I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
p
ute
r
E
ng
in
ee
ring
(
I
J
E
CE
)
Vo
l.
7
,
No
.
5
,
Octo
b
e
r
2
0
1
7
,
p
p
.
2
5
3
7
~
2
5
4
6
I
SS
N:
2
0
8
8
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v7
i
1
.
p
p
2
5
3
7
-
2546
2537
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ia
e
s
jo
u
r
n
a
l.c
o
m/o
n
lin
e/in
d
ex
.
p
h
p
/I
JE
C
E
India
n Clas
sica
l
Da
nce Mudra
C
la
ss
ificatio
n Using
H
O
G
Features an
d
SV
M
Clas
sifier
K
.
V.
V.
K
u
m
a
r
1
,
P
.
V.
V.
K
is
h
o
re
2
De
p
a
rtme
n
t
o
f
El
e
c
tro
n
ics
a
n
d
C
o
m
m
u
n
ica
ti
o
n
s E
n
g
i
n
e
e
rin
g
,
K.L
.
Un
iv
e
rsity
,
In
d
ia
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Dec
9
,
2
0
1
6
R
ev
i
s
ed
J
u
n
1
7
,
2
0
1
7
A
cc
ep
ted
Sep
1
7
,
2
0
1
7
Dig
it
a
l
u
n
d
e
rsta
n
d
in
g
o
f
In
d
ia
n
c
las
sic
a
l
d
a
n
c
e
is
lea
st
stu
d
ied
w
o
rk
,
th
o
u
g
h
it
h
a
s
b
e
e
n
a
p
a
rt
o
f
In
d
ia
n
C
u
lt
u
re
f
ro
m
a
ro
u
n
d
2
0
0
BC.
T
h
is
w
o
r
k
e
x
p
lo
re
s
th
e
p
o
ss
ib
il
it
ies
o
f
re
c
o
g
n
izin
g
c
las
sic
a
l
d
a
n
c
e
m
u
d
ra
s
in
v
a
r
io
u
s
d
a
n
c
e
f
o
r
m
s
in
In
d
ia.
T
h
e
im
a
g
e
s
o
f
h
a
n
d
m
u
d
ra
s
o
f
v
a
rio
u
s
c
las
sic
a
l
d
a
n
c
e
s
a
re
c
o
ll
e
c
ted
f
o
r
m
th
e
in
tern
e
t
a
n
d
a
d
a
t
a
b
a
se
is
c
re
a
t
e
d
f
o
r
th
is
jo
b
.
Histo
g
ra
m
o
f
o
rien
ted
(HO
G
)
fe
a
tu
re
s
o
f
h
a
n
d
m
u
d
ra
s
in
p
u
t
th
e
c
las
sif
ie
r.
S
u
p
p
o
r
t
v
e
c
to
r
m
a
c
h
in
e
(S
V
M
)
c
las
sif
ies
th
e
HO
G
f
e
a
tu
re
s
in
to
m
u
d
r
a
s
a
s
tex
t
m
e
ss
a
g
e
s.
T
h
e
m
u
d
ra
re
c
o
g
n
it
io
n
f
re
q
u
e
n
c
y
(M
RF
)
is
c
a
lcu
late
d
f
o
r
e
a
c
h
mu
d
ra
u
sin
g
g
ra
p
h
ica
l
u
se
r
in
t
e
rfa
c
e
(
G
UI)
d
e
v
e
lo
p
e
d
f
ro
m
th
e
m
o
d
e
l.
P
o
p
u
lar
f
e
a
tu
re
v
e
c
to
rs
su
c
h
a
s
S
IF
T
,
S
URF,
L
BP
a
n
d
HA
A
R
a
re
tes
ted
a
g
a
in
st
HO
G
f
o
r
p
re
c
isio
n
a
n
d
s
w
i
f
tn
e
ss
.
T
h
is
w
o
rk
h
e
lp
s
n
e
w
lea
rn
e
rs
a
n
d
d
a
n
c
e
e
n
th
u
sia
stic
p
e
o
p
le
to
lea
rn
a
n
d
u
n
d
e
rsta
n
d
d
a
n
c
e
f
o
rm
s
a
n
d
re
late
d
in
f
o
rm
a
ti
o
n
o
n
th
e
ir
m
o
b
il
e
d
e
v
ice
s.
K
ey
w
o
r
d
:
I
n
d
ian
clas
s
ical
d
an
ce
m
u
d
r
as
HOG
f
ea
t
u
r
es
Mu
d
r
a
r
ec
o
g
n
itio
n
f
r
eq
u
e
n
c
y
Scale
in
v
ar
ian
t f
ea
t
u
r
es
SVM
c
lass
if
ier
Co
p
y
rig
h
t
©
2
0
1
7
In
stit
u
te o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
rin
g
a
n
d
S
c
ien
c
e
.
Al
l
rig
h
ts
re
se
rv
e
d
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Dr
.
P
.
V.
V.
Kis
h
o
r
e,
Dep
ar
t
m
en
t o
f
E
lectr
o
n
ics a
n
d
C
o
m
m
u
n
icat
io
n
s
E
n
g
in
ee
r
i
n
g
,
Mu
lti
m
ed
ia
an
d
Vi
s
io
n
C
o
m
p
u
ti
n
g
R
esear
c
h
C
e
n
ter
,
K
L
U
n
iv
er
s
it
y
,
Gr
ee
n
Field
s
,
Vad
d
es
w
ar
a
m
,
Gu
n
tu
r
DT
,
An
d
h
r
a
P
r
ad
esh
,
I
n
d
ia
-
522502.
E
m
ail:
p
v
v
k
is
h
o
r
e@
k
l
u
n
i
v
er
s
i
t
y
.
in
1.
I
NT
RO
D
UCT
I
O
N
I
n
d
ian
d
an
ce
f
o
r
m
s
ar
e
a
m
ir
r
o
r
to
r
ich
cu
ltu
r
al
h
er
ita
g
e
t
h
a
t
ex
is
ted
f
r
o
m
t
h
e
p
ast
5
0
0
0
y
ea
r
s
.
T
h
e
n
a
m
e
f
o
r
th
ese
cla
s
s
ical
d
a
n
c
e
f
o
r
m
s
i
s
ca
lled
‘
Nat
y
a
R
asa
’
as
p
o
r
tr
ay
ed
i
n
th
e
b
ib
le
o
f
I
n
d
ian
d
an
ce
‘
Nat
y
a
Sh
as
tr
a’
.
A
cc
o
r
d
in
g
to
Nat
y
a
Sh
as
tr
a
th
er
e
ar
e
1
0
8
k
ar
an
as
[
1
]
m
ea
n
in
g
ac
tio
n
o
f
h
a
n
d
s
,
f
ee
t a
n
d
b
o
d
y
.
T
h
ese
p
o
s
es
s
y
m
b
o
lize
v
ar
io
u
s
p
h
y
s
ical
m
ea
n
in
g
r
elate
d
to
n
atu
r
e,
g
o
d
an
d
ac
tio
n
s
.
A
f
e
w
h
asta
m
u
d
r
as
ar
e
s
h
o
w
n
in
Fi
g
u
r
e
1
f
o
r
r
ef
er
en
ce
[
2
]
.
Fig
u
r
e.
1
.
Hasta
m
u
d
r
as o
f
i
n
d
ian
class
ical
d
an
ce
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
J
E
C
E
Vo
l.
7
,
No
.
5
,
Octo
b
er
2
0
1
7
:
2
5
3
7
–
2
5
4
6
2538
In
th
i
s
w
o
r
k
w
e
s
ee
k
to
class
i
f
y
h
a
s
ta
m
u
d
r
as
u
s
ed
in
v
ar
io
u
s
I
n
d
ian
cla
s
s
ical
d
an
ce
f
o
r
m
s
.
C
r
ea
tio
n
o
f
d
ataset
f
o
r
th
is
ta
s
k
is
a
co
m
p
lex
p
r
o
ce
s
s
.
Fo
r
th
e
in
it
ial
p
h
ase
o
f
test
i
n
g
,
w
e
w
o
r
k
o
n
i
m
ag
e
s
av
ai
lab
le
o
n
in
ter
n
e
t
d
an
ce
s
ites
.
A
ca
r
ef
u
l
s
ec
tio
n
o
f
p
r
ec
is
el
y
ca
p
tu
r
ed
i
m
a
g
es
ar
e
g
r
o
u
p
ed
to
g
eth
er
t
o
f
o
r
m
o
u
r
tr
ain
in
g
an
d
test
i
n
g
d
ataset
s
f
o
r
th
e
cl
ass
i
f
ier
.
Ou
r
d
ataset
is
h
a
v
i
n
g
5
d
if
f
er
en
t
s
ets
o
f
2
4
h
asta
m
u
d
r
as
f
r
o
m
v
ar
io
u
s
d
an
ce
f
o
r
m
s
.
B
aile
y
,
H.
et.
al
co
n
s
id
er
s
th
e
r
o
le
an
d
i
m
p
ac
t
o
f
n
e
w
a
n
d
em
er
g
i
n
g
e
-
Scien
ce
to
o
ls
o
n
p
r
ac
tice
-
led
r
esear
ch
in
d
an
ce
.
Sp
ec
i
f
icall
y
,
it
d
r
a
w
s
o
n
f
i
n
d
i
n
g
s
f
r
o
m
t
h
e
e
-
Dan
ce
p
r
o
j
ec
t.
T
h
is
t
w
o
-
y
ea
r
p
r
o
j
ec
t
b
r
in
g
s
to
g
eth
er
an
i
n
ter
d
is
cip
li
n
ar
y
tea
m
co
m
b
i
n
i
n
g
r
esear
c
h
asp
ec
ts
o
f
ch
o
r
eo
g
r
ap
h
y
,
n
e
x
t
g
en
er
atio
n
o
f
v
id
eo
co
n
f
er
en
c
in
g
,
an
d
H
u
m
an
-
c
o
m
p
u
ter
i
n
ter
ac
tio
n
a
n
al
y
s
is
i
n
co
r
p
o
r
atin
g
h
y
p
er
m
ed
ia
an
d
n
o
n
-
l
in
ea
r
an
n
o
tatio
n
s
f
o
r
r
ec
o
r
d
in
g
an
d
d
o
cu
m
en
tatio
n
[
3
],
[
4
].
T
h
e
k
e
y
f
ea
t
u
r
e
o
f
t
h
e
e
-
Dan
ce
p
r
o
j
ec
t
is
th
e
cr
ea
tiv
e
an
d
cr
itical
en
g
ag
e
m
e
n
t
w
it
h
e
-
Sc
ien
ce
an
d
s
p
ec
if
icall
y
A
G.
I
t
f
o
cu
s
e
s
o
n
th
e
Gr
id
i
n
ter
m
s
o
f
it
s
v
is
u
al
co
m
m
u
n
icatio
n
al
ca
p
ac
it
y
,
an
d
f
r
o
m
a
n
ar
ts
p
er
s
p
ec
tiv
e
it
is
p
ar
ticu
lar
l
y
c
o
n
ce
r
n
ed
w
i
th
m
ea
n
in
g
p
r
o
d
u
ctio
n
in
t
h
is
v
is
u
al,
tele
-
co
m
m
u
n
ica
tio
n
al
co
n
te
x
t.
T
h
is
f
o
cu
s
o
n
m
ea
n
i
n
g
b
r
id
g
es
th
e
d
is
cip
li
n
ar
y
d
i
v
id
e
th
r
o
u
g
h
u
s
er
i
n
ter
f
ac
e
d
es
ig
n
an
d
‘
s
en
s
e
m
a
k
i
n
g
’
[
5
]
o
n
th
e
e
-
Sc
ien
ce
s
id
e,
an
d
s
p
ec
tato
r
/p
ar
ticip
an
t
en
g
a
g
e
m
en
t
an
d
in
ter
p
r
etatio
n
f
r
o
m
t
h
e
p
er
s
p
ec
tiv
e
o
f
th
e
ar
ts
[
6
].
T
h
e
r
eg
u
lar
i
m
a
g
e
s
eg
m
e
n
tati
o
n
alg
o
r
ith
m
s
s
u
c
h
as
t
h
r
es
h
o
ld
in
g
[
7
]
,
ed
g
e
[
8
]
an
d
c
o
lo
r
clu
s
ter
i
n
g
[
9
]
f
ail
to
ex
tr
ac
t
f
u
ll
f
ea
t
u
r
es
o
f
th
e
m
u
d
r
as.
T
h
is
i
s
d
u
e
to
o
cc
lu
s
io
n
s
o
f
f
in
g
er
s
d
u
r
in
g
ca
p
tu
r
e
an
d
co
lo
u
r
i
n
g
u
s
ed
f
o
r
f
in
g
er
s
d
u
r
i
n
g
p
er
f
o
r
m
an
ce
s
.
Gen
er
al
f
ea
t
u
r
e
ex
tr
a
ctio
n
m
o
d
els
i
n
liter
at
u
r
e
ar
e
u
s
ed
to
r
ep
r
esen
t
th
e
d
an
ce
m
u
d
r
as.
T
h
es
e
f
ea
t
u
r
es
ar
e
h
i
s
to
g
r
a
m
o
f
o
r
ien
ted
g
r
a
d
ien
ts
(
H
OG)
,
s
p
ee
d
u
p
r
o
b
u
s
t
f
ea
t
u
r
es
(
SUR
F),
s
ca
le
in
v
ar
ian
t f
ea
t
u
r
e
tr
an
s
f
o
r
m
(
SIFT
)
,
lo
ca
l b
in
ar
y
p
atter
n
s
(
L
B
P
)
an
d
h
aa
r
w
a
v
elet
f
ea
t
u
r
es
(
H
AAR).
A
co
m
p
ar
is
o
n
o
f
t
h
e
s
e
f
ea
tu
r
es
o
n
i
m
a
g
es
o
f
d
a
n
ce
m
u
d
r
as
f
r
o
m
w
e
b
a
n
d
s
el
f
-
ca
p
tu
r
ed
d
ata
is
p
er
f
o
r
m
ed
.
Su
p
p
o
r
t
v
ec
to
r
m
a
ch
in
e
s
(
S
VM
)
ar
e
u
s
ed
as
cla
s
s
i
f
ier
o
f
t
h
ese
f
ea
t
u
r
es
to
id
e
n
ti
f
y
a
m
u
d
r
a
c
lass
in
k
u
ch
ip
u
d
i
d
an
ce
m
u
d
r
as.
W
e
f
o
u
n
d
t
h
at
t
h
e
s
e
2
4
m
u
d
r
as
ar
e
t
h
e
b
as
is
f
o
r
all
8
clas
s
ical
d
an
ce
f
o
r
m
s
i
n
I
n
d
ia.
Hen
ce
,
class
if
ica
tio
n
i
n
K
u
c
h
ip
u
d
i c
an
b
e
ex
te
n
d
ed
to
o
th
er
d
an
ce
f
o
r
m
s
as
w
ell.
T
h
e
id
ea
is
to
r
ep
r
esen
t
I
n
d
ian
class
ical
d
an
ce
o
n
a
d
ig
ital
p
latf
o
r
m
.
I
n
d
ian
c
u
lt
u
r
al
d
an
c
e
f
o
r
m
s
ar
e
m
o
s
t
co
m
p
le
x
h
u
m
a
n
g
es
tu
r
es
to
b
e
r
ep
r
esen
ted
in
d
ig
i
tal
f
o
r
m
at.
Fea
tu
r
e
e
x
tr
ac
tio
n
i
s
m
o
s
t c
o
m
p
l
icate
d
tas
k
as
t
h
e
i
m
ag
e
s
ar
e
f
u
ll
o
f
co
lo
r
,
o
cc
lu
s
io
n
s
a
n
d
f
in
g
er
clo
s
e
n
ess
.
T
h
i
s
ca
n
b
e
o
b
s
er
v
ed
i
n
i
m
ag
e
s
i
n
F
ig
u
r
e
1
.
T
h
e
r
eg
u
lar
i
m
a
g
e
p
r
o
ce
s
s
i
n
g
s
eg
m
en
tatio
n
m
o
d
els
s
u
ch
a
s
th
r
es
h
o
ld
in
g
an
d
ed
g
e
d
etec
ti
o
n
f
ail
to
r
ep
r
esen
t
th
e
co
r
r
e
ct
s
h
ap
es
a
s
f
o
u
n
d
i
n
th
e
o
r
i
g
i
n
al
i
m
ag
e
s
.
T
h
is
ca
n
b
e
o
b
s
er
v
ed
in
Fi
g
u
r
e
2
f
o
r
a
m
u
d
r
a
f
r
o
m
k
u
c
h
ip
u
d
i d
an
ce
f
o
r
m
.
Fig
u
r
e
2
.
(
a)
Or
ig
in
al
m
u
d
r
a
n
a
m
ed
‘
k
an
g
u
la
’
in
k
u
c
h
ip
u
d
i (
b
)
E
d
g
e
s
eg
m
e
n
ted
an
d
(
c)
th
r
esh
o
ld
ed
i
m
a
g
e
T
h
e
m
o
s
t
v
ib
r
an
t
a
n
d
h
i
g
h
l
y
u
s
ed
m
o
d
els
o
f
s
e
g
m
en
tat
io
n
i
n
r
ec
en
t
ti
m
e
s
ar
e
ac
tiv
e
co
n
to
u
r
s
.
Ma
n
y
m
o
d
el
s
h
av
e
b
ee
n
p
r
o
p
o
s
ed
in
li
ter
atu
r
e
[
1
0
,
11
]
.
B
u
t
th
e
b
asic
m
o
d
el
s
u
f
f
er
er
s
f
r
o
m
m
an
y
d
r
a
w
b
ac
k
s
s
u
ch
as
illu
m
i
n
atio
n
,
p
o
s
itio
n
o
f
t
h
e
m
a
s
k
a
n
d
n
u
m
b
er
o
f
iter
atio
n
s
.
W
e
b
eliev
e
f
o
cu
s
ed
ac
t
iv
e
co
n
to
u
r
m
o
d
el
s
w
it
h
m
o
r
e
s
p
atial
i
n
f
o
r
m
atio
n
u
s
i
n
g
co
lo
r
,
tex
tu
r
e
an
d
s
h
ap
e
h
av
e
p
r
o
f
o
u
n
d
ef
f
ec
t
o
n
ex
t
r
ac
tin
g
t
h
e
co
r
r
ec
t
s
eg
m
e
n
ts
[
1
2
]
.
Fig
u
r
e
3
s
h
o
w
s
th
e
r
esu
l
t o
f
ac
ti
v
e
co
n
to
u
r
m
o
d
el
o
n
h
an
d
m
u
d
r
a
in
Fi
g
u
r
e
2
(
a)
.
A
n
u
m
b
er
o
f
m
et
h
o
d
s
i
n
liter
a
tu
r
e
h
elp
i
n
d
eter
m
in
e
s
h
ap
e
f
ea
tu
r
es.
T
h
is
p
ap
er
test
s
5
s
u
c
h
f
ea
t
u
r
es
an
d
p
r
o
v
id
e
an
in
d
icato
r
tellin
g
t
h
e
d
an
ce
m
u
d
r
a
class
i
f
ic
atio
n
alg
o
r
it
h
m
f
o
r
b
est
m
atc
h
in
g
s
co
r
e.
A
t
w
o
-
d
ec
ad
e
lo
n
g
ch
alle
n
g
e
f
o
r
p
r
o
d
u
cin
g
a
n
i
m
ag
i
n
g
f
ea
tu
r
e
th
at
is
i
m
m
u
n
e
to
ill
u
m
i
n
a
t
io
n
,
n
o
is
e,
s
ca
le,
o
r
ien
tatio
n
,
p
ar
tial o
cc
lu
s
io
n
s
g
iv
in
g
ac
cu
r
ac
y
an
d
s
p
ee
d
is
c
o
m
in
g
g
o
o
d
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
E
C
E
I
SS
N:
2
0
8
8
-
8708
I
n
d
ia
n
C
la
s
s
ica
l D
a
n
ce
Mu
d
r
a
C
la
s
s
ifica
tio
n
Usi
n
g
HOG
F
ea
tu
r
es a
n
d
S
V
M C
la
s
s
ifier
(
K
.
V
.
V
.
K
u
ma
r
)
2539
Fig
u
r
e
3
.
A
cti
v
e
co
n
to
u
r
b
ased
im
a
g
e
s
e
g
m
e
n
tatio
n
m
o
d
el
f
o
r
8
0
0
iter
atio
n
s
o
n
h
a
n
d
m
u
d
r
a.
T
h
e
liter
atu
r
e
h
as
SIFT
[1
3
]
,
HAAR
[1
4
]
,
Featu
r
es
f
r
o
m
ac
ce
ler
ated
s
eg
m
e
n
t
test
(
F
AST
)
[
1
5
]
,
SUR
F
[1
6
]
,
HOG
[1
7
]
an
d
m
a
n
y
m
o
r
e.
A
f
o
r
m
al
co
m
p
ar
is
o
n
o
f
t
h
ese
m
e
th
o
d
s
in
d
icate
th
at
ea
ch
o
n
e
h
as
g
o
t
th
eir
p
r
o
s
an
d
co
n
s
.
T
ab
le
1
ch
ar
ac
ter
izes
th
e
s
e
f
ea
t
u
r
es
b
ased
o
n
th
e
p
ar
am
eter
s
r
eq
u
ir
ed
f
o
r
g
o
o
d
f
ea
tu
r
e
d
escr
ip
to
r
.
T
ab
le
1
.
Featu
r
e
Descr
ip
to
r
C
h
ar
ac
ter
izatio
n
f
o
r
Dan
ce
Mu
d
r
a
C
lass
i
f
icatio
n
M
e
t
h
o
d
s
T
i
mi
n
g
T
r
a
n
sf
o
r
mat
i
o
n
S
c
a
l
i
n
g
R
o
t
a
t
i
o
n
B
l
u
r
r
i
n
g
I
l
l
u
mi
n
a
t
i
o
n
F
A
S
T
C
o
mm
o
n
C
o
mm
o
n
C
o
mm
o
n
B
a
d
B
a
d
B
a
d
S
I
F
T
B
a
d
G
o
o
d
G
o
o
d
B
e
st
B
e
st
B
e
st
S
U
R
F
G
o
o
d
B
e
st
B
e
st
G
o
o
d
G
o
o
d
G
o
o
d
HOG
G
o
o
d
B
e
st
B
e
st
B
e
st
B
e
st
B
e
st
L
B
P
good
G
o
o
d
G
o
o
d
G
o
o
d
C
o
mm
o
n
G
o
o
d
Fro
m
T
ab
le
1
,
it
ca
n
b
e
u
n
d
er
s
to
o
d
th
at
th
e
b
est
f
ea
tu
r
e
d
escr
ip
to
r
is
HOG.
T
h
er
e
ar
e
m
a
n
y
v
ar
iatio
n
s
f
o
r
HOG
s
u
c
h
as H
OG
-
L
B
P
,
HOG
-
L
S
S (
lo
ca
l sel
f
s
i
m
ilar
it
y
),
lo
ca
l g
r
ad
ien
t
(
LG
)
-
HOG
a
n
d
s
o
o
n
.
I
n
th
i
s
p
ap
er
w
e
tr
ied
all
o
f
th
e
m
a
n
d
r
esu
lt
s
m
atch
to
th
at
i
n
d
icate
d
in
T
ab
le
1
.
P
lease
r
ef
er
th
e
co
r
r
esp
o
n
d
in
g
liter
atu
r
e
f
o
r
ad
d
itio
n
al
in
f
o
r
m
atio
n
r
eg
ar
d
in
g
lo
w
lev
e
l
i
m
a
g
e
f
ea
t
u
r
e
d
escr
ip
to
r
s
in
th
e
r
ef
er
en
ce
s
p
r
o
v
id
ed
ad
j
ac
en
t to
th
e
m
ab
o
v
e.
C
h
u
n
g
-
w
ei
[
1
8
]
is
p
r
o
p
o
s
ed
m
o
v
i
n
g
o
b
j
ec
t
class
i
f
icatio
n
l
ik
es:
ca
r
s
,
m
o
to
r
c
y
cle
s
,
p
ed
es
tr
ian
s
an
d
b
icy
c
le
b
y
u
s
in
g
lo
ca
l
s
h
ap
e
a
n
d
w
a
v
elet
tr
an
s
f
o
r
m
HOG
f
e
atu
r
es
w
it
h
h
ier
ar
c
h
ical
SVM
class
i
f
icatio
n
.
T
h
e
p
r
o
p
o
s
ed
m
et
h
o
d
u
s
ed
to
test
in
s
i
x
v
id
eo
s
eq
u
e
n
ce
s
f
o
r
class
i
f
icatio
n
.
T
h
e
co
m
p
u
ter
p
r
o
ce
s
s
i
n
g
ti
m
e
s
o
f
th
e
o
b
j
ec
t
s
eg
m
e
n
tatio
n
in
7
9
m
s
,
o
b
j
ec
t
tr
ac
k
in
g
i
n
2
1
1
m
s
,
f
ea
t
u
r
e
ex
tr
ac
tio
n
a
n
d
class
i
f
icatio
n
in
0
.
0
1
m
s
r
esp
ec
tiv
el
y
.
I
n
r
ec
en
t
y
ea
r
s
,
SVM
cla
s
s
i
f
ier
w
it
h
HO
G
f
ea
tu
r
es
ar
e
t
h
e
m
o
s
t
p
o
p
u
lar
tech
n
iq
u
e
s
f
o
r
v
eh
icl
e
d
etec
tio
n
[
1
9
]
.
I
n
r
ea
l
tim
e
i
m
p
le
m
en
ta
tio
n
w
h
ich
is
i
m
p
o
r
tan
t
f
o
r
ad
v
an
ce
d
d
r
iv
er
ass
is
tan
ce
s
y
s
te
m
s
ap
p
licatio
n
s
.
T
o
r
e
d
u
ce
th
e
co
m
p
le
x
it
y
o
f
t
h
e
S
VM
is
t
o
r
ed
u
ce
th
e
d
i
m
e
n
s
io
n
s
o
f
HOG
f
ea
t
u
r
es.
T
h
e
p
r
o
p
o
s
ed
m
et
h
o
d
SVM
cla
s
s
if
ica
tio
n
f
o
r
v
e
h
icle
d
etec
ti
o
n
is
t
h
r
ee
ti
m
es
s
p
ee
d
-
u
p
in
o
th
er
d
etec
tio
n
p
er
f
o
r
m
a
n
ce
.
T
h
e
r
est
o
f
th
e
p
ap
er
is
o
r
g
an
ized
as:
s
ec
tio
n
2
d
escr
ib
es
th
e
f
o
llo
w
ed
m
et
h
o
d
o
lo
g
y
f
o
r
m
u
d
r
a
class
i
f
icatio
n
.
R
es
u
lt
s
an
d
d
is
c
u
s
s
io
n
is
p
r
ese
n
te
d
in
s
ec
tio
n
3
w
it
h
co
n
cl
u
s
io
n
s
i
n
s
ec
tio
n
4
.
2.
M
E
T
H
O
DO
L
O
G
Y
T
h
e
ex
p
er
i
m
en
t
i
n
v
o
l
v
es
o
n
l
y
d
an
ce
m
u
d
r
a
s
f
r
o
m
k
u
c
h
i
p
u
d
i
d
an
ce
f
o
r
m
as
th
e
y
ar
e
th
e
b
asic
s
tr
u
ct
u
r
es
f
o
r
f
o
r
m
atio
n
o
f
an
y
d
a
n
ce
.
Me
th
o
d
o
lo
g
y
i
n
v
o
lv
es
t
w
o
p
h
a
s
es:
t
r
ai
n
i
n
g
p
h
ase
an
d
t
esti
n
g
p
h
a
s
e.
Du
r
in
g
tr
ain
i
n
g
p
h
a
s
e
2
4
d
an
ce
m
u
d
r
as
ar
e
u
s
ed
to
th
e
tr
a
in
t
h
e
SVM
c
lass
if
ier
.
T
h
e
ca
p
ab
ilit
ies
o
f
SV
M
class
i
f
ier
ar
e
m
ap
p
ed
to
m
u
lti
p
le
class
es to
f
o
r
m
a
m
u
lti
-
c
la
s
s
SVM.
2
.
1
.
Da
t
a
s
et
f
o
r
k
uchi
pu
d
i d
a
nce
f
o
r
m
T
h
e
d
ataset
is
m
ad
e
f
r
o
m
a
co
m
b
i
n
atio
n
o
f
lab
d
an
ce
m
u
d
r
as
an
d
d
an
ce
m
u
d
r
as
i
m
a
g
es
o
n
t
h
e
w
eb
s
i
tes
o
f
I
n
d
ian
ar
t
an
d
cu
l
tu
r
e.
Fo
r
ea
c
h
m
u
d
r
a
,
w
e
h
a
v
e
m
ad
e
a
s
et
o
f
5
i
m
a
g
e
s
f
r
o
m
5
d
i
f
f
er
en
t
ar
tis
t
s
.
Fig
u
r
e
4
s
h
o
w
s
t
h
e
s
et
u
s
ed
f
o
r
ca
p
tu
r
in
g
t
h
e
d
atase
t
at
u
n
iv
er
s
it
y
m
u
l
ti
m
ed
ia
ce
n
tr
.
A
m
i
x
t
u
r
e
o
f
5
×2
4
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
J
E
C
E
Vo
l.
7
,
No
.
5
,
Octo
b
er
2
0
1
7
:
2
5
3
7
–
2
5
4
6
2540
i
m
a
g
es
i
s
u
s
ed
f
o
r
tr
ain
i
n
g
a
n
d
test
i
n
g
.
W
e
h
a
v
e
2
s
ets
o
f
i
m
ag
e
s
f
r
o
m
o
u
r
d
an
ce
r
s
,
2
s
ets
f
r
o
m
d
a
n
ce
w
eb
s
i
tes a
n
d
1
s
et
f
r
o
m
Yo
u
t
u
b
e
v
id
eo
f
r
a
m
es.
Fig
u
r
e
4
.
Setu
p
u
s
ed
f
o
r
ca
p
tu
r
in
g
d
an
ce
i
m
a
g
es o
f
k
u
ch
ip
u
d
i m
u
d
r
as
2
.
2
.
B
a
na
liza
t
io
n
o
f
i
m
a
g
es
a
nd
f
ea
t
ure
ex
t
ra
ct
io
n
P
r
o
ce
s
s
in
g
ea
s
i
n
es
s
f
o
r
f
ea
t
u
r
e
ex
tr
ac
tio
n
ca
ll
s
f
o
r
th
is
s
te
p
.
T
h
e
d
im
e
n
s
io
n
ali
t
y
is
r
ed
u
ce
d
to
r
ed
p
lan
e
an
d
lo
ca
l
m
a
x
i
m
a
ar
e
co
m
p
u
ted
.
T
h
e
lo
ca
l
m
a
x
i
m
a
i
n
a
1
6
×1
6
b
lo
ck
is
u
s
ed
as
a
th
r
es
h
o
ld
f
o
r
th
a
t
p
ar
ticu
lar
b
lo
ck
m
a
k
i
n
g
t
h
e
p
r
o
ce
s
s
in
v
ar
ian
t
to
b
r
i
g
h
t
n
e
s
s
a
n
d
co
n
tr
ast.
A
s
et
o
f
b
i
n
a
r
y
s
i
g
n
i
m
a
g
es
ar
e
co
u
p
led
in
Fi
g
u
r
e
4
in
th
e
tr
ai
n
in
g
s
id
e.
S
h
ap
e
f
ea
t
u
r
es
ar
e
m
o
d
elled
f
r
o
m
t
h
ese
b
i
n
ar
y
i
m
ag
e
s
.
F
iv
e
f
ea
tu
r
e
v
ec
to
r
s
an
d
th
e
ir
co
m
b
in
atio
n
ar
e
u
s
ed
to
ex
tr
ac
t f
ea
t
u
r
es
f
r
o
m
th
e
m
u
d
r
a
s
.
2
.
3
.
Su
pp
o
rt
v
ec
t
o
r
m
a
chine
s
SVM’
s
an
al
y
ze
d
ata
an
d
p
r
o
d
u
ce
s
b
in
ar
y
r
esp
o
n
s
e
s
f
o
r
class
i
f
icatio
n
co
m
e
u
n
d
er
a
class
o
f
s
u
p
er
v
i
s
ed
lear
n
i
n
g
.
T
h
e
b
asi
c
SVM
class
if
ie
s
o
n
l
y
t
w
o
cla
s
s
p
r
o
b
lem
s
b
y
p
r
o
j
ec
tin
g
a
h
y
p
er
p
lan
e
b
et
w
ee
n
d
ata
d
u
r
in
g
tr
ain
in
g
p
h
ase.
T
h
e
h
y
p
er
p
lan
e
i
s
ch
ar
ac
ter
iz
ed
b
y
a
s
u
b
s
et
o
f
d
ata
p
o
in
ts
ac
tin
g
as
s
u
p
p
o
r
t
v
ec
to
r
s
.
D
u
r
in
g
tr
ai
n
i
n
g
th
e
S
VM
is
p
r
ese
n
ted
w
it
h
e
x
a
m
p
l
e
v
ec
to
r
s
x
i
ϵ
ℜ
n
,
i
=
1
….
.
l
;
l
tr
ain
i
n
g
s
a
m
p
le,
to
lab
el
ea
ch
d
ata
s
am
p
le
as
eit
h
er
as
class
lab
el
+1
o
r
−
1
w
h
i
ch
f
o
r
m
s
th
e
i
n
d
icato
r
v
ec
to
r
y
i
ϵ
{+1,
−
1
}
.
SVM
f
o
r
m
u
lates t
h
e
o
p
ti
m
izatio
n
p
r
o
b
lem
a
s
a
d
ec
is
io
n
b
o
u
n
d
ar
y
D
(
x
)
s
u
c
h
th
a
t
,
.
1
1
m
i
n
2
l
T
i
wb
i
D
x
w
w
C
Su
b
j
ec
t to
1
T
i
i
i
y
w
x
b
w
i
th
0
,
1
,
2
,
.
.
.
.
,
;
i
il
(
1
)
W
h
er
e
C
is
a
p
o
s
iti
v
e
co
n
s
tan
t
d
ef
in
in
g
r
e
g
u
lar
izat
io
n
.
T
h
e
ter
m
s
w
an
d
b
ar
e
w
eig
h
t
a
n
d
b
ias.
is
th
e
m
i
s
clas
s
if
icatio
n
h
a
n
d
ler
.
T
h
e
f
u
n
ctio
n
m
(
x
)
:
x
→
ɸ
(
x
)
m
ap
s
f
ea
t
u
r
e
v
ec
to
r
x
to
a
h
ig
h
er
d
i
m
e
n
s
io
n
a
l
s
p
ac
e.
T
h
e
m
ap
p
in
g
f
u
n
cti
o
n
m
(
x
)
m
ap
s
x
in
to
a
d
o
t
p
r
o
d
u
ct
o
f
f
ea
tu
r
e
s
p
ac
e
th
at
s
atis
f
ies
m
(
x
i
-
1
,
x
i
)
=
ɸ
T
(
x
i
-
1
)
ɸ
(
x
i
)
.
2
.
4
.
M
ulti
cl
a
s
s
SVM
T
h
e
m
o
s
t
w
id
el
y
u
s
ed
m
u
lti
c
lass
SVM
m
o
d
el
s
ar
e
On
e
Vs
A
ll
(
OV
A
)
,
On
e
Vs
O
n
e
(
OVO)
[
20
]
,
Dir
ec
ted
A
c
y
clic
Gr
ap
h
(
D
AG)
[
21
]
an
d
E
r
r
o
r
C
o
r
r
ec
tin
g
Ou
tp
u
t Co
d
es (
E
C
OC
)
[
22
]
.
OVA
cr
ea
tes N
b
i
n
ar
y
SVM’
s
f
o
r
all
ca
teg
o
r
ies
w
h
e
r
e
N
is
class
n
u
m
b
er
.
Fo
r
a
n
th
SVM,
o
n
l
y
e
x
a
m
p
le
s
in
t
h
a
t
class
ar
e
p
o
s
itiv
e
an
d
r
ea
m
i
n
g
ar
e
n
e
g
ati
v
e.
T
h
e
co
m
p
u
t
atio
n
ti
m
e
is
les
s
b
u
t
at
a
co
m
p
r
o
m
is
ed
e
f
f
ic
ien
c
y
.
OVO
cr
ea
tes
a
p
air
w
is
e
0
.
5
N(
N
-
1
)
SVM
’
s
an
d
p
air
w
i
s
e
v
o
ti
n
g
to
ac
co
m
m
o
d
ate
n
e
w
s
a
m
p
les
f
o
r
s
o
lv
in
g
m
u
lti
clas
s
p
r
o
b
lem
s
.
D
AG
tr
ai
n
i
n
g
is
f
r
o
m
OVO
m
o
d
el
a
n
d
test
in
g
i
s
f
r
o
m
b
i
n
ar
y
ac
y
clic
g
r
ap
h
m
o
d
el.
E
C
O
C
d
is
a
m
b
i
g
u
ate
s
o
u
tp
u
t
b
i
n
ar
y
co
d
es
to
co
n
s
tr
u
ct
a
co
d
e
w
o
r
d
m
a
tr
ix
w
h
ic
h
i
s
co
m
p
ar
ed
w
ith
g
e
n
er
ated
b
it
v
ec
to
r
s
b
y
s
elec
ti
n
g
r
o
w
a
s
a
clas
s
h
a
v
i
n
g
m
in
i
m
u
m
h
a
m
m
i
n
g
d
is
ta
n
ce
.
T
h
is
m
et
h
o
d
g
i
v
es
g
o
o
d
class
i
f
icatio
n
r
ates
co
m
p
ar
ed
to
o
th
er
f
o
u
r
at
th
e
co
s
t
o
f
s
p
ee
d
o
f
ex
ec
u
tio
n
.
T
h
e
s
lo
w
er
s
p
ee
d
is
d
u
e
to
th
e
in
cr
ea
s
ed
le
n
g
t
h
o
f
co
d
e
w
o
r
d
s
to
d
is
a
m
b
i
g
u
ates
N
clas
s
es.
T
h
e
m
in
i
m
u
m
co
d
e
w
o
r
d
s
i
n
E
C
OC
is
lo
g
2
N
to
a
m
ax
i
m
u
m
o
f
2
N
-
1
−1
b
its
.
C
o
m
p
ar
i
n
g
t
h
e
m
u
lti
cla
s
s
SVM
m
et
h
o
d
s
f
r
o
m
M
AL
A
B
i
m
p
le
m
en
tatio
n
,
w
e
f
o
u
n
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
E
C
E
I
SS
N:
2
0
8
8
-
8708
I
n
d
ia
n
C
la
s
s
ica
l D
a
n
ce
Mu
d
r
a
C
la
s
s
ifica
tio
n
Usi
n
g
HOG
F
ea
tu
r
es a
n
d
S
V
M C
la
s
s
ifier
(
K
.
V
.
V
.
K
u
ma
r
)
2541
E
C
OC
is
p
er
f
o
r
m
s
b
etter
a
t o
p
ti
m
u
m
s
p
ee
d
s
.
T
h
e
s
i
m
ilar
it
y
m
ea
s
u
r
e
f
o
r
2
4
d
if
f
er
en
t
k
u
c
h
i
p
u
d
i d
an
ce
m
u
d
r
as
u
s
i
n
g
co
m
p
u
ter
v
is
io
n
m
o
d
el
an
d
m
ac
h
in
e
lear
n
i
n
g
al
g
o
r
ith
m
is
e
x
ec
u
ted
.
3.
RE
SU
L
T
S
A
ND
D
I
SCU
SS
I
O
N
T
h
e
d
ataset
co
n
s
i
s
ts
o
f
2
4
×5
i
m
a
g
es
o
f
I
n
d
ia
n
d
an
ce
f
o
r
m
k
u
c
h
ip
u
d
i
ar
e
co
llected
f
r
o
m
v
ar
io
u
s
s
o
u
r
ce
s
.
T
h
ese
1
2
0
im
a
g
es
ar
e
co
n
tr
ast
en
h
an
ce
d
b
y
2
0
%
to
s
m
o
o
th
p
ix
el
v
alu
e
s
.
Feat
u
r
e
ex
tr
ac
tio
n
m
o
d
u
le
is
i
n
itiated
to
e
x
tr
ac
t
f
ea
t
u
r
es
f
o
r
t
h
e
1
2
0
m
u
d
r
as.
E
ac
h
m
u
d
r
a
f
ea
t
u
r
e
i
s
lab
elled
to
id
en
ti
f
y
t
h
e
m
w
it
h
a
p
ar
ticu
lar
class
.
T
h
e
class
la
b
els
ar
e
th
e
n
a
m
e
s
o
f
t
h
e
m
u
d
r
as
i
n
k
u
c
h
ip
u
d
i
d
an
ce
f
o
r
m
.
W
e
ca
m
e
to
u
n
d
er
s
ta
n
d
th
at
t
h
e
s
e
b
asic
m
u
d
r
as a
r
e
co
m
m
o
n
to
all
I
n
d
ian
d
an
ce
f
o
r
m
s
lis
ted
i
n
[
1
]
.
Fig
u
r
e
5
s
h
o
w
s
t
h
e
f
i
v
e
f
ea
tu
r
es u
s
ed
o
n
s
o
m
e
m
u
d
r
a
s
.
Vis
u
al
o
b
s
er
v
ati
o
n
s
o
f
th
e
F
i
g
u
r
e
5
p
r
o
v
id
e
a
b
etter
p
er
f
o
r
m
i
n
g
f
ea
t
u
r
e
v
ec
t
o
r
in
th
e
s
et
{HO
G,
SUR
F,
SI
FT
,
L
B
P
an
d
HA
AR
}.
T
h
is
p
ar
ticu
lar
ap
p
licatio
n
o
f
co
m
p
u
ter
v
i
s
io
n
o
n
h
u
m
a
n
h
a
n
d
s
i
n
d
icate
H
OG
as
t
h
e
b
etter
p
er
f
o
r
m
er
co
m
p
ar
ed
to
o
th
er
4
f
ea
t
u
r
e
ex
tr
ac
tio
n
m
o
d
els.
Fig
u
r
e
5
.
Featu
r
e
Vec
to
r
g
en
e
r
atio
n
s
f
r
o
m
5
f
ea
t
u
r
e
ex
tr
ac
ti
o
n
m
o
d
els
T
h
e
o
th
er
p
r
o
b
lem
en
co
u
n
ter
ed
is
in
t
h
e
n
u
m
b
er
o
f
f
ea
tu
r
e
v
ec
to
r
s
p
r
o
d
u
ce
d
p
er
im
a
g
e
m
u
d
r
a.
Dep
en
d
in
g
o
n
th
e
h
a
n
d
d
en
s
i
t
y
in
t
h
e
i
m
a
g
e
f
r
a
m
e,
th
e
n
u
m
b
er
o
f
f
ea
t
u
r
es
i
n
ea
ch
f
e
atu
r
e
v
ec
to
r
ar
e
o
f
d
if
f
er
e
n
t
s
ize.
T
o
tack
le
th
is
p
r
o
b
lem
,
w
e
m
o
d
i
f
ied
th
e
f
ea
t
u
r
e
v
ec
to
r
s
o
f
all
i
m
ag
e
s
to
a
n
o
r
m
alize
d
f
ea
t
u
r
e
s
ize.
T
h
e
n
o
r
m
aliza
tio
n
p
r
o
ce
s
s
i
n
v
o
lv
ed
s
elec
tio
n
o
f
i
m
p
o
r
tan
t
f
ea
t
u
r
es
b
a
s
ed
o
n
m
ag
n
i
tu
d
e
o
f
f
ea
t
u
r
e
v
ec
to
r
s
.
T
h
e
m
ax
p
o
o
lin
g
al
g
o
r
ith
m
i
s
u
s
ed
to
e
x
tr
ac
t
u
s
e
f
u
l
f
ea
t
u
r
es
f
r
o
m
a
s
et
o
f
f
ea
t
u
r
es.
T
h
e
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
i
n
e
is
s
u
p
p
lied
w
i
th
th
e
n
o
r
m
alize
d
f
ea
t
u
r
e
v
ec
t
o
r
s
f
r
o
m
a
s
et
o
f
m
u
d
r
as.
T
r
ai
n
in
g
o
n
th
e
s
e
s
et
o
f
m
u
d
r
as
is
g
i
v
e
n
to
t
h
e
SV
M.
A
s
et
o
f
2
4
m
u
l
tip
le
clas
s
es
o
f
k
u
c
h
ip
u
d
i
d
a
n
ce
m
u
d
r
as
i
s
th
e
tar
g
et
v
ec
to
r
.
T
r
ain
in
g
v
ec
to
r
co
n
s
is
t
s
o
f
o
n
l
y
o
n
e
m
u
d
r
a
s
et,
th
a
t
w
as p
er
f
ec
tl
y
ca
p
tu
r
ed
d
u
r
in
g
lab
tr
ials
.
T
esti
n
g
SVM
b
y
u
s
i
n
g
t
h
e
s
a
m
e
s
et
o
f
m
u
d
r
as
r
esu
lted
i
n
a
1
0
0
%
m
atc
h
to
t
h
e
c
lass
.
T
h
e
class
lab
els
ar
e
[
P
atak
a,
T
r
ip
atak
a,
A
r
d
h
ap
ata
k
a,
Kar
tar
i
m
u
k
h
a,
Ma
y
u
r
a,
A
r
d
h
ac
h
a
n
d
r
a,
A
r
ala,
S
h
u
k
at
u
n
d
a,
Mu
s
h
t
h
i,
Sh
ik
h
ar
a,
C
h
an
d
r
ak
ala,
Sar
p
ash
ir
s
h
a,
Si
m
h
a
m
u
k
h
a,
T
r
is
h
u
la,
S
w
a
s
ti
k
a
m
,
Ma
ts
y
a,
Ku
r
m
a,
Var
ah
a,
Gar
u
d
a,
S
h
iv
a
lin
g
a
m
,
P
u
s
h
p
ap
u
ta
m
,
Kar
k
ata
m
,
Kap
o
ta
m
,
B
h
er
u
n
d
a]
.
Fo
r
HOG
f
ea
t
u
r
es
t
h
e
c
o
n
f
u
s
io
n
m
atr
i
x
is
s
h
o
w
n
i
n
F
i
g
u
r
e
6
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
J
E
C
E
Vo
l.
7
,
No
.
5
,
Octo
b
er
2
0
1
7
:
2
5
3
7
–
2
5
4
6
2542
Fig
u
r
e
6
.
C
o
n
f
u
s
io
n
m
atr
ix
b
e
t
w
ee
n
lab
d
ataset
an
d
y
o
u
t
u
b
e
f
r
a
m
e
d
atas
et
f
o
r
HOG
f
ea
t
u
r
es
T
h
e
co
n
f
u
s
io
n
m
a
tr
ix
g
iv
e
s
th
e
m
atc
h
i
n
g
m
etr
ic
s
b
et
w
ee
n
th
e
tr
ai
n
ed
s
a
m
p
les
a
n
d
th
e
test
i
n
g
s
a
m
p
les.
T
h
e
tr
ai
n
i
n
g
s
a
m
p
le
s
f
o
r
SV
M
an
d
t
h
e
te
s
ti
n
g
s
a
m
p
les
w
er
e
d
i
f
f
er
en
t
i
n
F
ig
u
r
e
6
.
T
r
ain
in
g
w
a
s
g
iv
e
n
o
n
l
y
o
n
ce
u
s
i
n
g
th
e
la
b
ca
p
tu
r
ed
d
an
ce
i
m
a
g
es.
T
esti
n
g
i
s
d
o
n
e
w
it
h
lab
i
m
a
g
e
s
,
w
eb
i
m
a
g
es
a
n
d
Yo
u
T
u
b
e
i
m
ag
e
s
.
T
h
e
co
n
f
u
s
i
o
n
m
atr
i
x
in
d
icate
s
a
7
5
%
m
at
ch
b
et
w
ee
n
t
h
e
m
u
d
r
as o
f
u
n
s
ee
n
b
y
th
e
SVM.
T
h
e
m
u
d
r
a
r
ec
o
g
n
itio
n
f
r
eq
u
en
c
y
(
M
R
F)
i
s
d
ef
i
n
ed
as
n
u
m
b
er
o
f
co
r
r
ec
tl
y
r
ec
o
g
n
ized
m
u
d
r
as
to
to
tal
n
u
m
b
er
o
f
m
u
d
r
a
s
u
s
ed
f
o
r
clas
s
if
icatio
n
.
MRF
is
i
m
p
r
o
v
ed
b
y
i
n
cr
ea
s
i
n
g
th
e
tr
ai
n
i
n
g
v
ec
to
r
f
r
o
m
o
n
e
m
u
d
r
a
s
et
to
t
w
o
m
u
d
r
a
s
et
s
.
Fo
r
a
2
m
u
d
r
a
s
e
t
tr
ai
n
in
g
i
n
v
o
l
v
i
n
g
b
o
th
lab
ca
p
t
u
r
ed
a
n
d
w
eb
i
m
ag
e
s
t
h
e
class
i
f
icatio
n
o
f
u
n
s
ee
n
m
u
d
r
a
s
in
i
m
p
r
o
v
ed
b
y
1
4
%.
T
h
e
M
R
F i
n
th
i
s
ca
s
e
i
s
ar
o
u
n
d
8
9
%.
T
h
e
r
ea
s
o
n
f
o
r
m
i
s
clas
s
i
f
icati
o
n
b
et
w
ee
n
s
i
m
i
lar
m
u
d
r
as
ca
n
b
e
attr
ib
u
ted
to
t
h
e
h
a
n
d
s
h
ap
es,
h
a
n
d
co
lo
r
s
,
h
an
d
o
r
ien
tatio
n
s
an
d
h
an
d
te
x
tu
r
e
s
.
T
h
e
n
ex
t
b
est
f
ea
tu
r
e
v
ec
to
r
th
at
h
as
s
h
o
w
n
m
ax
i
m
u
m
MR
F
is
SIFT
.
Fo
r
s
am
e
tr
ain
i
n
g
an
d
t
esti
n
g
v
ec
to
r
s
,
SIFT
h
as a
M
R
F o
f
9
6
%.
Ho
w
ev
er
,
t
h
e
MR
F
is
d
o
w
n
to
7
1
% f
o
r
d
if
f
er
e
n
t
tr
ai
n
in
g
a
n
d
test
i
n
g
s
a
m
p
les.
I
n
ca
s
e
o
f
m
u
ltip
le
tr
ain
i
n
g
v
ec
to
r
s
,
SIFT
s
h
o
w
ed
an
i
n
cr
ea
s
e
i
n
MR
F
b
y
9
%
.
T
h
e
o
v
er
all
MR
F
’
s
f
o
r
all
5
f
ea
t
u
r
e
v
ec
to
r
s
u
s
ed
i
s
p
lo
tted
in
F
ig
u
r
e
7
.
P
lo
t
s
h
o
w
s
MR
F
v
alu
e
s
f
o
r
s
in
g
le
s
a
m
e
s
et
tr
ain
in
g
an
d
te
s
tin
g
(
1
STT
V)
,
s
in
g
le
d
if
f
er
en
t
s
et
tr
ain
in
g
an
d
test
in
g
(
1
D
T
T
V)
,
t
w
o
s
a
m
e
s
et
tr
ain
i
n
g
a
n
d
te
s
ti
n
g
(
2
STT
V)
an
d
t
w
o
d
if
f
er
e
n
t
s
e
t
tr
ai
n
in
g
an
d
tes
tin
g
(
2
DT
T
V)
.
A
lo
n
g
w
ith
t
h
e
p
r
o
p
o
s
ed
f
ea
t
u
r
e
ex
tr
ac
tio
n
m
o
d
els,
w
e
also
test
ed
SVM
w
it
h
m
u
lt
i
-
f
e
atu
r
e
ex
tr
ac
tio
n
m
o
d
el
s
w
i
th
HOG
co
m
b
in
atio
n
.
T
h
e
p
lo
ts
s
h
o
w
an
in
cr
ea
s
e
in
th
e
MR
F
v
al
u
e
f
o
r
HOG
co
m
b
i
n
ed
f
ea
t
u
r
es.
Oth
er
co
m
b
in
atio
n
s
o
f
f
ea
t
u
r
e
v
ec
to
r
s
ar
e
also
test
ed
b
u
t
th
e
r
esu
lt
s
w
er
e
n
o
t
e
n
co
u
r
ag
in
g
an
d
w
er
e
d
is
ca
r
d
ed
f
r
o
m
r
ep
o
r
tin
g
in
t
h
e
p
lo
t
o
f
F
ig
u
r
e
7
.
Fig
u
r
e
7
.
P
lo
t c
o
m
p
ar
in
g
v
ar
i
o
u
s
f
ea
t
u
r
e
v
ec
to
r
s
w
ith
r
esp
e
ct
to
MRF
Pa
t
Trip
A
r
d
p
Ka
r
M
a
y
A
r
d
c
A
r
a
Shu
Mus
Shi
C
h
a
S
a
r
S
i
m
Tri
s
S
w
a
M
a
t
Ku
r
V
a
r
Ga
r
Shi
Pus
h
Ka
r
k
Ka
p
B
h
e
r
Pa
t
a
k
a
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
T
r
i
p
a
t
a
k
a
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
A
r
d
h
a
p
a
t
a
k
a
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Ka
r
t
a
r
i
m
u
k
h
a
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
M
a
y
u
r
a
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
A
r
d
h
a
c
h
a
n
d
r
a
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
A
r
a
l
a
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
S
h
u
k
a
t
u
n
d
a
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
M
u
sh
t
h
i
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
S
h
i
k
h
a
r
a
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
C
h
a
n
d
r
a
k
a
l
a
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
S
a
r
p
a
sh
i
r
sh
a
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
S
i
m
h
a
m
u
k
h
a
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
T
r
i
sh
u
l
a
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
S
w
a
st
i
k
a
m
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
M
a
t
sy
a
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
Ku
r
m
a
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
V
a
r
a
h
a
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
Ga
r
u
d
a
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
S
h
i
v
a
l
i
n
g
a
m
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
Pus
h
p
a
p
u
t
a
m
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
Ka
r
k
a
t
a
m
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
Ka
p
o
t
a
m
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
B
h
e
r
u
n
d
a
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
E
C
E
I
SS
N:
2
0
8
8
-
8708
I
n
d
ia
n
C
la
s
s
ica
l D
a
n
ce
Mu
d
r
a
C
la
s
s
ifica
tio
n
Usi
n
g
HOG
F
ea
tu
r
es a
n
d
S
V
M C
la
s
s
ifier
(
K
.
V
.
V
.
K
u
ma
r
)
2543
T
h
e
r
esu
lts
ar
e
v
al
id
ated
b
y
m
ea
s
u
r
in
g
a
p
ar
a
m
eter
ca
ll
ed
s
tr
u
ctu
r
al
s
i
m
ilar
it
y
in
d
e
x
(
SS
I
M)
b
et
w
ee
n
th
e
m
u
d
r
as
u
s
in
g
a
g
r
ap
h
ical
u
s
er
in
ter
f
ac
e
(
GUI
)
.
T
h
e
GUI
h
as
o
n
o
n
e
s
id
e
a
q
u
er
y
m
u
d
r
a
an
d
th
e
o
th
er
s
id
e
i
n
a
m
u
d
r
a
o
f
d
i
f
f
er
en
t
i
m
a
g
e
alo
n
g
w
it
h
t
h
eir
n
a
m
es.
Fi
g
u
r
es
8
an
d
9
s
h
o
w
s
t
h
e
o
p
er
atio
n
o
f
GUI
f
o
r
s
a
m
e
m
u
d
r
a
i
m
ag
e
s
an
d
d
i
f
f
er
en
t
m
u
d
r
a
i
m
ag
e
s
f
o
r
th
e
s
a
m
e
clas
s
.
Fig
u
r
e
8
.
GUI
s
h
o
w
i
n
g
S
SIM
f
o
r
s
a
m
e
m
u
d
r
a
i
m
ag
e
o
f
a
cla
s
s
f
o
r
tr
ain
in
g
an
d
te
s
tin
g
Fig
u
r
e
9
.
GUI
s
h
o
w
i
n
g
S
SIM
f
o
r
d
if
f
er
e
n
t
m
u
d
r
a
i
m
a
g
e
o
f
a
class
f
o
r
tr
ain
in
g
an
d
tes
tin
g
HOG
f
ea
t
u
r
es
i
s
p
r
o
v
id
in
g
g
o
o
d
f
ea
tu
r
e
ex
tr
ac
tio
n
co
m
p
ar
e
d
to
o
th
er
ty
p
e
s
o
f
f
ea
tu
r
e
s
c
o
m
m
o
n
l
y
u
s
ed
i
n
co
m
p
u
ter
v
i
s
io
n
f
o
r
d
ac
e
m
u
d
r
a
class
if
ica
tio
n
.
I
n
th
i
s
w
o
r
k
,
w
e
a
ls
o
u
s
ed
t
w
o
m
o
r
e
s
eg
m
e
n
taio
n
m
et
h
o
d
s
p
o
p
u
lar
w
it
h
co
m
p
u
t
er
v
is
io
n
n
a
m
el
y
m
ar
k
er
co
n
t
r
o
lled
w
ater
s
h
ed
(
MCW
)
[
2
3
]
an
d
s
i
m
p
le
li
n
ea
r
iter
ativ
e
c
lu
s
ter
in
g
(
S
L
I
C
)
[
2
4
]
f
o
r
d
an
ce
m
u
d
r
a
s
e
g
m
e
n
tati
o
n
.
T
h
e
s
h
ap
e
o
r
ed
g
e
f
ea
t
u
r
e
s
ar
e
e
x
tr
ac
ted
f
r
o
m
th
e
m
u
d
r
a
i
m
a
g
es a
n
d
ar
e
clas
s
if
ied
u
s
i
n
g
SVM
clas
s
i
f
ier
.
T
h
e
s
i
m
ilar
it
y
es
ti
m
ate
b
et
w
e
en
th
e
HOG,
M
C
W
an
d
S
L
I
C
ar
e
co
m
p
ar
ed
u
s
in
g
Str
u
ct
u
r
al
s
i
m
ilar
it
y
in
d
ex
m
ea
s
u
r
e
(
SS
I
M)
[
2
5
]
o
n
th
e
e
x
t
r
ac
ted
f
ea
tu
r
e
s
f
r
o
m
a
h
an
d
s
e
g
m
e
n
ted
i
m
a
g
e.
T
h
e
s
i
m
ilar
it
y
i
n
d
ex
m
ea
s
u
r
es
ar
e
s
h
o
w
n
i
n
T
ab
le
2
.
W
e
f
in
d
th
at
H
OG
f
ea
t
u
r
es
ar
e
clo
s
el
y
r
elate
d
in
s
tr
u
ctu
r
al
co
n
n
ec
tiv
it
y
co
m
p
ar
ed
to
o
th
er
t
w
o
m
et
h
o
d
s
o
f
f
ea
t
u
r
e
ex
tr
ac
tio
n
.
He
n
c
e
w
h
en
cla
s
s
i
f
ied
u
s
i
n
g
M
u
lti
class
SVM,
HO
G
f
ea
t
u
r
es
r
eg
i
s
ter
ed
a
h
i
g
h
er
M
R
F
o
f
9
8
%.
Fo
r
MCW
an
d
S
L
I
C
M
R
F
i
s
ar
o
u
n
d
6
6
%,
w
h
i
ch
is
w
a
y
b
elo
w
th
e
HOG.
T
ab
le
2
.
SS
I
M
b
ased
C
o
m
p
ar
i
s
o
n
M
u
d
r
a
N
a
me
s
W
a
t
e
r
sh
e
d
B
a
se
d
P
o
se
Est
i
ma
t
i
o
n
S
l
i
c
B
a
se
sd
D
a
n
c
e
P
o
se
Est
i
ma
t
i
o
n
O
u
r
P
r
o
p
o
se
d
M
e
t
h
o
d
P
a
t
a
k
a
0
.
6
6
8
0
.
7
7
1
0
.
8
7
1
T
r
i
p
a
t
a
k
a
0
.
6
8
1
0
.
7
8
2
0
.
8
8
2
A
r
d
h
a
p
a
t
a
k
a
0
.
6
5
6
0
.
7
4
8
0
.
8
4
8
K
a
r
t
a
r
i
mu
k
h
a
0
.
6
4
9
0
.
7
5
7
0
.
8
5
7
M
a
y
u
r
a
0
.
6
5
2
0
.
7
8
7
0
.
8
8
7
A
r
d
h
a
c
h
a
n
d
r
a
0
.
6
5
9
0
.
7
6
2
0
.
8
6
2
A
r
a
l
a
0
.
6
6
1
0
.
7
5
1
0
.
8
5
1
S
h
u
k
a
t
u
n
d
a
0
.
6
7
4
0
.
7
7
9
0
.
8
7
9
M
u
s
h
t
h
i
0
.
6
7
8
0
.
7
7
3
0
.
8
7
3
S
h
i
k
h
a
r
a
0
.
6
6
1
0
.
7
7
1
0
.
8
7
1
C
h
a
n
d
r
a
k
a
l
a
0
.
5
0
1
0
.
6
8
1
0
.
7
8
1
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
J
E
C
E
Vo
l.
7
,
No
.
5
,
Octo
b
er
2
0
1
7
:
2
5
3
7
–
2
5
4
6
2544
S
a
r
p
a
s
h
i
r
s
h
a
0
.
5
2
4
0
.
6
9
6
0
.
7
9
6
S
i
mh
a
mu
k
h
a
0
.
5
9
8
0
.
6
7
1
0
.
7
7
1
T
r
i
sh
u
l
a
0
.
5
8
1
0
.
6
5
4
0
.
7
5
4
S
w
a
st
i
k
a
m
0
.
5
2
7
0
.
7
0
7
0
.
7
0
7
M
a
t
sy
a
0
.
5
1
1
0
.
6
9
4
0
.
7
9
4
K
u
r
ma
0
.
5
9
4
0
.
6
6
9
0
.
7
6
9
V
a
r
a
h
a
0
.
6
1
7
0
.
6
9
9
0
.
7
9
9
G
a
r
u
d
a
0
.
6
2
4
0
.
6
9
8
0
.
8
2
4
S
h
i
v
a
l
i
n
g
a
m
0
.
6
8
7
0
.
7
1
9
0
.
8
1
4
P
u
sh
p
a
p
u
t
a
m
0
.
6
4
8
0
.
7
5
1
0
.
8
0
1
K
a
r
k
a
t
a
m
0
.
6
2
4
0
.
7
6
4
0
.
7
9
9
K
a
p
o
t
a
m
0
.
6
2
1
0
.
7
4
1
0
.
8
0
6
B
h
e
r
u
n
d
a
0.
6
29
0
.
7
2
4
0
.
7
9
4
T
ab
le
3
.
P
SNR
b
ased
C
o
m
p
ar
is
o
n
M
u
d
r
a
N
a
me
s
W
a
t
e
r
sh
e
d
B
a
se
d
P
o
se
Est
i
ma
t
i
o
n
S
l
i
c
B
a
se
sd
D
a
n
c
e
P
o
se
Est
i
ma
t
i
o
n
O
u
r
P
r
o
p
o
se
d
M
e
t
h
o
d
P
a
t
a
k
a
8
.
3
5
9
.
1
5
1
2
.
7
4
T
r
i
p
a
t
a
k
a
8
.
1
7
9
.
2
7
1
2
.
5
9
A
r
d
h
a
p
a
t
a
k
a
8
.
0
1
8
.
8
1
1
2
.
4
6
K
a
r
t
a
r
i
mu
k
h
a
8
.
3
4
9
.
2
6
1
2
.
5
7
M
a
y
u
r
a
8
.
8
1
9
.
3
8
1
2
.
4
9
A
r
d
h
a
c
h
a
n
d
r
a
8
.
5
2
9
.
2
9
1
2
.
6
5
A
r
a
l
a
8
.
4
5
8
.
9
2
1
2
.
3
8
S
h
u
k
a
t
u
n
d
a
8
.
5
7
9
.
0
9
1
2
.
2
9
M
u
s
h
t
h
i
8
.
1
5
9
.
2
1
1
2
.
5
6
S
h
i
k
h
a
r
a
8
.
4
7
9
.
5
1
1
2
.
9
3
C
h
a
n
d
r
a
k
a
l
a
7
.
9
2
7
.
9
1
1
0
.
1
2
S
a
r
p
a
s
h
i
r
s
h
a
7
.
8
5
8
.
0
1
1
0
.
2
6
S
i
mh
a
mu
k
h
a
8
.
0
5
8
.
9
6
1
0
.
0
4
T
r
i
sh
u
l
a
7
.
9
5
8
.
8
9
9
.
9
8
S
w
a
st
i
k
a
m
7
.
9
9
8
.
1
2
1
0
.
4
6
M
a
t
sy
a
8
.
0
4
8
.
9
9
1
0
.
0
1
K
u
r
ma
8
.
1
1
8
.
1
4
1
0
.
1
7
V
a
r
a
h
a
7
.
6
5
8
.
0
7
1
0
.
3
1
G
a
r
u
d
a
7
.
2
4
7
.
9
4
1
0
.
5
6
S
h
i
v
a
l
i
n
g
a
m
8
.
0
1
7
.
9
9
1
1
.
6
5
P
u
sh
p
a
p
u
t
a
m
7
.
9
1
8
.
2
1
1
0
.
9
1
K
a
r
k
a
t
a
m
7
.
7
6
8
.
1
9
1
0
.
5
6
K
a
p
o
t
a
m
7
.
9
2
8
.
1
1
1
0
.
8
4
B
h
e
r
u
n
d
a
8
.
0
9
8
.
2
6
1
0
.
9
9
T
ab
le
4
.
I
QI
b
ased
co
m
p
ar
is
o
n
M
u
d
r
a
N
a
me
s
W
a
t
e
r
sh
e
d
B
a
se
d
P
o
se
Est
i
ma
t
i
o
n
S
l
i
c
B
a
se
sd
D
a
n
c
e
P
o
se
Est
i
ma
t
i
o
n
O
u
r
P
r
o
p
o
se
d
M
e
t
h
o
d
P
a
t
a
k
a
0
.
7
2
3
0
.
7
6
2
0
.
8
9
6
T
r
i
p
a
t
a
k
a
0
.
7
3
6
0
.
7
7
3
0
.
9
0
7
A
r
d
h
a
p
a
t
a
k
a
0
.
6
9
8
0
.
7
3
9
0
.
8
7
3
K
a
r
t
a
r
i
mu
k
h
a
0
.
6
9
1
0
.
7
4
8
0
.
8
8
2
M
a
y
u
r
a
0
.
6
9
4
0
.
7
7
8
0
.
9
1
2
A
r
d
h
a
c
h
a
n
d
r
a
0
.
7
0
1
0
.
7
5
3
0
.
8
8
7
A
r
a
l
a
0
.
7
0
3
0
.
7
4
2
0
.
8
7
6
S
h
u
k
a
t
u
n
d
a
0
.
7
1
6
0
.
7
7
0
0
.
9
0
4
M
u
s
h
t
h
i
0
.
7
2
0
0
.
7
6
4
0
.
8
9
8
S
h
i
k
h
a
r
a
0
.
7
0
3
0
.
7
6
2
0
.
8
9
6
C
h
a
n
d
r
a
k
a
l
a
0
.
5
4
3
0
.
6
7
2
0
.
8
0
6
S
a
r
p
a
s
h
i
r
s
h
a
0
.
5
6
6
0
.
6
8
7
0
.
8
2
1
S
i
mh
a
mu
k
h
a
0
.
6
4
0
0
.
6
6
2
0
.
7
9
6
T
r
i
sh
u
l
a
0
.
6
2
3
0
.
6
4
5
0
.
7
7
9
S
w
a
st
i
k
a
m
0
.
5
6
9
0
.
6
9
8
0
.
7
3
2
M
a
t
sy
a
0
.
5
5
3
0
.
6
8
5
0
.
8
1
9
K
u
r
ma
0
.
6
3
6
0
.
6
6
0
0
.
7
9
4
V
a
r
a
h
a
0
.
6
5
9
0
.
6
9
0
0
.
8
2
4
G
a
r
u
d
a
0
.
6
6
6
0
.
6
8
9
0
.
8
4
9
S
h
i
v
a
l
i
n
g
a
m
0
.
7
2
9
0
.
7
1
0
0
.
8
3
9
P
u
sh
p
a
p
u
t
a
m
0
.
6
9
0
0
.
7
4
2
0
.
8
2
6
K
a
r
k
a
t
a
m
0
.
6
6
6
0
.
7
5
5
0
.
8
2
4
K
a
p
o
t
a
m
0
.
6
6
3
0
.
7
3
2
0
.
8
3
1
B
h
e
r
u
n
d
a
0
.
6
7
1
0
.
7
1
5
0
.
8
1
9
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
E
C
E
I
SS
N:
2
0
8
8
-
8708
I
n
d
ia
n
C
la
s
s
ica
l D
a
n
ce
Mu
d
r
a
C
la
s
s
ifica
tio
n
Usi
n
g
HOG
F
ea
tu
r
es a
n
d
S
V
M C
la
s
s
ifier
(
K
.
V
.
V
.
K
u
ma
r
)
2545
P
ea
k
s
ig
n
al
to
n
o
i
s
e
r
atio
(
P
SNR
)
[
2
6
]
in
d
b
is
a
m
e
asu
r
e
o
f
d
i
f
f
er
e
n
tiati
n
g
t
h
e
a
m
o
u
n
t
o
f
u
n
w
n
ated
n
es
s
in
t
h
e
i
m
a
g
e
f
ea
tu
r
es.
W
h
e
n
P
SNR
i
s
co
m
p
u
ted
o
n
ex
tr
ac
ted
f
ea
tu
r
e
s
a
n
d
th
e
g
r
o
u
n
d
tr
u
t
h
f
ea
t
u
r
es,
HOG
f
ea
t
u
r
es
f
o
r
d
a
n
ce
m
u
d
r
a
i
m
a
g
es
r
e
g
is
ter
ed
h
ig
h
e
s
t
v
alu
e.
T
h
is
r
es
u
lt
s
ar
e
s
h
o
w
n
i
n
T
ab
le
3
.
P
SNR
s
h
o
w
s
t
h
at
t
h
er
e
is
a
h
i
g
h
r
elat
iv
i
t
y
b
et
w
ee
n
HOG
f
e
atu
r
es o
f
o
r
i
g
in
a
l
m
u
d
r
as a
n
d
th
e
Ha
n
d
s
eg
m
e
n
ted
g
r
o
u
n
d
tr
u
th
m
u
d
r
a
s
,
co
m
p
ar
ed
to
MCW
an
d
SL
I
C
f
ea
t
u
r
es.
I
m
ag
e
q
u
alit
y
i
n
d
ex
(
I
QI
)
[
2
7
]
m
ea
s
u
r
es
th
e
q
u
alit
y
o
f
f
ea
t
u
r
es
f
r
o
m
a
ll
t
h
r
ee
al
g
o
r
ith
m
s
co
m
p
ar
ed
to
g
r
o
u
n
d
tr
u
t
h
i
m
a
g
e.
T
ab
le
4
g
iv
es
th
e
m
ea
s
u
r
es
o
f
co
m
p
ar
i
s
o
n
o
n
d
if
f
er
en
t
d
an
ce
m
u
d
r
a
i
m
a
g
e
s
.
T
h
e
v
alu
e
s
s
h
o
w
th
a
t
HOG
m
ati
n
s
g
o
o
d
f
ea
tu
r
e
q
u
alit
y
co
m
p
ar
ed
to
th
e
o
t
h
er
t
w
o
a
lg
o
r
it
h
m
s
MCW
an
d
SL
I
C
.
Fo
r
v
id
eo
d
an
ce
c
lass
if
icatio
n
a
n
d
an
a
l
y
s
is
,
S
L
I
C
an
d
Activ
e
C
o
n
to
u
r
s
ca
n
e
x
tr
ac
t
b
etter
s
e
g
m
e
n
tatio
n
o
u
tp
u
ts
w
h
e
n
m
o
r
e
in
f
o
r
m
a
tio
n
in
t
h
e
f
o
r
m
o
f
co
lo
r
,
s
h
ap
e
an
d
tex
t
u
r
e
in
f
o
r
m
atio
n
is
p
r
o
v
id
ed
as
an
in
p
u
t
to
th
e
co
n
v
e
x
o
p
ti
m
izatio
n
f
u
n
ctio
n
.
I
n
f
u
t
u
r
e
w
o
r
k
s
,
w
e
p
r
o
p
o
s
e
to
u
s
e
f
ea
tu
r
es
d
esc
r
ib
ed
in
[
2
8
]
w
it
h
A
r
ti
f
icial
n
e
u
r
al
n
et
w
o
r
k
m
o
d
els
i
n
[
2
9
]
r
elate
d
to
b
ig
d
ata
ap
p
licatio
n
s
m
o
d
u
les
d
is
c
u
s
s
e
d
in
[
3
0
]
o
n
I
n
d
ia
n
class
ical
d
an
ce
v
id
eo
s
av
ai
lab
le
in
o
u
r
d
atab
ase
an
d
th
e
i
n
ter
n
et.
4.
CO
NCLU
SI
O
N
An
atte
m
p
t
i
s
m
ad
e
to
f
i
n
d
s
i
m
ilar
it
y
b
et
w
e
en
d
a
n
ce
m
u
d
r
as
o
f
I
n
d
ia
n
cla
s
s
ical
d
an
ce
f
o
r
m
k
u
c
h
ip
u
d
i
b
a
s
ed
o
n
i
m
ag
e
p
r
o
ce
s
s
in
g
m
o
d
els
a
n
d
p
atter
n
class
i
f
ier
s
.
Fiv
e
f
ea
t
u
r
e
ex
tr
ac
tio
n
tec
h
n
iq
u
es
ar
e
co
m
p
ar
ed
f
o
r
t
h
is
w
o
r
k
.
M
u
lti
class
S
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
class
i
f
ied
t
h
ese
f
ea
tu
r
e
s
an
d
t
h
e
p
er
f
o
r
m
a
n
ce
o
f
th
e
clas
s
i
f
ier
w
it
h
r
esp
ec
t
to
a
p
ar
ticu
lar
f
ea
t
u
r
e
is
m
ea
s
u
r
ed
.
Vis
u
al
v
er
if
icatio
n
a
n
d
s
tr
u
ctu
r
al
v
er
i
f
icatio
n
u
s
i
n
g
S
SIM
ar
e
p
r
ef
o
r
m
ed
to
ch
ec
k
t
h
e
cla
s
s
i
f
ier
s
p
er
f
o
r
m
an
ce
.
T
h
e
SVM
clas
s
i
f
ier
r
eg
is
ter
ed
an
a
v
er
a
g
e
MRF
o
f
9
0
%
w
it
h
HOG
f
ea
t
u
r
e
v
ec
to
r
an
d
th
e
r
e
m
ai
n
i
n
g
f
e
atu
r
e
v
ec
to
r
s
p
r
o
d
u
ce
d
les
s
th
an
8
0
%
m
a
tch
i
n
g
.
A
GUI
is
b
u
i
lt
to
v
alid
ate
th
e
r
esu
lt
s
p
r
o
d
u
ce
d
b
y
f
ea
t
u
r
e
v
e
cto
r
s
u
s
i
n
g
S
SIM
in
d
icato
r
.
T
h
e
SS
I
M
in
d
icato
r
h
as
c
lo
s
el
y
r
elate
d
t
h
e
f
ea
t
u
r
e
m
o
d
el
s
o
f
HOG
an
d
SIFT
f
o
r
s
a
m
e
an
d
d
i
f
f
er
en
t
d
ata
s
et
s
.
Mo
s
t
o
f
th
e
m
u
d
r
as
w
it
h
t
w
o
h
a
n
d
s
p
r
o
d
u
ce
d
o
cc
lu
s
io
n
s
t
h
at
i
n
d
u
ce
d
b
o
ttlen
ec
k
’
s
d
u
r
in
g
f
ea
tu
r
e
e
x
tr
ac
tio
n
s
t
ag
e.
T
h
is
m
o
d
el
o
f
m
u
d
r
a
clas
s
i
f
icatio
n
w
i
ll h
e
lp
en
h
a
n
ce
t
h
e
lear
n
i
n
g
ca
p
ac
it
y
o
f
a
f
ir
s
t ti
m
e
lear
n
er
.
RE
F
E
R
E
NC
E
S
[1
]
A
K
a
ra
n
a
in
d
a
n
c
e
is
d
e
f
in
e
d
a
s "
T
h
e
c
o
o
rd
in
a
ti
o
n
o
f
th
e
m
o
v
e
m
e
n
ts
o
f
th
e
h
a
n
d
s an
d
f
e
e
t"
,
Na
id
u
,
N.
Na
id
u
,
B.
V
.
P
a
n
tu
l
u
,
P
.
R.
T
a
n
d
a
v
a
L
a
k
sh
a
n
a
m
:T
h
e
F
u
n
d
a
m
e
n
tals
o
f
A
n
c
i
e
n
t
Hin
d
u
Da
n
c
in
g
.
Ne
w
De
lh
i,
:
M
u
n
sh
iram
M
a
n
o
h
a
rlal,
1
9
7
1
,
1
9
8
0
,
p
.
1
9
.
[2
]
h
tt
p
:
//
n
a
ty
a
n
jali.
b
lo
g
sp
o
t.
i
n
/
[3
]
Ok
a
d
a
,
A
.
,
Bu
c
k
in
g
h
a
m
S
h
u
m
,
S
.
&
S
h
e
rb
o
rn
e
,
T
.
2
0
0
8
K
n
o
w
led
g
e
Ca
rto
g
ra
p
h
y
:
S
o
f
t
w
a
re
T
o
o
ls
a
n
d
M
a
p
p
in
g
T
e
c
h
n
iq
u
e
s.
(e
d
s.)
L
o
n
d
o
n
:
S
p
ri
n
g
e
r.
[4
]
Ba
il
e
y
,
H,
Bu
c
k
in
g
h
a
m
S
h
u
m
,
S
,
L
e
Blan
c
,
A
,
P
o
p
a
t,
S
,
Ro
w
ley
,
A
a
n
d
T
u
rn
e
r,
M
(2
0
0
9
)
Da
n
c
i
n
g
o
n
th
e
G
rid
:
Us
in
g
e
-
S
c
ien
c
e
to
o
ls
to
e
x
ten
d
c
h
o
re
o
g
ra
p
h
ic
re
se
a
rc
h
.
Ph
il
o
s
o
p
h
ica
l
T
ra
n
sa
c
ti
o
n
s
o
f
t
h
e
Ro
y
a
l
S
o
c
iety
A
,
1
8
9
8
;
3
6
7
:
2
7
9
3
–
2
8
0
6
.
[5
]
W
e
ic
k
,
Ka
rl
E.
S
e
n
se
m
a
k
in
g
in
o
rg
a
n
iza
ti
o
n
s.
V
o
l
.
3
.
S
a
g
e
,
1
9
9
5
.
[6
]
Ko
z
e
l,
S
u
sa
n
.
Clo
se
r:
p
e
rf
o
rm
a
n
c
e
,
tec
h
n
o
l
o
g
ies
,
p
h
e
n
o
m
e
n
o
lo
g
y
.
M
IT
P
re
ss
,
2
0
0
7
.
[7
]
Ch
a
tt
e
rjee
,
S
.
,
M
a
tri
x
e
stim
a
ti
o
n
b
y
u
n
iv
e
rsa
l
sin
g
u
lar
v
a
lu
e
th
re
sh
o
l
d
i
n
g
.
T
h
e
An
n
a
ls
o
f
S
t
a
ti
stics
,
2
0
1
5
;
4
3
(
1
):
1
7
7
-
2
1
4
.
[8
]
IKG
D
P
u
tra,
E
Erd
iaw
a
n
,
"
Hig
h
p
e
rf
o
rm
a
n
c
e
p
a
lm
p
rin
t
id
e
n
ti
f
ica
ti
o
n
sy
ste
m
b
a
se
d
o
n
tw
o
d
im
e
n
sio
n
a
l
g
a
b
o
r
,
”
T
EL
KOM
NIKA
T
e
lec
o
mm
u
n
ic
a
ti
o
n
C
o
mp
u
ti
n
g
El
e
c
tro
n
ics
a
n
d
Co
n
tro
l.
,
v
o
l.
8
,
n
o
.
3
,
p
p
.
3
0
9
-
3
1
8
,
2
0
1
0
.
[9
]
Ce
len
k
,
M
.
,
A
c
o
lo
r
c
l
u
ste
rin
g
tec
h
n
iq
u
e
f
o
r
im
a
g
e
se
g
m
e
n
ta
ti
o
n
.
Co
m
p
u
ter
Vi
si
o
n
,
Gr
a
p
h
i
c
s,
a
n
d
Ima
g
e
Pro
c
e
ss
in
g
,
1
9
9
0
;
5
2
(2
),
1
4
5
-
1
7
0
.
[1
0
]
Kish
o
re
,
P
.
V.
V
.
,
D.
A
n
il
Ku
m
a
r,
E.
N.
D.
G
o
u
th
a
m
,
a
n
d
M
.
M
a
n
ik
a
n
ta.
Co
n
ti
n
u
o
u
s
sig
n
l
a
n
g
u
a
g
e
re
c
o
g
n
it
i
o
n
fro
m
tra
c
k
in
g
a
n
d
sh
a
p
e
f
e
a
t
u
re
s
u
sin
g
F
u
zz
y
In
fer
e
n
c
e
En
g
i
n
e
.
In
W
irele
ss
Co
m
m
u
n
ica
ti
o
n
s,
S
i
g
n
a
l
P
r
o
c
e
ss
in
g
a
n
d
Ne
tw
o
rk
in
g
(W
iS
P
NET
),
I
n
t
e
rn
a
ti
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
,
2
0
1
6
;
2
1
6
5
-
21
7
0
.
IEE
E.
[1
1
]
Kish
o
re
,
P
.
V.
V
.
,
S
.
R.
C.
Kish
o
re
,
a
n
d
M
.
V
.
D.
P
ra
sa
d
.
Co
n
g
lo
m
e
ra
ti
o
n
o
f
h
a
n
d
sh
a
p
e
s
a
n
d
tex
tu
re
in
f
o
rm
a
ti
o
n
f
o
r
re
c
o
g
n
izin
g
g
e
stu
re
s
o
f
In
d
ian
sig
n
lan
g
u
a
g
e
u
sin
g
f
e
e
d
f
o
rw
a
rd
n
e
u
ra
l
n
e
tw
o
rk
s.
In
ter
n
a
ti
o
n
a
l
J
o
u
r
n
a
l
o
f
e
n
g
in
e
e
rin
g
a
n
d
T
e
c
h
n
o
l
o
g
y
(
IJ
ET
)
,
IS
S
N
(2
0
1
3
):
0
9
7
5
-
4
0
2
4
.
[1
2
]
A
n
a
n
d
h
,
A
.
,
K.
M
a
la,
a
n
d
S
.
S
u
g
a
n
y
a
.
Co
n
ten
t
b
a
se
d
im
a
g
e
re
triev
a
l
sy
ste
m
b
a
se
d
o
n
se
ma
n
ti
c
in
f
o
rm
a
ti
o
n
u
si
n
g
c
o
lo
r,
tex
tu
re
a
n
d
sh
a
p
e
fea
t
u
re
s
.
In
Co
m
p
u
ti
n
g
T
e
c
h
n
o
lo
g
ies
a
n
d
In
telli
g
e
n
t
Da
ta
En
g
in
e
e
ri
n
g
(ICCT
IDE),
In
tern
a
ti
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
,
2
0
1
6
;
1
-
8
.
I
EE
E.
[1
3
]
L
o
we
,
D.
G
.
,
Distin
c
ti
v
e
i
m
a
g
e
fe
a
tu
re
s
f
ro
m
sc
a
l
e
-
in
v
a
rian
t
k
e
y
p
o
i
n
ts.
In
ter
n
a
ti
o
n
a
l
jo
u
rn
a
l
o
f
c
o
mp
u
ter
v
isio
n
,
2
0
0
4
;
6
0
(
2
):
91
-
1
1
0
.
[1
4
]
V
io
la,
P
.
a
n
d
M
.
Jo
n
e
s.
Ra
p
id
o
b
j
e
c
t
d
e
tec
ti
o
n
u
sin
g
a
b
o
o
ste
d
c
a
s
c
a
d
e
o
f
simp
le
fea
tu
re
s
.
in
Co
m
p
u
ter
V
isi
o
n
a
n
d
P
a
tt
e
r
n
Re
c
o
g
n
it
i
o
n
,
2
0
0
1
.
C
VP
R
2
0
0
1
.
P
r
o
c
e
e
d
in
g
s
o
f
th
e
2
0
0
1
IEE
E
Co
m
p
u
ter
S
o
c
iety
Co
n
f
e
re
n
c
e
o
n
.
2
0
0
1
.
IEE
E.
[1
5
]
Ro
ste
n
,
E.
a
n
d
T
.
Dru
m
m
o
n
d
.
M
a
c
h
in
e
lea
rn
i
n
g
f
o
r
h
ig
h
-
sp
e
e
d
c
o
rn
e
r
d
e
tec
ti
o
n
.
i
n
Eu
r
o
p
e
a
n
c
o
n
f
e
re
n
c
e
o
n
c
o
m
p
u
ter v
isio
n
.
2
0
0
6
.
S
p
r
in
g
e
r.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
J
E
C
E
Vo
l.
7
,
No
.
5
,
Octo
b
er
2
0
1
7
:
2
5
3
7
–
2
5
4
6
2546
[1
6
]
Ba
y
,
H.,
e
t
a
l.
,
S
p
e
e
d
e
d
-
u
p
ro
b
u
s
t
f
e
a
tu
re
s
(S
URF).
Co
mp
u
ter
v
isio
n
a
n
d
ima
g
e
u
n
d
e
rs
ta
n
d
in
g
,
2
0
0
8
;
1
1
0
(
3
):
3
4
6
-
3
5
9
.
[1
7
]
Ha
m
e
d
,
A
.
,
N.A
.
Be
l
a
l,
a
n
d
K.M
.
M
a
h
a
r.
Ara
b
ic
S
ig
n
L
a
n
g
u
a
g
e
A
lp
h
a
b
e
t
Rec
o
g
n
it
io
n
Ba
se
d
o
n
H
OG
-
PCA
Us
in
g
M
icr
o
so
ft
Kin
e
c
t
in
C
o
mp
lex
B
a
c
k
g
ro
u
n
d
s
.
i
n
2
0
1
6
IEE
E
6
t
h
I
n
t
e
rn
a
ti
o
n
a
l
Co
n
f
e
re
n
c
e
o
n
A
d
v
a
n
c
e
d
Co
m
p
u
ti
n
g
(IA
CC).
2
0
1
6
.
IEE
E
.
[1
8
]
L
ian
g
,
C.
-
W
.
a
n
d
C.
-
F
.
Ju
a
n
g
,
M
o
v
in
g
o
b
jec
t
c
las
sif
ic
a
ti
o
n
u
si
n
g
lo
c
a
l
sh
a
p
e
a
n
d
HO
G
fe
a
tu
re
s
in
w
a
v
e
let
-
tran
sf
o
r
m
e
d
sp
a
c
e
w
it
h
h
iera
rc
h
ic
a
l
S
VM
c
las
sif
iers
.
Ap
p
li
e
d
S
o
f
t
Co
mp
u
t
in
g
,
2
0
1
5
;
2
8
:
4
8
3
-
4
9
7
.
[1
9
]
L
e
e
,
S
.
-
H.,
e
t
a
l.
An
e
ff
icie
n
t
se
lec
ti
o
n
o
f
HO
G
fea
tu
re
fo
r
S
VM
c
la
ss
if
ica
ti
o
n
o
f
v
e
h
icle
.
in
2
0
1
5
In
tern
a
ti
o
n
a
l
S
y
m
p
o
siu
m
o
n
Co
n
su
m
e
r
El
e
c
tro
n
ics
(IS
CE).
2
0
1
5
.
IEE
E
.
[2
0
]
G
a
l
a
r,
M
.
,
e
t
a
l.
,
A
n
o
v
e
r
v
ie
w
o
f
e
n
se
m
b
le
m
e
th
o
d
s
f
o
r
b
in
a
ry
c
las
sif
ier
s
in
m
u
lt
i
-
c
las
s
p
ro
b
lem
s
:
Ex
p
e
rime
n
tal
stu
d
y
o
n
o
n
e
-
vs
-
o
n
e
a
n
d
o
n
e
-
vs
-
a
ll
sc
h
e
m
e
s.
Pa
tt
e
rn
Rec
o
g
n
i
ti
o
n
,
2
0
1
1
;
4
4
(
8
):
1
7
6
1
-
1
7
7
6
.
[2
1
]
Zh
a
n
g
,
X
.
,
S
.
Din
g
,
a
n
d
T
.
S
u
n
,
M
u
lt
i
-
c
las
s
L
S
T
M
S
V
M
b
a
se
d
o
n
o
p
ti
m
a
l
d
irec
ted
a
c
y
c
li
c
g
ra
p
h
a
n
d
s
h
u
f
f
led
f
ro
g
lea
p
in
g
a
lg
o
rit
h
m
.
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
M
a
c
h
i
n
e
L
e
a
r
n
in
g
a
n
d
Cy
b
e
rn
e
ti
c
s
,
2
0
1
6
;
7
(
2
):
2
4
1
-
2
5
1
.
[2
2
]
Ba
i,
X
.
,
e
t
a
l.
,
L
e
a
rn
in
g
ECOC
c
o
d
e
m
a
tri
x
f
o
r
m
u
lt
icla
ss
c
las
si
f
ica
ti
o
n
w
it
h
a
p
p
li
c
a
ti
o
n
to
g
lau
c
o
m
a
d
iag
n
o
sis.
J
o
u
rn
a
l
o
f
me
d
ica
l
sy
ste
ms
,
2
0
1
6
;
4
0
(
4
):
1
-
1
0
.
[2
3
]
L
u
,
S
iy
u
a
n
,
S
h
u
ih
u
a
W
a
n
g
,
a
n
d
Yu
d
o
n
g
Zh
a
n
g
.
A
n
o
te
o
n
t
h
e
m
a
rk
e
r
-
b
a
se
d
w
a
t
e
rsh
e
d
m
e
th
o
d
f
o
r
X
-
ra
y
i
m
a
g
e
se
g
m
e
n
tatio
n
.
Co
mp
u
ter
M
e
th
o
d
s a
n
d
Pro
g
ra
ms
i
n
Bi
o
me
d
icin
e
,
2
0
1
7
;
1
4
1
:
1
-
2.
[2
4
]
Ra
n
tala
n
k
il
a
,
P
e
k
k
a
,
Ju
h
o
Ka
n
n
a
la,
a
n
d
Esa
Ra
h
tu
.
Ge
n
e
ra
ti
n
g
o
b
jec
t
se
g
me
n
ta
t
io
n
p
r
o
p
o
sa
ls
u
si
n
g
g
l
o
b
a
l
a
n
d
lo
c
a
l
se
a
rc
h
.
In
P
ro
c
e
e
d
i
n
g
s o
f
th
e
IEE
E
c
o
n
f
e
re
n
c
e
o
n
c
o
m
p
u
ter v
is
io
n
a
n
d
p
a
tt
e
rn
re
c
o
g
n
i
ti
o
n
,
2
0
1
4
;
2
4
1
7
-
2
4
2
4
.
[2
5
]
W
a
n
g
,
Zh
o
u
,
A
lan
C.
Bo
v
ik
,
Ha
m
id
R.
S
h
e
ik
h
,
a
n
d
Eero
P
.
S
i
m
o
n
c
e
ll
i.
I
m
a
g
e
q
u
a
li
ty
a
ss
e
ss
m
e
n
t:
f
ro
m
e
rro
r
v
isib
il
it
y
to
stru
c
tu
ra
l
sim
il
a
rit
y
.
I
EE
E
tra
n
sa
c
ti
o
n
s
o
n
ima
g
e
p
ro
c
e
ss
in
g
,
2
0
0
4
;
1
3
(
4
):
6
0
0
-
6
1
2
.
[2
6
]
T
a
n
c
h
e
n
k
o
,
A
le
x
a
n
d
e
r.
V
isu
a
l
-
P
S
NR
m
e
a
su
re
o
f
i
m
a
g
e
q
u
a
li
t
y
.
J
o
u
rn
a
l
o
f
Vi
su
a
l
Co
mm
u
n
ica
t
i
o
n
a
n
d
Ima
g
e
Rep
re
se
n
ta
ti
o
n
,
2
0
1
4
;
25
(
5
)
:
8
7
4
-
8
7
8
.
[2
7
]
P
ra
sh
a
n
th
,
H.
S
.
,
H.
L
.
S
h
a
sh
i
d
h
a
ra
,
a
n
d
Ba
las
u
b
ra
m
a
n
y
a
M
u
rth
y
KN
.
Ima
g
e
sc
a
li
n
g
c
o
m
p
a
riso
n
u
sin
g
u
n
ive
rs
a
l
ima
g
e
q
u
a
l
it
y
in
d
e
x
.
In
A
d
v
a
n
c
e
s
in
Co
m
p
u
ti
n
g
,
Co
n
tr
o
l,
&
T
e
lec
o
m
m
u
n
ica
ti
o
n
T
e
c
h
n
o
lo
g
ies
,
2
0
0
9
.
A
CT
'
0
9
.
I
n
tern
a
ti
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
,
2
0
0
9
;8
5
9
-
8
6
3
.
IE
EE
.
[2
8
]
Ra
o
,
G
.
A
n
a
n
th
,
a
n
d
P
.
V
.
V
.
Kish
o
re
.
"
S
ig
n
L
a
n
g
u
a
g
e
Re
c
o
g
n
it
io
n
S
y
ste
m
S
i
m
u
late
d
f
o
r
V
id
e
o
Ca
p
tu
re
d
w
it
h
S
m
a
rt
P
h
o
n
e
F
ro
n
t
Ca
m
e
ra
.
"
In
te
rn
a
ti
o
n
a
l
Jo
u
rn
a
l
o
f
El
e
c
tri
c
a
l
a
n
d
Co
m
p
u
ter E
n
g
in
e
e
rin
g
(IJECE)
6
,
n
o
.
5
(2
0
1
6
)
:
2
1
7
6
-
2
1
8
7
.
[2
9
]
Ra
o
,
D.
Na
ra
sim
h
a
,
T
.
S
u
re
n
d
ra
,
a
n
d
S
.
T
a
ra
Ka
ly
a
n
i.
"
DP
F
C
P
e
rf
o
r
m
a
n
c
e
w
it
h
th
e
Co
m
p
a
riso
n
o
f
P
I
a
n
d
A
NN
Co
n
tr
o
ll
e
r.
"
In
tern
a
t
io
n
a
l
Jo
u
r
n
a
l
o
f
El
e
c
tri
c
a
l
a
n
d
Co
m
p
u
ter E
n
g
in
e
e
rin
g
(IJECE)
6
,
n
o
.
5
(2
0
1
6
):
2
0
8
0
-
2
0
8
7
.
[3
0
]
T
h
a
n
e
k
a
r,
S
a
c
h
in
A
ru
n
,
K.
S
u
b
ra
h
m
a
n
y
a
m
,
a
n
d
A
.
B.
Ba
g
w
a
n
.
"
Bi
g
Da
ta
a
n
d
M
a
p
Re
d
u
c
e
Ch
a
ll
e
n
g
e
s,
Op
p
o
r
tu
n
it
ies
a
n
d
T
re
n
d
s."
In
tern
a
ti
o
n
a
l
Jo
u
rn
a
l
o
f
El
e
c
tri
c
a
l
a
n
d
Co
m
p
u
ter E
n
g
in
e
e
rin
g
6
,
n
o
.
6
(
2
0
1
6
):
2
9
1
1
.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
K
.
V
.
V
.
K
u
m
a
r
b
o
rn
in
In
d
i
a
1
9
8
7
.
Ob
tain
e
d
h
is
B.
T
e
c
h
,
fro
m
V
RS
&
YRN
Co
ll
e
g
e
o
f
En
g
in
e
e
rin
g
a
n
d
T
e
c
h
n
o
lo
g
y
Ch
irala
.
Du
rin
g
2
0
0
7
-
2
0
1
0
.
&
M
.
T
e
c
h
f
ro
m
K.
L
Un
iv
e
r
sit
y
in
th
e
S
p
e
c
ializa
ti
o
n
o
f
Co
m
m
u
n
ica
ti
o
n
a
n
d
Ra
d
a
r
sy
ste
m
s
d
u
rin
g
2
0
1
0
-
2
0
1
2
.
He
is
h
a
v
in
g
M
o
re
th
a
n
5
y
e
a
rs
t
e
a
c
h
in
g
e
x
p
e
rien
c
e
a
n
d
h
a
v
in
g
8
in
tern
a
ti
o
n
a
l
a
n
d
n
a
ti
o
n
a
l
jo
u
rn
a
ls
/
Co
n
f
e
re
n
c
e
p
a
p
e
rs
L
ik
e
IEE
E
a
n
d
S
P
RING
ER
He
is
A
ss
o
c
iate
m
e
m
b
e
r
o
f
I.
E.
T
.
E
a
n
d
I
S
S
E
o
th
e
r
b
o
d
ies
L
ik
e
I.
S
.
T
.
E.
His
re
se
a
rc
h
in
tere
s
ted
a
re
a
s
in
c
lu
d
e
s
Im
a
g
e
p
ro
c
e
ss
i
n
g
a
n
d
S
ig
n
a
l
P
r
o
c
e
ss
in
g
.
a
n
d
V
id
e
o
P
r
o
c
e
ss
in
g
.
P
.
V.V.Kish
o
re
is
h
a
v
in
g
P
h
.
D
d
e
g
re
e
in
E
lec
tro
n
ics
a
n
d
C
o
m
m
u
n
ica
ti
o
n
s
En
g
in
e
e
rin
g
f
ro
m
A
n
d
h
ra
Un
iv
e
rsit
y
Co
ll
e
g
e
o
f
e
n
g
in
e
e
rin
g
in
2
0
1
3
.
He
re
c
e
i
v
e
d
M
.
T
e
c
h
d
e
g
re
e
f
ro
m
Co
c
h
in
Un
iv
e
rsit
y
o
f
sc
ien
c
e
a
n
d
tec
h
n
o
l
o
g
y
in
th
e
y
e
a
r
2
0
0
3
.
He
re
c
e
iv
e
d
B
.
T
e
c
h
d
e
g
r
e
e
in
e
lec
tro
n
ics
a
n
d
c
o
m
m
u
n
ica
ti
o
n
s
e
n
g
in
e
e
rin
g
f
ro
m
J
NT
U,
H
y
d
.
in
2
0
0
0
.
He
i
s
c
u
rre
n
tl
y
f
u
ll
p
ro
f
e
ss
o
r
a
n
d
Im
a
g
e
,
si
g
n
a
l
a
n
d
sp
e
e
c
h
p
ro
c
e
ss
i
n
g
Re
se
a
rc
h
He
a
d
a
t
K.
L
.
Un
iv
e
r
sity
,
ECE
De
p
a
rt
m
e
n
t.
He
is
c
u
rre
n
tl
y
e
n
g
a
g
e
d
in
a
p
ro
jec
t
sp
o
n
se
re
d
b
y
De
p
a
rt
m
e
n
t
o
f
S
c
ien
c
e
a
n
d
T
e
c
h
n
o
lo
g
y
,
S
EE
D,
G
o
v
e
rn
m
e
n
t
o
f
In
d
ia,
u
n
d
e
r
t
h
e
T
e
c
h
n
o
lo
g
y
f
o
r
Disa
b
led
a
n
d
El
d
e
rly
s
c
h
e
m
e
,
re
late
d
to
sig
n
lan
g
u
a
g
e
3
D
in
v
a
rien
t
m
o
d
e
ls
f
o
r
In
d
ian
sig
n
lan
g
u
a
g
e
.
His
re
se
a
rc
h
in
tere
sts
a
re
d
ig
it
a
l
si
g
n
a
l
a
n
d
im
a
g
e
p
ro
c
e
ss
in
g
,
A
rti
fi
c
ial
In
telli
g
e
n
c
e
a
n
d
h
u
m
a
n
o
b
jec
t
i
n
tera
c
ti
o
n
s.
He
is
c
u
rre
n
t
ly
a
m
e
m
b
e
r
o
f
IEE
E.
He
h
a
s
p
u
b
li
s
h
e
d
7
0
re
se
a
rc
h
p
a
p
e
rs
i
n
V
a
rio
u
s
Na
ti
o
n
a
l
a
n
d
In
tern
a
t
io
n
a
l
jo
u
r
n
a
ls
a
n
d
c
o
n
f
e
re
n
c
e
s in
c
lu
d
i
n
g
IEE
E,
S
p
rin
g
e
r
a
n
d
El
se
v
ier.
Evaluation Warning : The document was created with Spire.PDF for Python.