
ep
r
esen
tatio
n
s
o
f
v
is
u
al
-
wo
r
d
d
is
tr
ib
u
tio
n
s
an
d
g
en
er
ates
h
is
to
g
r
am
r
ep
r
ese
n
tatio
n
s
f
r
o
m
lo
ca
l
f
ea
tu
r
e
v
ec
to
r
s
th
at
ca
n
b
e
clu
s
ter
ed
an
d
a
r
e
th
e
n
b
etter
s
u
ited
to
r
e
p
r
esen
tin
g
t
h
e
c
o
n
ten
t
o
f
im
ag
es
[
1
4
]
,
[
1
5
]
.
DC
T
u
s
es
co
s
in
e
f
u
n
ctio
n
s
to
d
ec
o
m
p
o
s
e
im
ag
es
in
to
f
r
e
q
u
en
c
y
c
o
m
p
o
n
en
ts
,
h
ig
h
lig
h
tin
g
th
e
lo
w
-
f
r
eq
u
en
cy
co
ef
f
icien
ts
th
at
ca
p
tu
r
e
th
e
g
lo
b
al
s
tr
u
ctu
r
e
o
f
an
im
ag
e
w
h
ile
f
ilter
in
g
o
u
t
n
o
is
e
[
1
6
]
.
L
B
P
en
co
d
es
r
el
atio
n
s
h
ip
s
b
etwe
en
p
ix
els
(
n
eig
h
b
o
r
h
o
o
d
s
)
in
to
b
in
ar
y
p
atter
n
s
b
ased
o
n
i
n
ten
s
ity
co
m
p
ar
is
o
n
s
alo
n
g
s
id
e
th
eir
n
eig
h
b
o
r
s
,
p
r
o
d
u
cin
g
h
is
to
g
r
am
s
th
at
ar
e
r
o
b
u
s
t
to
illu
m
in
atio
n
ch
an
g
es
[
1
7
]
.
SVD
d
ec
o
m
p
o
s
es
th
e
im
ag
e
m
atr
ix
in
to
s
in
g
u
lar
v
alu
es
th
at
co
r
r
esp
o
n
d
to
th
e
d
o
m
in
an
t
s
tr
u
ctu
r
al
co
m
p
o
n
e
n
ts
o
f
th
e
im
ag
es,
p
r
o
d
u
cin
g
co
m
p
ac
t,
d
is
cr
im
in
ativ
e
d
escr
ip
to
r
s
[
1
8
]
,
[
1
9
]
.
HOG
u
s
es
g
r
ad
ien
t
-
o
r
ien
tatio
n
s
f
r
o
m
ce
lls
,
wh
ile
c
ap
tu
r
in
g
s
h
ap
es
o
f
o
b
jects
th
r
o
u
g
h
s
p
atially
-
a
g
g
r
eg
ated
h
is
to
g
r
am
-
b
ased
g
r
a
d
ien
t
o
r
ie
n
tatio
n
s
to
e
n
co
d
e
th
e
f
ea
tu
r
es,
an
d
n
o
r
m
alizin
g
t
h
e
s
p
atial
g
r
ad
ie
n
t o
r
ien
tatio
n
s
h
is
to
g
r
am
to
as
s
u
r
e
r
o
b
u
s
tn
ess
to
illu
m
in
atio
n
ch
an
g
es
[
2
0
]
.
Af
ter
co
m
p
u
tin
g
t
h
ese
co
m
p
lem
en
tar
y
d
escr
ip
t
o
r
s
,
to
f
u
r
th
er
im
p
r
o
v
e
r
etr
iev
al
r
o
b
u
s
tn
ess
,
th
e
p
r
o
ce
s
s
o
f
f
ea
tu
r
e
f
u
s
io
n
in
teg
r
ates
m
u
ltip
le
d
escr
ip
to
r
s
b
y
e
ith
er
co
n
ca
te
n
atin
g
t
h
e
r
esu
lta
n
t
f
ea
tu
r
e
v
ec
to
r
o
r
co
m
b
in
in
g
th
e
d
ec
is
io
n
o
f
d
ir
ec
tin
g
m
o
d
els,
cr
ea
tin
g
th
e
ab
ilit
y
f
o
r
o
n
e
d
escr
ip
to
r
to
b
ala
n
ce
o
r
co
m
p
en
s
ate
f
o
r
th
e
lim
itatio
n
s
/ef
f
ec
ts
o
f
an
o
th
er
d
escr
ip
to
r
[
2
1
]
,
[
2
2
]
.
Fo
r
th
e
s
y
s
tem
ev
alu
atio
n
,
we
ad
o
p
t
th
e
m
ea
n
av
er
ag
e
p
r
ec
is
io
n
(m
AP)
e
v
a
lu
ated
o
n
th
e
to
p
p
ed
ten
,
a
v
e
r
ag
ed
ac
r
o
s
s
q
u
e
r
ies
an
d
im
a
g
es
f
o
r
r
eliab
ilit
y
,
co
m
p
u
tatio
n
o
f
m
AP
m
etr
ic
(
m
AP,
to
p
-
1
0
)
.
T
h
e
m
o
d
el
is
e
v
alu
ated
o
n
th
e
I
n
tel
I
m
ag
e
C
lass
if
icatio
n
d
ataset,
wh
ich
co
n
s
is
ts
o
f
1
4
0
3
4
tr
ain
in
g
im
ag
es
an
d
3
0
0
0
test
in
g
im
ag
es
ac
r
o
s
s
s
eg
r
eg
ated
in
t
o
s
ev
e
n
ca
teg
o
r
ies
-
b
u
ild
in
g
s
,
f
o
r
est,
g
lacie
r
,
m
o
u
n
tain
,
s
ea
,
a
n
d
s
tr
ee
t
-
ca
p
tu
r
in
g
v
ar
ied
tex
t
u
r
es,
lig
h
tin
g
,
an
d
v
iewp
o
in
ts
[
2
3
]
.
Fig
u
r
e
2
s
h
o
ws r
ep
r
esen
tativ
e
s
am
p
le
im
ag
es f
r
o
m
ea
c
h
o
f
th
ese
ca
teg
o
r
ies.
Fig
u
r
e
2
.
I
m
ag
es f
r
o
m
th
e
in
te
l im
ag
e
class
if
icatio
n
d
ataset
3
.
2
.
F
e
a
t
ure
ex
t
r
a
ct
io
n
T
o
en
s
u
r
e
a
co
m
p
r
eh
en
s
iv
e
r
ep
r
esen
tatio
n
,
s
ev
en
f
ea
tu
r
e
ex
tr
ac
tio
n
m
o
d
u
les
ar
e
ap
p
li
ed
to
ea
ch
im
ag
e,
ea
ch
ca
p
tu
r
in
g
d
is
tin
ct
v
is
u
al
p
r
o
p
e
r
ties
:
-
T
h
e
B
o
VW
m
o
d
el
u
s
es
lo
ca
l
f
ea
tu
r
es
th
at
wer
e
o
b
tain
ed
b
y
ex
t
r
ac
tin
g
SIFT
f
ea
tu
r
es
t
o
d
etec
t
e
x
tr
em
a
with
in
th
e
Do
G
s
p
ac
e,
g
en
e
r
atin
g
1
2
8
-
d
im
e
n
s
io
n
al
d
escr
ip
to
r
s
.
T
h
e
B
o
VW
m
o
d
el
u
s
es
Min
i
-
B
atch
K
-
m
ea
n
s
clu
s
ter
in
g
(
k
=2
0
0
)
t
o
cr
ea
te
a
v
is
u
al
v
o
ca
b
u
lar
y
,
with
ea
ch
im
ag
e
s
u
b
s
eq
u
en
tly
r
ep
r
esen
ted
b
y
a
n
o
r
m
alis
ed
h
is
to
g
r
a
m
o
f
v
is
u
a
l w
o
r
d
o
cc
u
r
r
en
ce
s
.
-
T
h
e
SW
T
ex
tr
ac
ts
tr
an
s
latio
n
-
in
v
ar
ian
t
f
ea
t
u
r
es
v
ia
th
e
u
s
e
o
f
m
u
lti
-
s
ca
le
f
ilter
in
g
(
ap
p
r
o
x
im
ately
J
=4
)
ac
r
o
s
s
m
u
ltip
le
o
r
ie
n
tatio
n
s
t
o
p
r
o
d
u
ce
f
ea
tu
r
es.
C
o
ef
f
icie
n
ts
ar
e
p
r
o
d
u
ce
d
a
n
d
a
v
er
ag
e
d
to
p
r
o
d
u
ce
a
f
ea
tu
r
e
v
ec
to
r
th
at
r
e
p
r
esen
ts
th
e
s
tr
u
ctu
r
al
p
atter
n
s
co
n
tain
e
d
with
in
th
e
o
r
ig
in
al
im
ag
es.
-
T
h
e
DC
T
tr
an
s
f
o
r
m
s
im
ag
es
to
a
f
r
eq
u
en
cy
-
b
ased
r
ep
r
esen
tatio
n
.
W
h
en
wo
r
k
in
g
with
R
GB
ch
an
n
els,
th
e
1
6
x
1
6
lo
w
-
f
r
e
q
u
en
c
y
co
ef
f
icien
ts
in
th
e
to
p
lef
t
co
r
n
er
we
r
e
s
elec
ted
f
r
o
m
ea
c
h
o
f
th
e
R
GB
ch
an
n
els,
av
er
ag
ed
a
n
d
co
n
ca
te
n
ated
t
o
p
r
o
d
u
ce
a
co
m
p
ac
t
d
escr
ip
to
r
.
T
h
e
co
m
p
ac
t
d
escr
ip
to
r
was
cr
ea
ted
b
y
en
s
u
r
in
g
th
at
it is
r
o
b
u
s
t to
n
o
is
e.
-
PC
A
wa
s
p
er
f
o
r
m
ed
o
n
a
c
o
lo
u
r
h
is
to
g
r
a
m
(
R
GB
h
is
to
g
r
am
,
5
1
2
-
d
im
en
s
io
n
al
s
p
ac
e
c
r
e
ated
v
ia
8
x
8
x
8
b
in
s
)
;
u
s
in
g
PC
A
th
e
h
is
to
g
r
am
was
r
ed
u
ce
d
f
r
o
m
a
5
1
2
-
d
i
m
en
s
io
n
al
s
p
ac
e
to
a
6
4
-
d
im
e
n
s
io
n
al
s
p
ac
e
b
y
s
elec
tin
g
th
e
p
r
in
cip
al
co
m
p
o
n
en
ts
th
at
p
r
eser
v
e
th
e
m
aj
o
r
i
ty
o
f
th
e
v
a
r
ian
ce
f
o
u
n
d
with
in
th
at
s
p
ac
e
to
cr
ea
te
a
co
m
p
ac
t c
o
lo
u
r
d
escr
i
p
to
r
.
-
L
B
P
p
r
o
v
id
es
a
way
t
o
en
c
o
d
e
th
e
lo
ca
l
tex
tu
r
e
v
ia
th
r
esh
o
ld
in
g
o
f
th
e
n
ei
g
h
b
o
u
r
h
o
o
d
p
i
x
els
in
o
r
d
er
to
p
r
o
d
u
ce
b
i
n
ar
y
p
atter
n
s
,
a
n
d
s
u
b
s
eq
u
en
tly
ag
g
r
eg
ate
th
o
s
e
i
n
to
h
is
to
g
r
a
m
s
th
at
ar
e
r
o
b
u
s
t
to
illu
m
in
atio
n
ch
an
g
es,
th
u
s
p
r
o
d
u
cin
g
a
n
L
B
P f
ea
tu
r
e.
-
SVD
d
ec
o
m
p
o
s
es
th
e
in
p
u
t
i
m
ag
e
m
atr
i
x
in
to
s
in
g
u
lar
v
al
u
es;
SVD
allo
ws
f
o
r
t
h
e
id
e
n
tific
atio
n
o
f
th
e
d
o
m
in
an
t
s
tr
u
ctu
r
al
ch
a
r
ac
ter
i
s
tics
o
f
th
e
o
r
ig
in
al
m
atr
ix
b
y
r
etain
in
g
th
e
lar
g
est
s
in
g
u
lar
v
alu
es.
As
s
u
ch
,
SVD
p
r
o
d
u
ce
s
a
co
m
p
ac
t f
ea
t
u
r
e
d
escr
ip
to
r
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
15
,
No
.
3
,
Sep
tem
b
er
20
26
:
1
3
7
6
-
1
3
8
4
1380
-
HOG
is
u
s
e
d
to
d
eter
m
in
e
t
h
e
s
h
ap
e
o
f
an
o
b
ject
u
s
in
g
th
e
d
ir
ec
tio
n
(
o
r
ie
n
tatio
n
)
o
f
th
e
g
r
ad
ie
n
t
b
y
ca
lcu
latin
g
an
o
r
ie
n
tatio
n
-
h
is
t
o
g
r
am
f
o
r
ar
ea
s
n
ea
r
a
l
o
ca
lized
r
eg
io
n
a
n
d
n
o
r
m
alizin
g
t
h
e
v
alu
es
s
o
th
at
ch
an
g
es in
am
b
ien
t lig
h
t d
o
n
o
t in
ter
f
er
e
with
th
e
ca
p
tu
r
e
d
v
alu
es.
T
h
ese
d
escr
ip
to
r
s
c
o
llectiv
ely
en
c
o
d
e
c
o
m
p
lem
e
n
tar
y
asp
ec
ts
o
f
v
is
u
al
co
n
ten
t
,
in
clu
d
in
g
lo
ca
l
s
tr
u
ctu
r
e,
g
lo
b
al
f
r
e
q
u
en
c
y
,
co
l
o
r
d
is
tr
ib
u
tio
n
,
an
d
tex
tu
r
e
p
atter
n
s
.
3
.
3
.
F
us
io
n a
nd
re
-
ra
nk
ing
f
ra
m
e
wo
r
k
T
h
e
ce
n
tr
al
m
ec
h
an
is
m
to
t
h
is
s
y
s
tem
i
s
a
late
f
u
s
io
n
tech
n
iq
u
e
wh
er
e
b
y
d
ata
f
r
o
m
m
u
ltip
le
d
escr
ip
to
r
s
co
n
tr
ib
u
te
to
war
d
im
p
r
o
v
i
n
g
ac
cu
r
ac
y
ass
o
ciate
d
with
r
etr
iev
ed
im
a
g
es.
-
W
eig
h
ted
v
o
tin
g
is
t
h
e
b
asis
f
o
r
t
h
e
im
p
lem
en
tatio
n
o
f
f
ea
tu
r
e
-
b
ased
d
escr
ip
to
r
s
(
B
o
VW
,
SW
T
,
DC
T
,
PC
A)
,
wh
er
e
f
o
u
r
f
ea
tu
r
es
ar
e
o
r
ig
in
ally
r
an
k
e
d
to
p
r
o
d
u
c
e
in
itial
li
s
ts
o
f
ten
r
an
k
in
g
im
ag
es
(
N=
1
0
)
.
T
h
ese
lis
ts
ar
e
m
er
g
ed
i
n
to
a
ca
n
d
id
ate
p
o
o
l.
E
ac
h
im
ag
e
i
s
ass
ig
n
ed
a
s
co
r
e
b
ased
o
n
weig
h
ted
v
o
tin
g
,
wh
er
e
weig
h
ts
ar
e
d
er
iv
ed
f
r
o
m
th
e
em
p
ir
ical
p
er
f
o
r
m
an
ce
(
m
AP,
to
p
-
1
0
)
o
f
ea
ch
d
escr
ip
t
o
r
(
e.
g
.
,
SW
T
=
0
.
4
3
,
PC
A
=
0
.
1
7
)
,
a
n
d
th
e
f
in
al
s
co
r
e
is
co
m
p
u
ted
as th
e
s
u
m
o
f
co
n
tr
ib
u
tin
g
weig
h
ts
.
-
T
o
f
u
r
th
er
im
p
r
o
v
e
p
r
ec
is
io
n
,
a
s
ec
o
n
d
-
s
tag
e
r
e
-
r
an
k
in
g
is
p
er
f
o
r
m
ed
u
s
in
g
a
co
m
b
in
ed
d
escr
ip
to
r
co
n
s
is
tin
g
o
f
L
B
P
,
HOG
,
an
d
SVD
.
E
ac
h
d
escr
ip
to
r
f
ac
ilit
ates
th
e
id
en
tific
atio
n
o
f
lo
ca
l
tex
tu
r
e
(
L
B
P),
s
h
ap
e
(
HOG)
,
an
d
s
tr
u
ctu
r
e
(
SVD)
ch
ar
ac
ter
is
tics
.
C
o
s
in
e
s
im
ilar
i
ty
is
co
m
p
u
ted
b
etwe
en
th
e
co
n
ca
ten
ated
L
B
P
-
HOG
-
SVD
d
escr
ip
to
r
s
o
f
th
e
q
u
e
r
y
i
m
ag
e
an
d
ca
n
d
id
ate
im
ag
es.
Fin
al
r
an
k
in
g
is
d
eter
m
in
ed
u
s
in
g
a
two
-
lev
el
cr
iter
io
n
:
(
i
)
d
escen
d
in
g
wei
g
h
ted
-
v
o
te
s
co
r
e
an
d
(
ii
)
asce
n
d
in
g
s
im
ilar
ity
d
is
tan
ce
.
T
h
is
en
s
u
r
es
th
at
i
m
ag
es
co
n
s
is
ten
tly
s
elec
ted
b
y
m
u
ltip
le
d
escr
ip
to
r
s
an
d
ex
h
ib
itin
g
h
ig
h
v
is
u
al
s
im
ilar
ity
ar
e
p
r
io
r
itized
.
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
is
s
ec
tio
n
ex
p
lain
s
t
h
e
v
is
u
al
o
u
tco
m
es
o
f
t
h
e
p
r
o
p
o
s
ed
s
y
s
tem
with
th
e
p
e
r
f
o
r
m
an
ce
r
esu
lts
o
f
th
e
s
y
s
tem
as
illu
s
tr
ated
in
T
ab
le
2
,
c
o
m
p
ar
is
o
n
with
th
e
r
elate
d
s
y
s
tem
s
,
an
ab
latio
n
s
t
u
d
y
,
a
n
aly
s
is
o
f
th
e
r
esu
lts
,
ad
v
an
tag
es o
f
t
h
e
p
r
o
p
o
s
ed
s
y
s
tem
,
lim
itatio
n
s
o
f
th
e
s
y
s
tem
,
an
d
f
in
ally
,
th
e
s
u
g
g
e
s
ted
r
elate
d
wo
r
k
.
T
ab
le
2
.
R
esu
lts
o
f
th
e
p
r
o
p
o
s
ed
s
y
s
tem
C
l
a
s
s
I
n
p
u
t
i
m
a
g
e
R
e
t
r
i
e
v
a
l
i
ma
g
e
s
Ef
f
i
c
i
e
n
c
y
met
r
i
c
s
S
t
r
e
e
t
P
r
e
c
i
s
i
o
n
o
f
6
:
1
.
0
0
S
e
a
F
i
n
a
l
p
r
e
c
i
s
i
o
n
o
f
5
:
0
.
8
0
M
o
u
n
t
a
i
n
F
i
n
a
l
p
r
e
c
i
s
i
o
n
o
f
8
:
0
.
7
5
G
l
a
c
i
e
r
F
i
n
a
l
p
r
e
c
i
s
i
o
n
o
f
7
:
1
.
0
0
F
o
r
e
st
F
i
n
a
l
p
r
e
c
i
s
i
o
n
o
f
9
:
1
.
0
0
S
t
r
e
e
t
P
r
e
c
i
s
i
o
n
o
f
6
:
1
.
0
0
B
u
i
l
d
i
n
g
s
F
i
n
a
l
p
r
e
c
i
s
i
o
n
o
f
5
:
1
.
0
0
T
h
e
p
r
o
p
o
s
e
d
h
i
e
r
a
r
c
h
i
c
a
l
f
u
s
i
o
n
s
y
s
t
e
m
d
e
m
o
n
s
t
r
a
t
e
d
h
i
g
h
l
y
c
o
m
p
e
t
i
t
i
v
e
p
e
r
f
o
r
m
a
n
c
e
,
a
c
h
i
e
v
i
n
g
a
m
e
a
n
a
v
e
r
a
g
e
p
r
e
c
i
s
i
o
n
f
o
r
t
h
e
t
o
p
1
0
r
e
t
r
i
e
v
e
d
i
m
a
g
e
s
o
f
0
.
8
8
o
n
t
h
e
i
n
t
e
l
i
m
a
g
e
c
l
a
s
s
i
f
i
c
a
t
i
o
n
d
a
t
a
s
e
t
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
A
mu
lti
-
ex
p
ert a
p
p
r
o
a
c
h
to
c
o
n
ten
t
-
b
a
s
ed
im
a
g
e
r
etri
ev
a
l
… (
A
li A
b
d
u
l
a
z
ee
z
Mo
h
a
mme
d
B
a
q
er
Qa
z
z
a
z
)
1381
T
h
e
p
e
r
f
o
r
m
a
n
c
e
o
f
e
a
c
h
o
f
t
h
e
f
i
v
e
c
o
r
e
f
e
a
t
u
r
e
d
e
s
c
r
i
p
t
o
r
s
w
i
l
l
b
e
e
v
a
l
u
a
t
e
d
.
T
h
i
s
i
s
o
l
a
t
e
s
t
h
e
i
n
h
e
r
e
n
t
d
i
s
c
r
i
m
i
n
a
t
i
v
e
p
o
w
e
r
o
f
e
a
c
h
f
e
a
t
u
r
e
.
T
h
e
r
e
s
u
l
t
s
,
p
r
e
s
e
n
t
e
d
i
n
T
a
b
l
e
3
,
e
s
t
a
b
l
i
s
h
a
p
e
r
f
o
r
m
a
n
c
e
b
a
s
e
l
i
n
e
f
o
r
e
a
c
h
m
e
t
h
o
d
a
n
d
i
n
d
i
c
a
t
e
a
r
o
b
u
s
t
a
b
i
l
i
t
y
t
o
r
e
t
r
i
e
v
e
s
e
m
a
n
t
i
c
a
l
l
y
r
e
l
e
v
a
n
t
i
m
a
g
e
s
c
o
n
s
i
s
t
e
n
t
l
y
w
i
t
h
i
n
t
h
e
t
o
p
r
e
s
u
l
t
s
.
T
ab
le
3
.
Per
f
o
r
m
an
ce
o
f
in
d
iv
id
u
al
f
ea
tu
r
e
d
escr
ip
to
r
s
F
e
a
t
u
r
e
d
e
scri
p
t
o
r
R
o
l
e
i
n
s
y
st
e
m
mA
P
,
t
o
p
-
1
0
(
I
n
d
i
v
i
d
u
a
l
p
e
r
f
o
r
m
a
n
c
e
)
B
o
V
W
B
r
o
a
d
f
i
l
t
e
r
i
n
g
0
.
7
2
SWT
B
r
o
a
d
f
i
l
t
e
r
i
n
g
0
.
6
8
D
C
T
B
r
o
a
d
f
i
l
t
e
r
i
n
g
0
.
6
5
P
C
A
o
n
c
o
l
o
r
h
i
st
o
g
r
a
ms
B
r
o
a
d
f
i
l
t
e
r
i
n
g
0
.
7
9
S
V
D
F
i
n
e
-
g
r
a
i
n
e
d
r
a
n
k
i
n
g
0
.
7
6
LB
P
F
i
n
e
-
g
r
a
i
n
e
d
r
a
n
k
i
n
g
0
.
6
3
4
.
1
.
Co
m
pa
riso
n
I
n
o
r
d
e
r
to
ap
p
r
o
p
r
iately
lo
ca
t
e
th
e
p
r
o
p
o
s
ed
m
o
d
el
with
in
t
h
e
cu
r
r
en
t
s
tate
o
f
ac
tiv
ity
tak
in
g
p
lace
u
tili
zin
g
C
B
I
R
s
y
s
tem
s
,
a
co
m
p
ar
is
o
n
was
p
er
f
o
r
m
ed
ag
a
in
s
t
o
th
er
c
u
r
r
en
tl
y
av
ailab
le
tech
n
iq
u
es
ac
tiv
e
with
in
th
is
ar
ea
em
p
lo
y
i
n
g
v
a
r
io
u
s
m
o
d
el
ar
c
h
itectu
r
e
ty
p
es a
s
s
ee
n
in
T
ab
le
4
.
T
ab
le
4
.
C
o
m
p
a
r
ativ
e
an
aly
s
is
o
f
r
etr
iev
al
p
er
f
o
r
m
an
ce
with
s
tate
-
of
-
th
e
-
ar
t m
eth
o
d
s
#
M
e
t
h
o
d
o
l
o
g
y
s
u
mm
a
r
y
D
a
t
a
s
e
t
B
e
st
r
e
s
u
l
t
s
[
5
]
D
u
a
l
-
C
N
N
e
n
c
o
d
e
r
s
o
n
e
n
h
a
n
c
e
d
LD
P
i
ma
g
e
s
.
C
o
r
e
l
-
1K
A
v
e
r
a
g
e
p
r
e
c
i
s
i
o
n
:
9
4
.
5
%
[
6
]
D
e
e
p
C
N
N
f
e
a
t
u
r
e
s
w
i
t
h
D
D
R
a
n
d
l
o
c
V
LA
D
a
g
g
r
e
g
a
t
i
o
n
.
P
a
r
i
s6
K
mA
P
:
0
.
7
1
8
8
[
8
]
F
i
n
e
-
t
u
n
e
d
R
e
sN
e
t
-
5
0
w
i
t
h
si
m
i
l
a
r
i
t
y
mat
c
h
i
n
g
.
C
u
s
t
o
m
(
2
,
1
0
0
)
P
r
e
c
i
s
i
o
n
:
9
0
.
1
8
%
[
1
0
]
H
y
b
r
i
d
mo
d
e
l
:
Ef
f
i
c
i
e
n
t
N
e
t
w
i
t
h
H
O
G
a
n
d
co
-
a
t
t
e
n
t
i
o
n
.
C
o
r
e
l
-
1K
mA
P
:
0
.
8
9
P
r
o
p
o
se
d
s
y
st
e
m
H
i
e
r
a
r
c
h
i
c
a
l
fu
s
i
o
n
:
(
B
o
V
W
+
S
W
T+
D
C
T+
L
B
P
)
a
n
d
(
P
C
A
+
S
V
D
)
I
n
t
e
l
i
ma
g
e
mA
P
,
t
o
p
-
1
0
:
0
.
8
8
4.
2
.
Ana
ly
s
is
o
f
c
o
m
p
a
riso
n
T
h
e
r
esu
lts
d
em
o
n
s
tr
ate
th
at
th
e
p
r
o
p
o
s
ed
m
u
lti
-
ex
p
er
t
f
r
am
ewo
r
k
ac
h
iev
es
a
s
tr
o
n
g
b
alan
ce
b
etwe
en
ac
cu
r
ac
y
,
ef
f
icien
cy
,
an
d
i
n
ter
p
r
etab
ilit
y
.
T
h
e
p
r
o
p
o
s
ed
m
u
lti
-
ex
p
er
t
f
r
am
ew
o
r
k
allo
ws
f
o
r
th
e
g
en
er
atio
n
o
f
ca
n
d
id
ate
im
a
g
es
f
r
o
m
b
r
o
a
d
d
escr
ip
to
r
s
(
e.
g
.
,
B
o
VW
,
S
W
T
,
DC
T
)
,
an
d
th
e
r
ef
in
em
en
t
o
f
im
ag
e
r
an
k
in
g
f
r
o
m
f
in
e
-
g
r
ain
ed
d
escr
ip
to
r
s
(
e.
g
.
,
L
B
P
,
HOG
,
SVD
)
.
A
m
u
lti
-
ex
p
er
t
C
B
I
R
s
y
s
tem
b
a
s
ed
o
n
h
ier
ar
ch
ical
f
u
s
io
n
a
n
d
s
u
b
-
r
a
n
k
in
g
allo
ws
f
o
r
an
ef
f
ec
tiv
e
m
ea
n
s
b
y
wh
ich
t
o
ca
p
t
u
r
e
c
o
m
p
lem
en
tar
y
v
is
u
al
f
ea
tu
r
es.
Alth
o
u
g
h
h
y
b
r
i
d
an
d
d
ee
p
m
o
d
els
m
a
y
ac
h
ie
v
e
s
lig
h
tly
h
ig
h
er
ac
cu
r
ac
y
in
s
o
m
e
ca
s
es
[
5
]
,
[
1
0
]
,
th
eir
p
er
f
o
r
m
an
ce
d
ep
e
n
d
s
o
n
h
ig
h
co
m
p
u
tatio
n
al
r
eso
u
r
ce
s
an
d
tr
ain
in
g
d
ata.
4.
3
.
Abla
t
io
n study
T
o
ass
ess
t
h
e
c
o
n
tr
ib
u
t
io
n
o
f
v
a
r
i
o
u
s
i
n
d
iv
id
u
al
c
o
m
p
o
n
e
n
ts
,
as
well
as
t
o
ass
ess
e
f
f
ic
a
cy
as
f
u
l
l,
we
co
n
d
u
ct
ed
a
f
u
l
l
ab
lat
io
n
s
tu
d
y
u
t
ili
zi
n
g
t
h
e
t
h
e
i
m
a
g
e
cl
ass
i
f
ic
ati
o
n
d
a
tas
et
f
r
o
m
t
h
e
i
n
t
el
d
atas
et
.
All
e
v
al
u
ati
o
n
s
we
r
e
p
e
r
f
o
r
m
e
d
u
s
in
g
t
h
e
m
e
an
av
e
r
a
g
e
p
r
e
cisi
o
n
f
o
r
to
p
1
0
r
e
tr
ie
v
e
d
i
m
a
g
es
m
et
r
i
cs.
T
h
e
r
es
u
lts
ar
e
b
r
ie
f
l
y
s
u
m
m
a
r
iz
e
d
i
n
T
a
b
le
4
,
wi
th
f
in
e
-
g
r
a
in
e
d
r
a
n
k
i
n
g
f
ea
tu
r
es
(
PC
A
a
n
d
SVD
)
h
a
v
i
n
g
t
h
e
b
est
ac
c
u
r
ac
y
p
e
r
f
o
r
m
a
n
c
e
o
n
th
ei
r
o
wn
,
w
h
i
c
h
c
o
n
f
ir
m
s
t
h
eir
u
s
e
i
n
a
f
in
al
r
e
f
i
n
e
m
e
n
t
s
tep
.
T
a
b
l
e
5
s
h
o
ws
in
d
i
v
i
d
u
al
c
o
n
t
r
i
b
u
ti
o
n
s
a
n
d
i
n
c
r
e
m
e
n
t
al
i
m
p
r
o
v
e
m
e
n
ts
a
t
e
ac
h
h
i
e
r
a
r
c
h
y
le
v
el
,
w
h
i
ch
s
h
o
ws
t
h
e
c
u
m
u
la
ti
v
e
ch
an
g
es
f
r
o
m
ar
c
h
ite
ct
u
r
al
d
e
cisi
o
n
s
o
f
s
i
ze
a
n
d
s
c
al
e
(
o
b
je
cti
v
el
y
b
r
o
a
d
-
f
ilt
er
i
n
g
m
o
d
els
to
a
f
i
n
al
c
o
n
s
e
n
s
u
s
s
tep
)
th
at
ac
h
i
ev
e
o
p
ti
m
al
r
etr
i
ev
al
p
e
r
f
o
r
m
a
n
c
e.
T
h
e
r
es
u
l
ts
y
ie
ld
tw
o
i
m
p
o
r
t
a
n
t
o
b
s
e
r
v
ati
o
n
s
:
F
ir
s
t
,
as
s
h
o
wn
in
T
ab
le
3
,
n
o
s
in
g
l
e
d
esc
r
i
p
to
r
p
er
f
o
r
m
s
we
ll
e
n
o
u
g
h
as
a
s
t
an
d
-
al
o
n
e
t
o
p
i
c
d
esc
r
i
p
t
o
r
–
e
v
e
n
t
h
e
b
est
in
d
i
v
i
d
u
al
f
ea
tu
r
e,
PC
A,
s
t
ill
p
er
f
o
r
m
s
s
ig
n
i
f
i
ca
n
tl
y
w
o
r
s
e
t
h
an
t
h
e
e
n
t
ir
e
s
y
s
t
em
;
S
ec
o
n
d
,
t
h
e
r
es
u
lts
h
i
g
h
li
g
h
t
a
c
lea
r
h
ie
r
a
r
c
h
y
o
f
s
t
r
e
n
g
t
h
wit
h
c
o
m
b
i
n
i
n
g
t
h
r
e
e
m
et
h
o
d
s
f
o
r
b
r
o
a
d
-
f
ilt
er
in
g
t
o
b
u
il
d
an
i
n
it
ial
p
o
o
l
o
f
ca
n
d
id
ates
,
w
h
i
c
h
h
a
d
t
h
e
l
ar
g
es
t
p
er
f
o
r
m
an
ce
in
cr
ea
s
e
(
+
0
.
0
9
m
AP
)
j
u
s
ti
f
y
i
n
g
t
h
at
th
ese
f
e
at
u
r
es
a
r
e
ca
p
t
u
r
in
g
v
is
u
a
l
c
u
es
t
h
at
a
r
e
co
m
p
le
m
e
n
t
a
r
y
to
e
ac
h
o
t
h
e
r
.
Af
te
r
t
h
e
c
o
n
s
tr
u
cte
d
p
o
o
l
o
f
ca
n
d
i
d
a
tes
was
f
ilt
e
r
e
d
a
n
d
th
e
c
an
d
i
d
ates
w
e
r
e
f
i
n
e
-
g
r
ai
n
e
d
r
e
-
r
a
n
k
e
d
b
ase
d
o
n
p
air
in
g
PC
A
,
t
h
is
p
e
r
f
o
r
m
e
d
b
ett
er
b
y
a
n
o
t
h
e
r
+0
.
0
4
m
AP,
t
o
th
e
p
ea
k
p
er
f
o
r
m
a
n
c
e
o
f
0
.
8
8
f
r
o
m
a
r
a
n
k
c
o
n
s
en
s
u
s
b
etw
ee
n
PC
A
a
n
d
SVD
r
an
k
i
n
g
s
.
4
.
4
.
L
im
it
a
t
io
ns
Desp
ite
at
tai
n
i
n
g
n
o
ta
b
l
y
h
i
g
h
r
e
tr
ie
v
al
p
r
ec
is
io
n
in
o
u
r
p
r
o
p
o
s
e
d
s
y
s
te
m
,
e
x
is
ti
n
g
li
m
it
at
io
n
s
e
x
is
t.
First
,
f
u
s
i
o
n
wei
g
h
ts
a
m
o
n
g
d
esc
r
i
p
t
o
r
s
a
r
e
b
as
e
d
o
n
e
m
p
ir
ica
l
r
at
h
e
r
th
an
ad
a
p
ti
v
e
o
p
ti
m
iz
ati
o
n
,
lik
el
y
c
o
n
s
tr
ai
n
i
n
g
g
e
n
e
r
aliz
at
io
n
to
u
n
s
ee
n
d
at
asets
.
Se
co
n
d
,
c
o
m
b
i
n
ati
o
n
s
o
f
h
a
n
d
-
c
r
af
t
ed
f
ea
tu
r
es
m
a
y
n
o
t
ca
p
t
u
r
e
s
e
m
a
n
ti
c
v
a
r
ia
n
ce
o
r
h
i
g
h
-
l
e
v
el
co
n
t
e
x
t
u
al
cu
es
as
well
as
d
ja
n
g
o
-
s
t
y
l
e
o
r
o
t
h
e
r
d
e
ep
-
l
ea
r
n
i
n
g
b
as
e
d
task
s
.
T
h
i
r
d
,
t
h
e
p
r
o
p
o
s
e
d
f
r
am
ew
o
r
k
r
el
ies
u
p
o
n
f
i
x
e
d
f
ea
t
u
r
e
d
i
m
e
n
s
i
o
n
ali
ty
f
o
r
e
v
e
r
y
d
esc
r
i
p
t
o
r
,
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
15
,
No
.
3
,
Sep
tem
b
er
20
26
:
1
3
7
6
-
1
3
8
4
1382
wh
i
ch
m
a
y
li
m
i
t
s
c
ala
b
ilit
y
to
m
u
c
h
la
r
g
er
i
m
a
g
e
d
at
asets
.
Fin
all
y
,
w
h
il
e
t
h
e
a
r
c
h
i
te
ct
u
r
e
is
t
r
a
in
in
g
f
r
e
e,
it
is
in
c
a
p
a
b
le
o
f
o
n
l
in
e
a
d
a
p
t
a
b
ili
t
y
-
an
d
s
l
ig
h
t
s
h
i
f
ts
i
n
d
at
ase
t
d
o
m
ai
n
,
o
r
v
is
u
al
c
o
n
d
iti
o
n
s
,
m
a
y
r
e
q
u
ir
e
m
an
u
a
l
f
e
at
u
r
es
n
o
r
m
a
liz
ati
o
n
a
n
d
d
es
cr
i
p
t
o
r
wei
g
h
t
u
p
d
a
tes
.
T
ab
le
5
.
I
n
cr
em
en
tal
p
er
f
o
r
m
a
n
ce
o
f
th
e
h
ier
ar
c
h
ical
s
y
s
tem
S
y
st
e
m c
o
n
f
i
g
u
r
a
t
i
o
n
D
e
scri
p
t
i
o
n
mA
P
,
t
o
p
-
10
P
e
r
f
o
r
ma
n
c
e
g
a
i
n
S
t
a
g
e
1
o
n
l
y
R
e
s
u
l
t
s fr
o
m
t
h
e
b
e
st
si
n
g
l
e
b
r
o
a
d
-
f
i
l
t
e
r
(
B
o
V
W
)
.
0
.
7
2
(
B
a
se
l
i
n
e
)
S
t
a
g
e
1
+
p
o
o
l
M
e
r
g
e
d
c
a
n
d
i
d
a
t
e
p
o
o
l
f
r
o
m
a
l
l
t
h
r
e
e
b
r
o
a
d
-
f
i
l
t
e
r
s
(
B
o
V
W
+
S
W
T
+
D
C
T)
.
0
.
8
1
+
0
.
0
9
S
t
a
g
e
1
+
R
e
-
r
a
n
k
(
P
C
A
o
n
l
y
)
C
a
n
d
i
d
a
t
e
p
o
o
l
r
e
-
r
a
n
k
e
d
u
s
i
n
g
o
n
l
y
t
h
e
P
C
A
f
e
a
t
u
r
e
.
0
.
8
5
+
0
.
0
4
P
r
o
p
o
se
d
s
y
st
e
m
F
i
n
a
l
h
i
e
r
a
r
c
h
i
c
a
l
m
o
d
e
l
w
i
t
h
c
o
n
s
e
n
s
u
s (P
C
A
∩
S
V
D
)
.
0
.
8
8
+
0
.
0
3
4
.
5
.
F
uture
wo
r
k
s
Fu
tu
r
e
wo
r
k
will
p
lace
an
em
p
h
asis
o
n
ex
ten
d
in
g
th
e
p
r
o
p
o
s
ed
f
r
am
ewo
r
k
to
f
ac
ilit
ate
g
r
ea
ter
f
lex
ib
ilit
y
an
d
s
ca
lab
ilit
y
.
A
p
o
ten
tial
way
f
o
r
war
d
is
to
cr
e
ate
an
ad
a
p
tiv
e
weig
h
tin
g
ap
p
r
o
ac
h
to
ad
ju
s
t
th
e
co
n
tr
ib
u
tio
n
s
o
f
ev
e
r
y
d
escr
ip
to
r
b
ased
o
n
th
e
in
d
i
v
id
u
al
q
u
er
y
im
ag
es.
As
well,
an
o
th
er
d
ir
ec
tio
n
o
f
f
u
tu
r
e
wo
r
k
will
b
e
to
ev
al
u
ate
wh
eth
er
o
r
n
o
t
lig
h
tweig
h
t
d
ee
p
f
ea
tu
r
es
an
d
/
o
r
s
elf
-
s
u
p
er
v
i
s
ed
r
ep
r
esen
tatio
n
s
co
u
ld
b
e
b
r
o
u
g
h
t
in
to
th
e
h
y
b
r
id
f
u
s
io
n
p
r
o
ce
s
s
in
an
in
ter
p
r
etab
le
way
.
I
n
ad
d
itio
n
,
a
th
ir
d
f
u
tu
r
e
d
ir
ec
ti
o
n
is
to
p
u
r
s
u
e
r
ea
l
-
tim
e
in
te
g
r
atio
n
o
n
ed
g
e
d
ev
ices
an
d
d
e
m
o
n
s
tr
atio
n
s
o
v
er
a
v
ar
iety
o
f
C
B
I
R
b
en
ch
m
ar
k
s
to
s
h
o
w
u
tili
ty
ab
ilit
y
in
r
ea
l
-
wo
r
ld
s
ce
n
ar
io
s
.
Fu
tu
r
e
f
r
a
m
ewo
r
k
s
also
m
ay
i
n
clu
d
e
m
ec
h
an
i
s
m
s
th
at
allo
w
f
o
r
u
s
er
s
to
co
n
tr
o
l
a
n
y
th
in
g
f
r
o
m
g
u
id
ed
to
in
ter
ac
tiv
e
r
etr
iev
al
o
f
im
ag
es
a
n
d
r
ef
i
n
em
en
t.
Fr
o
m
a
n
im
p
lem
en
tatio
n
p
er
s
p
ec
tiv
e,
u
tili
zin
g
FP
GA
ar
ch
itectu
r
es
to
class
if
y
h
ar
d
war
e
ac
ce
ler
atio
n
h
o
ld
s
p
r
o
m
is
e
f
o
r
en
ab
lin
g
r
ea
l
-
tim
e
f
ea
tu
r
es
an
d
p
r
o
ce
s
s
in
g
as
h
ig
h
lig
h
t
ed
in
s
o
m
e
d
o
m
ain
s
o
f
r
ec
en
t
n
eu
r
al
s
y
s
tem
s
id
en
tific
atio
n
wo
r
k
[
2
4
]
.
Als
o
,
co
s
t
-
ef
f
ec
tiv
e
p
ar
ad
i
g
m
s
th
at
u
tili
ze
lear
n
in
g
at
th
e
e
d
g
e
wo
u
ld
ass
is
t
in
ex
ten
d
in
g
C
B
I
R
ap
p
licatio
n
s
i
n
ed
u
ca
tio
n
,
o
r
em
b
ed
d
ed
s
ettin
g
s
[
2
5
]
.
5.
CO
NCLU
SI
O
N
T
h
is
p
ap
er
p
r
o
p
o
s
ed
a
m
u
l
ti
-
ex
p
er
t
f
r
am
ewo
r
k
f
o
r
C
B
I
R
b
ased
o
n
h
ier
ar
ch
ical
f
u
s
io
n
an
d
re
-
r
an
k
i
n
g
o
f
s
ev
e
n
h
an
d
-
cr
af
t
ed
d
e
s
cr
ip
to
r
s
(
B
o
VW
,
SW
T
,
DC
T
,
P
C
A,
L
B
P,
HOG,
an
d
SVD)
.
E
v
alu
atio
n
s
in
d
icate
d
th
at
s
o
m
e
in
d
iv
id
u
al
d
escr
ip
to
r
s
y
ield
ed
m
AP
(
to
p
-
1
0
)
s
co
r
es
r
an
g
in
g
f
r
o
m
0
.
6
3
to
0
.
7
9
,
with
PC
A
as
th
e
b
est
s
in
g
le
d
escr
ip
t
o
r
.
T
h
e
p
r
o
p
o
s
ed
f
u
s
ed
s
y
s
tem
r
e
tu
r
n
ed
an
m
AP
s
co
r
e
o
f
0
.
8
8
.
T
h
e
p
r
o
p
o
s
ed
s
tu
d
y
o
f
f
er
s
a
co
m
p
etitiv
e
lev
el
o
f
p
r
ec
is
io
n
co
m
p
ar
ed
with
s
tate
-
of
-
th
e
-
ar
t
d
ee
p
a
n
d
h
y
b
r
id
C
B
I
R
m
o
d
els
wh
ile
r
em
ain
in
g
tr
an
s
p
a
r
en
t,
co
s
t
-
e
f
f
icien
t,
an
d
in
d
e
p
en
d
e
n
t
o
f
lar
g
e
-
s
ca
le
tr
ain
in
g
d
ata.
T
h
e
p
r
ec
ed
in
g
r
esu
lts
s
u
b
s
tan
tiate
th
e
ab
ilit
y
o
f
t
r
ain
in
g
-
f
r
ee
,
i
n
ter
p
r
etab
le
r
e
tr
iev
al
s
y
s
tem
s
to
b
e
u
tili
ze
d
in
co
n
s
tr
ain
ed
en
v
ir
o
n
m
en
ts
,
s
u
ch
as e
m
b
e
d
d
ed
o
r
m
o
b
ile
p
latf
o
r
m
s
.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
T
h
e
a
u
th
o
r
s
s
tate
n
o
f
u
n
d
in
g
was
in
v
o
lv
ed
.
AUTHO
R
CO
NT
RI
B
UT
I
O
NS ST
A
T
E
M
E
N
T
T
h
is
jo
u
r
n
al
u
s
es
th
e
C
o
n
tr
ib
u
to
r
R
o
les
T
ax
o
n
o
m
y
(
C
R
ed
iT)
to
r
ec
o
g
n
ize
in
d
iv
id
u
al
au
th
o
r
co
n
tr
ib
u
tio
n
s
,
r
ed
u
ce
au
th
o
r
s
h
ip
d
is
p
u
tes,
an
d
f
ac
ilit
ate
co
llab
o
r
atio
n
.
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
Ali A
b
d
u
laze
ez
Mo
h
am
m
ed
B
aq
er
Qazza
z
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Yo
u
s
if
Sam
er
Mu
d
h
af
ar
✓
✓
✓
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
A
mu
lti
-
ex
p
ert a
p
p
r
o
a
c
h
to
c
o
n
ten
t
-
b
a
s
ed
im
a
g
e
r
etri
ev
a
l
… (
A
li A
b
d
u
l
a
z
ee
z
Mo
h
a
mme
d
B
a
q
er
Qa
z
z
a
z
)
1383
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
Au
th
o
r
s
s
tate
n
o
co
n
f
lict o
f
in
t
er
est.
DATA AV
AI
L
AB
I
L
I
T
Y
No
t a
p
p
licab
le,
as n
o
n
ew
d
ata
wer
e
g
en
er
ate
d
o
r
a
n
aly
ze
d
i
n
th
is
s
tu
d
y
.
RE
F
E
R
E
NC
E
S
[
1
]
M
.
S
.
S
a
y
e
d
,
A
.
A
.
A
.
G
a
d
-
El
r
a
b
,
K
.
A
.
F
a
t
h
y
,
a
n
d
K
.
R
.
R
a
sl
a
n
,
“
U
n
su
p
e
r
v
i
s
e
d
c
o
n
t
e
n
t
b
a
s
e
d
i
ma
g
e
r
e
t
r
i
e
v
a
l
u
si
n
g
p
r
e
-
t
r
a
i
n
e
d
C
N
N
a
n
d
P
C
N
N
f
e
a
t
u
r
e
s
e
x
t
r
a
c
t
o
r
s,”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
I
n
t
e
l
l
i
g
e
n
t
En
g
i
n
e
e
ri
n
g
a
n
d
S
y
st
e
m
s
,
v
o
l
.
1
6
,
n
o
.
1
,
p
p
.
5
8
4
–
5
9
6
,
F
e
b
.
2
0
2
3
,
d
o
i
:
1
0
.
2
2
2
6
6
/
i
j
i
e
s
2
0
2
3
.
0
2
2
8
.
5
0
.
[
2
]
S
.
R
.
D
u
b
e
y
,
“
A
d
e
c
a
d
e
s
u
r
v
e
y
o
f
c
o
n
t
e
n
t
b
a
se
d
i
ma
g
e
r
e
t
r
i
e
v
a
l
u
s
i
n
g
d
e
e
p
l
e
a
r
n
i
n
g
,
”
I
E
EE
T
ra
n
s
a
c
t
i
o
n
s
o
n
C
i
r
c
u
i
t
s a
n
d
S
y
st
e
m
s
f
o
r
Vi
d
e
o
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
3
2
,
n
o
.
5
,
p
p
.
2
6
8
7
–
2
7
0
4
,
M
a
y
2
0
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
T
C
S
V
T.
2
0
2
1
.
3
0
8
0
9
2
0
.
[
3
]
V
.
K
.
I
n
d
u
,
M
.
C
h
a
r
m
i
l
a
,
A
.
S
w
a
r
o
o
p
,
a
n
d
C
.
L
.
S
u
c
h
a
r
i
t
h
a
,
“
C
o
n
t
e
n
t
-
b
a
se
d
i
m
a
g
e
r
e
t
r
i
e
v
a
l
s
y
s
t
e
m
u
s
i
n
g
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
C
r
e
a
t
i
v
e
Re
s
e
a
r
c
h
T
h
o
u
g
h
t
s
,
v
o
l
.
1
3
,
n
o
.
3
,
p
p
.
5
5
6
–
5
6
3
,
2
0
2
5
.
[
4
]
S
.
M
.
C
h
a
v
d
a
a
n
d
M
.
M
.
G
o
y
a
n
i
,
“
R
o
b
u
st
c
o
n
t
e
n
t
-
b
a
s
e
d
i
ma
g
e
r
e
t
r
i
e
v
a
l
u
s
i
n
g
I
C
C
V
,
G
L
C
M
,
a
n
d
D
W
T
-
M
S
LB
P
d
e
s
c
r
i
p
t
o
r
s,
”
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
2
3
,
n
o
.
1
,
M
a
r
.
2
0
2
2
,
d
o
i
:
1
0
.
7
4
9
4
/
c
sc
i
.
2
0
2
2
.
2
3
.
1
.
3
8
2
1
.
[
5
]
C
.
P
a
l
a
i
,
P
.
K
.
Je
n
a
,
S
.
R
.
P
a
t
t
a
n
a
i
k
,
T.
P
a
n
i
g
r
a
h
i
,
a
n
d
T.
K
.
M
i
sh
r
a
,
“
C
o
n
t
e
n
t
-
b
a
se
d
i
m
a
g
e
r
e
t
r
i
e
v
a
l
u
s
i
n
g
e
n
c
o
d
e
r
b
a
se
d
R
G
B
a
n
d
t
e
x
t
u
r
e
f
e
a
t
u
r
e
f
u
s
i
o
n
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
A
d
v
a
n
c
e
d
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
1
4
,
n
o
.
3
,
2
0
2
3
,
d
o
i
:
1
0
.
1
4
5
6
9
/
I
JA
C
S
A
.
2
0
2
3
.
0
1
4
0
3
2
8
.
[
6
]
F
.
M
a
g
l
i
a
n
i
,
T.
F
o
n
t
a
n
i
n
i
,
a
n
d
A
.
P
r
a
t
i
,
“
A
d
e
n
s
e
-
d
e
p
t
h
r
e
p
r
e
se
n
t
a
t
i
o
n
f
o
r
V
LA
D
d
e
s
c
r
i
p
t
o
r
s
i
n
c
o
n
t
e
n
t
-
b
a
se
d
i
mag
e
r
e
t
r
i
e
v
a
l
,
”
i
n
L
e
c
t
u
re
N
o
t
e
s
i
n
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
(
i
n
c
l
u
d
i
n
g
s
u
b
ser
i
e
s
L
e
c
t
u
r
e
N
o
t
e
s
i
n
Art
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
a
n
d
L
e
c
t
u
re
N
o
t
e
s
i
n
Bi
o
i
n
f
o
rm
a
t
i
c
s)
,
v
o
l
.
1
1
2
4
1
LN
C
S
,
2
0
1
8
,
p
p
.
6
6
2
–
6
7
1
.
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
0
3
0
-
038
01
-
4
_
5
8
.
[
7
]
V
.
G
a
y
a
t
h
r
i
,
“
A
n
a
n
a
l
y
s
i
s
o
f
c
o
n
t
e
n
t
-
b
a
s
e
d
i
ma
g
e
r
e
t
r
i
e
v
a
l
sy
s
t
e
m
,
”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
R
e
se
a
rc
h
i
n
Ad
v
a
n
c
e
d
E
n
g
i
n
e
e
r
i
n
g
a
n
d
T
e
c
h
n
o
l
o
g
,
v
o
l
.
9
,
n
o
.
4
,
p
p
.
1
9
–
2
3
,
2
0
2
3
.
[
8
]
G
.
G
a
u
t
a
m
a
n
d
A
.
K
h
a
n
n
a
,
“
C
o
n
t
e
n
t
b
a
s
e
d
i
m
a
g
e
r
e
t
r
i
e
v
a
l
s
y
st
e
m
u
s
i
n
g
C
N
N
b
a
s
e
d
d
e
e
p
l
e
a
r
n
i
n
g
m
o
d
e
l
s
,
”
Pr
o
c
e
d
i
a
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
2
3
5
,
p
p
.
3
1
3
1
–
3
1
4
1
,
2
0
2
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
p
r
o
c
s.
2
0
2
4
.
0
4
.
2
9
6
.
[
9
]
A
.
G
h
i
t
a
a
n
d
R
.
I
o
n
e
s
c
u
,
“
C
l
a
ss
a
n
c
h
o
r
mar
g
i
n
l
o
ss
f
o
r
c
o
n
t
e
n
t
-
b
a
se
d
i
ma
g
e
r
e
t
r
i
e
v
a
l
,
”
i
n
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
1
6
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
A
g
e
n
t
s
a
n
d
Ar
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
,
S
C
I
TEPRE
S
S
-
S
c
i
e
n
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
P
u
b
l
i
c
a
t
i
o
n
s
,
2
0
2
4
,
p
p
.
8
4
8
–
853
,
d
o
i
:
1
0
.
5
2
2
0
/
0
0
1
2
4
0
0
5
0
0
0
0
3
6
3
6
.
[
1
0
]
R
.
K
u
m
a
r
a
n
d
N
.
M
.
M
S
,
“
R
e
l
e
v
a
n
c
e
-
a
w
a
r
e
c
o
n
t
e
n
t
-
b
a
se
d
i
m
a
g
e
r
e
t
r
i
e
v
a
l
u
s
i
n
g
d
e
e
p
h
y
b
r
i
d
f
e
a
t
u
r
e
e
x
t
r
a
c
t
i
o
n
,
”
E
n
g
i
n
e
e
ri
n
g
,
T
e
c
h
n
o
l
o
g
y
&
A
p
p
l
i
e
d
S
c
i
e
n
c
e
Re
s
e
a
rch
,
v
o
l
.
1
5
,
n
o
.
3
,
p
p
.
2
2
9
7
6
–
2
2
9
8
2
,
Ju
n
.
2
0
2
5
,
d
o
i
:
1
0
.
4
8
0
8
4
/
e
t
a
sr
.
1
0
7
6
7
.
[
1
1
]
A
.
M
a
h
a
j
a
n
,
N
.
R
a
n
e
,
a
n
d
P
.
P
a
t
i
l
,
“
C
o
n
t
e
n
t
-
b
a
se
d
i
ma
g
e
r
e
t
r
i
e
v
a
l
s
y
st
e
m
u
t
i
l
i
z
i
n
g
v
i
si
o
n
t
r
a
n
sf
o
r
mer,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
f
o
r
Re
se
a
rc
h
T
r
e
n
d
s
a
n
d
I
n
n
o
v
a
t
i
o
n
,
v
o
l
.
1
0
,
n
o
.
2
,
p
p
.
7
8
6
–
7
9
0
,
2
0
2
5
.
[
1
2
]
C
.
Zh
a
n
g
a
n
d
J.
Li
u
,
“
C
o
n
t
e
n
t
b
a
se
d
d
e
e
p
l
e
a
r
n
i
n
g
i
m
a
g
e
r
e
t
r
i
e
v
a
l
:
A
s
u
r
v
e
y
,
”
i
n
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
2
0
2
3
9
t
h
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
C
o
m
m
u
n
i
c
a
t
i
o
n
a
n
d
I
n
f
o
rm
a
t
i
o
n
Pr
o
c
e
ssi
n
g
,
N
e
w
Y
o
r
k
,
N
Y
,
U
S
A
:
A
C
M
,
D
e
c
.
2
0
2
3
,
p
p
.
1
5
8
–
1
6
3
,
d
o
i
:
1
0
.
1
1
4
5
/
3
6
3
8
8
8
4
.
3
6
3
8
9
0
8
.
[
1
3
]
D
.
S
r
i
v
a
st
a
v
a
,
S
.
S
.
S
i
n
g
h
,
B
.
R
a
j
i
t
h
a
,
M
.
V
e
r
m
a
,
M
.
K
a
u
r
,
a
n
d
H
.
-
N
.
Le
e
,
“
C
o
n
t
e
n
t
-
b
a
s
e
d
i
m
a
g
e
r
e
t
r
i
e
v
a
l
:
A
s
u
r
v
e
y
o
n
l
o
c
a
l
a
n
d
g
l
o
b
a
l
f
e
a
t
u
r
e
s
se
l
e
c
t
i
o
n
,
e
x
t
r
a
c
t
i
o
n
,
r
e
p
r
e
s
e
n
t
a
t
i
o
n
,
a
n
d
e
v
a
l
u
a
t
i
o
n
p
a
r
a
m
e
t
e
r
s,
”
I
EE
E
A
c
c
e
ss
,
v
o
l
.
1
1
,
p
p
.
9
5
4
1
0
–
9
5
4
3
1
,
2
0
2
3
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
3
.
3
3
0
8
9
1
1
.
[
1
4
]
A
.
D
o
n
g
,
Y
.
Zh
a
n
g
,
W
.
Li
,
a
n
d
M
.
C
h
e
n
,
“
A
w
e
i
g
h
t
e
d
b
a
g
o
f
v
i
s
u
a
l
w
o
r
d
s
mo
d
e
l
f
o
r
p
r
e
d
i
c
t
i
n
g
f
e
t
a
l
g
r
o
w
t
h
r
e
st
r
i
c
t
i
o
n
a
t
a
n
e
a
r
l
y
st
a
g
e
,
”
Fr
o
n
t
i
e
rs
i
n
M
e
d
i
c
i
n
e
,
v
o
l
.
1
2
,
Ju
n
.
2
0
2
5
,
d
o
i
:
1
0
.
3
3
8
9
/
f
me
d
.
2
0
2
5
.
1
5
2
9
6
6
6
.
[
1
5
]
P
o
o
j
a
a
n
d
M
.
A
r
o
r
a
,
“
T
e
x
t
a
n
d
i
ma
g
e
c
l
a
ss
i
f
i
c
a
t
i
o
n
u
si
n
g
sh
a
p
e
c
o
n
t
e
x
t
a
n
d
b
a
g
o
f
v
i
su
a
l
w
o
r
d
s
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
S
c
i
e
n
t
i
f
i
c
Re
se
a
r
c
h
& E
n
g
i
n
e
e
ri
n
g
T
r
e
n
d
s
,
v
o
l
.
1
0
,
n
o
.
3
,
p
p
.
7
8
5
–
7
8
9
,
2
0
2
4
.
[
1
6
]
R
.
A
.
A
smar
a
,
R
.
A
g
u
st
i
n
a
,
a
n
d
H
i
d
a
y
a
t
u
l
l
o
h
,
“
C
o
m
p
a
r
i
so
n
o
f
d
i
s
c
r
e
t
e
c
o
s
i
n
e
t
r
a
n
sf
o
r
ms
(
D
C
T)
,
d
i
scr
e
t
e
F
o
u
r
i
e
r
t
r
a
n
sf
o
r
m
s
(
D
F
T)
,
a
n
d
d
i
scr
e
t
e
w
a
v
e
l
e
t
t
r
a
n
sf
o
r
ms
(
D
W
T)
i
n
d
i
g
i
t
a
l
i
ma
g
e
w
a
t
e
r
m
a
r
k
i
n
g
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
A
d
v
a
n
c
e
d
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
8
,
n
o
.
2
,
2
0
1
7
,
d
o
i
:
1
0
.
1
4
5
6
9
/
I
JA
C
S
A
.
2
0
1
7
.
0
8
0
2
3
2
.
[
1
7
]
S
.
A
.
B
.
P
a
c
h
e
c
o
,
M
.
G
o
y
a
n
i
,
Z
.
G
.
R
e
h
ma
n
,
S
.
F
.
R
e
h
ma
n
,
T.
C
h
a
m
p
a
n
e
r
i
a
,
a
n
d
S
.
G
o
y
a
n
i
,
“
E
n
h
a
n
c
e
d
c
o
n
t
e
n
t
-
b
a
s
e
d
i
m
a
g
e
r
e
t
r
i
e
v
a
l
u
si
n
g
mu
l
t
i
v
i
s
u
a
l
f
e
a
t
u
r
e
s
f
u
si
o
n
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
C
o
m
p
u
t
e
rs
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
4
7
,
n
o
.
1
0
,
p
p
.
8
3
5
–
8
5
6
,
2
0
2
5
,
d
o
i
:
1
0
.
1
0
8
0
/
1
2
0
6
2
1
2
X
.
2
0
2
5
.
2
5
3
6
5
7
7
.
[
1
8
]
S
.
C
h
a
v
d
a
a
n
d
M
.
G
o
y
a
n
i
,
“
M
u
l
t
i
-
l
a
b
e
l
r
e
m
o
t
e
se
n
si
n
g
sc
e
n
e
c
l
a
ss
i
f
i
c
a
t
i
o
n
u
si
n
g
t
w
o
-
l
e
v
e
l
d
o
u
b
l
e
c
h
a
n
n
e
l
sp
a
t
i
a
l
a
t
t
e
n
t
i
o
n
r
e
si
d
u
a
l
b
l
o
c
k
s
,
”
J
o
u
r
n
a
l
o
f
A
p
p
l
i
e
d
R
e
m
o
t
e
S
e
n
si
n
g
,
v
o
l
.
1
8
,
n
o
.
3
,
p
.
0
3
6
5
1
1
,
S
e
p
.
2
0
2
4
,
d
o
i
:
1
0
.
1
1
1
7
/
1
.
J
R
S
.
1
8
.
0
3
6
5
1
1
.
[
1
9
]
A
.
A
.
M
.
B
.
Q
a
z
z
a
z
a
n
d
N
.
E
.
K
a
d
h
i
m,
“
W
a
t
e
r
mark
b
a
se
d
o
n
s
i
n
g
u
l
a
r
v
a
l
u
e
d
e
c
o
mp
o
s
i
t
i
o
n
,
”
B
a
g
h
d
a
d
S
c
i
e
n
c
e
J
o
u
r
n
a
l
,
v
o
l
.
2
0
,
n
o
.
5
,
p
p
.
1
7
9
7
–
1
8
0
7
,
F
e
b
.
2
0
2
3
,
d
o
i
:
1
0
.
2
1
1
2
3
/
b
s
j
.
2
0
2
3
.
7
1
6
8
.
[
2
0
]
J.
D
.
S
o
l
e
r
e
t
a
l
.
,
“
H
i
st
o
g
r
a
m
o
f
o
r
i
e
n
t
e
d
g
r
a
d
i
e
n
t
s:
a
t
e
c
h
n
i
q
u
e
f
o
r
t
h
e
st
u
d
y
o
f
m
o
l
e
c
u
l
a
r
c
l
o
u
d
f
o
r
ma
t
i
o
n
,
”
A
st
r
o
n
o
m
y
&
Ast
r
o
p
h
y
s
i
c
s
,
v
o
l
.
6
2
2
,
p
.
A
1
6
6
,
F
e
b
.
2
0
1
9
,
d
o
i
:
1
0
.
1
0
5
1
/
0
0
0
4
-
6
3
6
1
/
2
0
1
8
3
4
3
0
0
.
[
2
1
]
C.
-
T.
Y
e
n
,
J
.
-
X
.
Li
a
o
,
a
n
d
Y
.
-
K
.
H
u
a
n
g
,
“
E
v
a
l
u
a
t
i
n
g
f
e
a
t
u
r
e
f
u
s
i
o
n
t
e
c
h
n
i
q
u
e
s
w
i
t
h
d
e
e
p
l
e
a
r
n
i
n
g
mo
d
e
l
s
f
o
r
c
o
r
o
n
a
v
i
r
u
s
d
i
s
e
a
s
e
2
0
1
9
c
h
e
st
X
-
r
a
y
s
e
n
s
o
r
i
ma
g
e
i
d
e
n
t
i
f
i
c
a
t
i
o
n
,
”
S
e
n
so
rs
a
n
d
M
a
t
e
ri
a
l
s
,
v
o
l
.
3
6
,
n
o
.
2
,
p
p
.
6
8
3
–
6
9
9
,
F
e
b
.
2
0
2
4
,
d
o
i
:
1
0
.
1
8
4
9
4
/
S
A
M
4
6
8
5
.
[
2
2
]
T.
A
.
A
l
-
A
sad
i
a
n
d
A
.
A
.
A
.
M
.
B
a
q
e
r
,
“
F
u
si
o
n
f
o
r
mu
l
t
i
p
l
e
l
i
g
h
t
s
o
u
r
c
e
s
i
n
t
e
x
t
u
r
e
ma
p
p
i
n
g
o
b
j
e
c
t
,
”
J
o
u
r
n
a
l
o
f
T
e
l
e
c
o
m
m
u
n
i
c
a
t
i
o
n
,
E
l
e
c
t
r
o
n
i
c
a
n
d
C
o
m
p
u
t
e
r E
n
g
i
n
e
e
ri
n
g
,
v
o
l
.
9
,
n
o
.
2
–
1
1
,
p
p
.
7
–
1
2
,
2
0
1
7
.
[
2
3
]
P
.
B
a
n
sa
l
,
“
I
n
t
e
l
i
ma
g
e
c
l
a
ssi
f
i
c
a
t
i
o
n
,
”
K
a
g
g
l
e
.
A
c
c
e
ss
e
d
:
A
p
r
.
1
8
,
2
0
2
6
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
w
w
w
.
k
a
g
g
l
e
.
c
o
m
/
d
a
t
a
se
t
s/
p
u
n
e
e
t
6
0
6
0
/
i
n
t
e
l
-
i
m
a
g
e
-
c
l
a
ssi
f
i
c
a
t
i
o
n
[
2
4
]
S
.
H
.
A
b
d
u
l
n
a
b
i
,
Y
.
S
.
M
u
d
h
a
f
a
r
,
A
.
A
.
K
a
d
h
i
m,
M
.
B
.
M
a
h
d
i
,
a
n
d
H
.
H
.
S
o
j
a
r
,
“
N
e
u
r
a
l
n
e
t
w
o
r
k
-
b
a
se
d
s
y
st
e
m
i
d
e
n
t
i
f
i
c
a
t
i
o
n
:
a
c
o
mp
r
e
h
e
n
s
i
v
e
F
P
G
A
d
e
s
i
g
n
a
n
d
i
mp
l
e
m
e
n
t
a
t
i
o
n
,
”
i
n
2
0
2
4
I
EEE
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
Art
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
a
n
d
Me
c
h
a
t
r
o
n
i
c
s
S
y
st
e
m
s (AI
M
S
)
,
I
EEE,
F
e
b
.
2
0
2
4
,
p
p
.
1
–
7
.
d
o
i
:
1
0
.
1
1
0
9
/
A
I
M
S
6
1
8
1
2
.
2
0
2
4
.
1
0
5
1
2
5
3
1
.
[
2
5
]
A
.
M
.
A
.
A
l
-
m
u
q
a
r
m,
Y
.
M
u
d
h
a
f
a
r
,
A
.
M
.
S
h
a
k
i
r
,
M
.
K
a
z
e
m
,
R
.
A
b
d
e
l
-
Y
a
h
i
y
a
,
a
n
d
B
.
S
.
A
.
Za
h
r
a
,
“
Lo
w
-
c
o
st
sm
a
r
t
l
e
a
r
n
i
n
g
w
i
t
h
M
o
o
d
l
e
-
b
a
se
d
R
a
sp
b
e
r
r
y
P
i
4
f
o
r
u
n
i
v
e
r
s
i
t
y
st
u
d
e
n
t
s
,
”
i
n
2
0
2
3
6
t
h
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
E
n
g
i
n
e
e
r
i
n
g
T
e
c
h
n
o
l
o
g
y
a
n
d
i
t
s
Ap
p
l
i
c
a
t
i
o
n
s
(
I
I
C
ETA)
,
I
EEE,
Ju
l
.
2
0
2
3
,
p
p
.
6
0
3
–
6
0
8
,
d
o
i
:
1
0
.
1
1
0
9
/
I
I
C
E
TA
5
7
6
1
3
.
2
0
2
3
.
1
0
3
5
1
2
6
6
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
15
,
No
.
3
,
Sep
tem
b
er
20
26
:
1
3
7
6
-
1
3
8
4
1384
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
Ali
Abd
u
l
a
z
e
e
z
Mo
h
a
m
m
e
d
B
a
q
e
r
Q
a
z
z
a
z
re
c
e
iv
e
d
th
e
M
.
S
c
.
a
n
d
P
h
.
D.
d
e
g
re
e
s
in
Co
m
p
u
ter
S
c
ien
c
e
fro
m
Ba
b
y
lo
n
Un
iv
e
rsity
,
Ira
q
,
in
2
0
1
2
a
n
d
2
0
1
8
,
re
sp
e
c
ti
v
e
ly
.
He
wo
rk
e
d
a
s
a
l
e
c
tu
re
r
a
t
t
h
e
Un
iv
e
rsit
y
o
f
Ku
fa
,
C
o
ll
e
g
e
o
f
Ed
u
c
a
ti
o
n
,
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter
S
c
ien
c
e
.
His
re
se
a
r
c
h
in
tere
sts
in
c
lu
d
e
ima
g
e
p
r
o
c
e
ss
in
g
,
c
o
m
p
u
ter
v
isi
o
n
,
in
fo
rm
a
ti
o
n
se
c
u
rit
y
,
d
e
e
p
lea
rn
in
g
,
AI
,
a
n
d
d
a
ta
m
in
i
n
g
.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
a
li
a
.
q
a
z
z
a
z
@u
o
k
u
fa
.
e
d
u
.
i
q
.
Yo
u
sif
S
a
m
e
r
Mu
d
h
a
f
a
r
e
a
rn
e
d
h
is
B.
S
c
.
i
n
Co
m
p
u
ter
Tec
h
n
iq
u
e
s
En
g
i
n
e
e
rin
g
fro
m
th
e
Isla
m
ic
Un
iv
e
rsity
in
Na
jaf
in
2
0
1
8
.
He
c
o
m
p
lete
d
h
is
M
.
S
c
.
in
Co
m
p
u
ter
S
c
ien
c
e
En
g
i
n
e
e
rin
g
a
t
th
e
U
n
iv
e
rsit
y
o
f
De
b
re
c
e
n
in
2
0
2
2
,
g
ra
d
u
a
ti
n
g
wi
th
h
o
n
o
rs
a
n
d
re
c
e
iv
in
g
t
h
e
Ou
tstan
d
i
n
g
S
t
u
d
e
n
t
c
e
r
ti
fica
te.
He
wo
rk
s
a
t
t
h
e
U
n
iv
e
rsit
y
o
f
Ku
fa
,
F
a
c
u
l
ty
o
f
E
d
u
c
a
ti
o
n
,
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter
S
c
ie
n
c
e
.
His
re
se
a
rc
h
i
n
tere
sts
in
c
l
u
d
e
c
o
m
p
u
ter
n
e
two
r
k
s,
I
o
T
,
a
n
d
AI
.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
y
o
u
sifs
.
m
u
d
h
a
fa
r@u
o
k
u
fa
.
e
d
u
.
i
q
.
Evaluation Warning : The document was created with Spire.PDF for Python.