I
nte
rna
t
io
na
l J
o
urna
l o
f
I
nfo
r
m
a
t
ics a
nd
Co
mm
u
n
ica
t
io
n T
ec
hn
o
lo
g
y
(
I
J
-
I
CT
)
Vo
l.
7
,
No
.
3
,
Dec
em
b
er
201
8
,
p
p
.
1
4
1
~
1
4
5
I
SS
N:
2252
-
8776
,
DOI
: 1
0
.
1
1
5
9
1
/i
j
i
ct.
v
7
i3
.
p
p
1
4
1
-
145
141
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ia
e
s
co
r
e.
co
m/jo
u
r
n
a
ls
/in
d
ex
.
p
h
p
/
I
JI
C
T
Address
ing
La
ten
cy
Iss
ues in 2D
to
3
D Conv
ersio
n:
D
eplo
y
ing
Av
a
ila
ble Synthet
ic Data
ba
se
N.
L
.
M
a
na
s
a
G
NI
T
S
,
JN
T
UH
,
H
y
d
e
ra
b
a
d
,
In
d
i
a
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
u
n
2
0
,
2
0
1
8
R
ev
i
s
ed
A
u
g
6
,
2
0
1
8
A
cc
ep
ted
A
u
g
2
0
,
2
0
1
8
Co
n
v
e
n
ti
o
n
a
l
2
D
to
3
D
re
n
d
e
rin
g
tec
h
n
iq
u
e
s
i
n
v
o
lv
e
a
se
q
u
e
n
ti
a
l
p
ro
c
e
ss
o
f
g
ro
u
p
i
n
g
o
f
th
e
in
p
u
t
im
a
g
e
s
b
a
se
d
o
n
e
d
g
e
in
f
o
rm
a
ti
o
n
a
n
d
p
re
d
ictiv
e
a
lg
o
rit
h
m
s
to
a
ss
ig
n
d
e
p
t
h
v
a
lu
e
s
to
p
ix
e
ls
w
i
th
sa
m
e
h
u
e
.
T
h
e
it
e
ra
ti
v
e
c
a
lcu
latio
n
s
a
n
d
v
o
l
u
m
e
o
f
d
a
ta
u
n
d
e
r
sc
ru
ti
n
y
to
a
ss
ig
n
‘re
a
l
-
ti
m
e
’
v
a
lu
e
s
ra
ise
late
n
c
y
is
su
e
s
a
n
d
c
o
st
c
o
n
sid
e
ra
ti
o
n
s.
F
o
r
c
o
m
m
e
r
c
ial
c
o
n
su
m
p
ti
o
n
,
w
h
e
re
sp
e
e
d
a
n
d
a
c
c
u
ra
c
y
d
e
f
in
e
th
e
v
iab
il
it
y
o
f
a
p
ro
d
u
c
t,
th
e
re
is
a
n
e
e
d
to
re
o
rien
t
th
e
a
p
p
ro
a
c
h
u
se
d
in
th
e
p
re
se
n
t
m
e
th
o
d
o
l
o
g
ies
.
In
p
re
d
ictiv
e
m
e
th
o
d
o
lo
g
ies
o
n
e
o
f
th
e
c
o
re
in
tere
sts
is
a
c
h
iev
in
g
th
e
in
it
ial
a
p
p
ro
x
im
a
ti
o
n
a
s
c
lo
se
to
th
e
‘re
a
l’
v
a
lu
e
a
s
p
o
ss
ib
le.
In
th
is
w
o
rk
,
‘s
y
n
th
e
ti
c
’
d
a
tab
a
se
is
u
se
d
to
p
r
o
v
id
e
th
e
f
irst
a
p
p
ro
x
i
m
a
ti
o
n
th
r
o
u
g
h
c
o
m
p
a
riso
n
tec
h
n
iq
u
e
s
a
n
d
f
e
d
to
th
e
p
re
d
ictiv
e
t
o
o
l.
It
is
b
e
l
iev
e
d
th
a
t
t
h
is
w
o
rk
w
il
l
p
ro
v
id
e
a
b
a
sis
f
o
r
d
e
v
e
lo
p
in
g
a
n
e
f
f
icie
n
t
2
D t
o
3
D
c
o
n
v
e
rsio
n
m
e
th
o
d
o
lo
g
y
.
K
ey
w
o
r
d
s
:
3
D
g
r
ap
h
ical
r
en
d
er
in
g
3
D
s
ter
eo
s
co
p
ic
v
id
eo
s
Dep
th
p
er
ce
p
tio
n
Gau
s
s
ia
n
Mix
t
u
r
e
m
o
d
el
Co
p
y
rig
h
t
©
2
0
1
8
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
:
N.
L
.
Ma
n
a
s
a
,
GNI
T
S,
J
NT
UH,
Hy
d
er
ab
ad
,
I
n
d
ia
E
m
ail:
Nad
ip
all
y
.
m
@
g
m
ail.
co
m
; N
ad
ip
all
y
.
m
@
g
n
i
ts
.
ac
.
in
.
1.
I
NT
RO
D
UCT
I
O
N
I
n
th
e
la
s
t
d
ec
ad
e,
th
e
n
ee
d
an
d
co
n
s
eq
u
e
n
t
d
e
m
a
n
d
f
o
r
3
D
(
3
Dim
e
n
s
io
n
al)
i
m
ag
e
s
an
d
v
id
eo
s
h
a
s
g
r
o
w
n
o
w
i
n
g
to
it
s
u
ti
lit
y
in
d
if
f
er
e
n
t
f
ield
s
,
s
u
ch
a
s
m
ed
ical
an
d
en
g
i
n
ee
r
in
g
d
iag
n
o
s
t
ics
an
d
r
esto
r
ativ
e
in
ter
v
e
n
tio
n
s
[
1
]
.
I
n
ad
d
itio
n
,
th
e
en
t
er
tain
m
e
n
t
in
d
u
s
tr
y
u
s
es
i
t
to
ad
d
t
h
e
‘
r
ea
l
f
ee
l
’
f
ac
to
r
f
o
r
i
m
p
r
o
v
ed
co
n
s
u
m
p
tio
n
.
A
s
s
u
c
h
,
th
e
n
e
ed
f
o
r
f
aster
co
n
v
er
s
io
n
tech
n
iq
u
es
o
f
co
n
v
e
n
tio
n
al
2
D
(
2
Di
m
e
n
s
io
n
al)
to
3
D
i
m
a
g
er
y
h
as
g
ain
ed
w
id
er
atte
n
tio
n
in
d
if
f
er
e
n
t
i
n
d
u
s
tr
ies.
A
lt
h
o
u
g
h
d
i
f
f
er
en
t
a
p
p
r
o
ac
h
e
s
h
a
v
e
b
ee
n
tr
ied
to
m
o
d
u
late
p
la
n
ar
i
m
a
g
e
s
to
3
D
i
m
a
g
es
f
o
r
b
etter
v
i
s
u
al
ex
p
er
ien
ce
to
co
n
s
id
er
ab
le
ex
te
n
t,
t
h
er
e
s
till
e
x
i
s
t
s
n
ee
d
f
o
r
b
etter
q
u
alit
y
t
h
at
is
b
o
th
,
f
aster
a
n
d
co
m
m
er
ciall
y
v
iab
le.
Am
o
n
g
s
t
th
e
p
r
i
n
cip
al
co
n
ce
r
n
s
i
n
v
o
lv
ed
in
th
e
co
n
v
er
s
io
n
o
f
p
lan
ar
,
2
D
i
m
a
g
es
to
3
D
v
i
s
u
al
s
is
to
r
ec
r
ea
te
th
e
‘
d
ep
t
h
’
an
d
‘
co
lo
r
’
(
h
u
e)
p
ar
a
m
eter
f
r
o
m
t
h
e
o
r
ig
in
a
l
i
m
ag
e
s
.
So
m
e
o
f
t
h
e
n
o
tab
le
an
d
w
id
el
y
u
s
ed
tec
h
n
iq
u
e
s
i
n
t
h
e
p
u
r
s
u
it
o
f
g
en
er
ati
n
g
t
h
e
‘
d
ep
th
asp
ec
t
f
r
o
m
co
n
v
e
n
tio
n
al
2
D
i
m
a
g
es
in
c
lu
d
e
s
ter
eo
s
co
p
ic
v
is
io
n
(
to
allo
w
f
o
r
tr
ian
g
u
latio
n
)
,
ac
ti
v
e
(
d
y
n
a
m
ic)
m
et
h
o
d
s
,
an
d
3
D
g
r
ap
h
ical
(
al
g
o
r
ith
m
ic
,
s
o
f
t
w
ar
e
-
b
ased
)
r
en
d
er
in
g
tec
h
n
iq
u
es.
W
h
ile
t
h
e
f
ir
s
t
t
w
o
t
ec
h
n
iq
u
es
m
en
tio
n
ed
(
s
ter
eo
s
co
p
ic
im
a
g
er
y
a
n
d
ac
tiv
e
s
e
n
s
o
r
s
r
eq
u
ir
e
s
p
ec
if
ic
,
an
d
co
n
te
m
p
o
r
ar
il
y
h
i
g
h
-
co
s
t
eq
u
ip
m
e
n
t
to
h
elp
ass
e
s
s
a
n
d
r
en
d
er
th
e
‘
d
ep
th
p
ar
am
eter
,
t
h
e
t
h
ir
d
(
3
D
g
r
ap
h
ical
r
en
d
er
in
g
)
tec
h
n
iq
u
e
i
n
v
o
l
v
es
t
i
m
e
-
co
n
s
u
m
in
g
iter
a
tiv
e
s
eq
u
en
ce
.
T
h
e
s
ter
eo
s
co
p
ic
m
et
h
o
d
s
r
eq
u
ir
e
m
u
ltip
le
ca
m
er
as
to
ca
p
tu
r
e
2
D
i
m
a
g
es
a
n
d
th
er
eb
y
a
s
s
e
s
s
t
h
e
d
ep
th
r
eq
u
ir
ed
to
cr
ea
te
th
e
v
is
u
all
y
co
m
p
lia
n
t
3
D
v
er
s
io
n
.
I
n
th
e
ac
ti
v
e
i
m
ag
e
ca
p
t
u
r
in
g
m
et
h
o
d
s
,
p
h
o
to
g
r
ap
h
er
s
an
d
v
id
eo
g
r
ap
h
er
s
u
s
e
d
i
f
f
er
e
n
t
s
en
s
o
r
s
s
u
c
h
as
ti
m
e
-
of
-
f
li
g
h
t
an
d
s
tr
u
ct
u
r
ed
lig
h
t
ca
p
ab
ili
ties
to
esti
m
ate
t
h
e
d
ep
th
f
ac
to
r
o
f
2
D
i
m
a
g
e
s
.
S
u
ch
r
ec
o
r
d
s
ar
e
n
o
t
co
m
m
er
cial
l
y
f
ea
s
ib
le[
2
]
o
w
i
n
g
to
s
k
il
ls
r
eq
u
ir
ed
to
o
p
er
ate
s
p
ec
iall
y
d
ev
is
ed
eq
u
ip
m
en
t
an
d
th
e
ec
o
n
o
m
ic
f
ea
s
ib
ili
t
y
ap
ar
t
f
r
o
m
ac
ce
s
s
to
s
u
c
h
eq
u
ip
m
e
n
t.
T
h
e
th
ir
d
m
et
h
o
d
o
f
g
r
ap
h
ical
r
e
n
d
er
in
g
tec
h
n
iq
u
es
u
s
ed
co
n
v
en
tio
n
all
y
i
s
u
s
u
all
y
a
s
k
illed
a
n
d
e
x
p
er
ien
tial
m
a
n
u
al
p
r
o
ce
s
s
r
eq
u
ir
in
g
ted
io
u
s
a
n
d
ti
m
e
-
co
n
s
u
m
i
n
g
en
g
a
g
e
m
en
t
,
th
er
eb
y
m
ak
in
g
it
u
n
f
ea
s
ib
l
e
f
o
r
co
m
m
er
cial,
lar
g
e
-
s
ca
le
co
n
s
u
m
p
t
io
n
.
O
f
late,
th
er
e
is
a
g
r
o
w
i
n
g
ac
ce
n
t
o
n
d
ev
e
lo
p
in
g
au
to
m
ated
p
r
o
ce
s
s
es
t
h
at
ca
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8776
IJ
-
I
C
T
Vo
l.
7
,
No
.
3
,
Dec
em
b
er
20
1
8
:
1
41
–
1
45
142
co
n
v
er
t
2
D
i
m
ag
e
s
to
3
D
v
i
s
u
als
b
y
in
teg
r
ati
n
g
t
h
e
m
i
s
s
i
n
g
th
ir
d
‘
d
ep
th
’
p
ar
a
m
eter
in
th
e
p
lan
ar
i
m
a
g
es
o
b
tain
ed
in
co
n
v
e
n
tio
n
al
p
h
o
t
o
/v
id
eo
.
I
n
o
r
d
er
to
i
m
p
r
o
v
e
th
e
q
u
ali
t
y
o
f
3
D
r
e
n
d
er
in
g
,
d
esi
g
n
er
s
ar
e
u
s
i
n
g
d
i
f
f
er
e
n
t
p
r
ed
icti
v
e
tech
n
iq
u
es
an
d
d
ev
elo
p
in
g
al
g
o
r
ith
m
s
t
h
at
w
o
r
k
f
aster
an
d
m
o
r
e
ac
cu
r
atel
y
an
d
ap
p
r
o
ac
h
‘
r
ea
l
-
t
i
m
e
’
ex
p
er
ien
ce
s
o
u
g
h
t
b
y
co
n
s
u
m
er
s
i
n
d
if
f
er
en
t
ap
p
licatio
n
s
.
T
h
e
h
u
m
an
e
y
e
es
ti
m
atio
n
o
f
d
ep
th
i
s
ac
h
iev
ed
th
r
o
u
g
h
co
m
b
i
n
ed
ef
f
ec
t
an
d
s
u
p
er
p
o
s
itio
n
o
f
b
in
o
cu
lar
an
d
m
o
n
o
cu
lar
d
ep
th
p
er
ce
p
tio
n
[
3
]
.
W
h
er
ea
s
,
in
th
e
b
in
o
cu
lar
v
is
u
al,
h
u
m
a
n
s
ar
e
ab
le
to
d
ec
o
d
e
th
e
th
ir
d
d
i
m
e
n
s
io
n
th
r
o
u
g
h
co
n
v
er
g
e
n
ce
o
f
d
i
f
f
er
e
n
tial
e
le
m
e
n
ts
f
r
o
m
b
o
th
v
is
io
n
s
a
n
d
co
r
r
ec
tl
y
e
s
ti
m
at
e
th
e
d
ep
th
,
th
e
m
o
n
o
cu
lar
p
ar
am
eter
s
ar
e
also
ca
p
ab
le
o
f
r
ep
r
o
d
u
cin
g
th
e
d
esire
d
in
f
o
r
m
atio
n
u
s
in
g
s
i
m
ilar
p
r
o
p
er
ties
an
d
in
d
u
ctin
g
t
h
e
d
if
f
er
en
tial
s
.
Su
c
h
cu
e
s
ar
e:
d
y
n
a
m
ic
p
ar
allax
,
p
lan
ar
p
r
o
p
o
r
tio
n
al
m
ea
s
u
r
es
,
th
e
r
elativ
e
te
x
t
u
r
e
ch
a
n
g
e
s
,
an
d
f
o
ca
l
d
is
ta
n
ce
s
.
T
h
e
i
m
p
licatio
n
o
f
t
h
e
d
is
cu
s
s
io
n
i
s
t
h
at
m
o
n
o
cu
lar
i
m
a
g
es/
v
id
eo
s
ca
n
b
e
u
s
ed
to
p
r
o
d
u
ce
3
D
v
is
u
al
s
.
E
x
p
er
i
m
en
tal
s
u
p
p
o
r
t
an
d
p
r
ac
tical
ap
p
licatio
n
o
f
V
is
u
al
p
ar
am
e
ter
s
a
n
d
d
is
p
ar
ities
th
a
t p
r
o
v
id
e
th
e
m
e
n
tio
n
ed
c
u
es
h
av
e
b
ee
n
p
r
o
v
id
ed
b
y
d
i
f
f
er
en
t st
u
d
ies to
g
en
er
at
e
3
D
f
r
o
m
co
n
v
e
n
tio
n
al
i
m
ag
e
s
.
T
h
e
v
ar
io
u
s
m
et
h
o
d
s
th
at
h
a
v
e
atte
m
p
ted
2
D
to
3
D
r
en
d
er
in
g
u
s
e
t
w
o
m
ain
t
y
p
e
s
:
Si
n
g
le
-
f
r
a
m
e
Me
th
o
d
an
d
M
u
ltip
le
-
f
r
a
m
e
Me
th
o
d
s
.
No
tab
le
a
m
o
n
g
Si
n
g
le
-
f
r
a
m
e
Me
t
h
o
d
s
is
th
e
a
tte
m
p
t
b
y
B
attiato
[
4
]
w
h
o
m
ad
e
u
s
e
o
f
m
u
ltip
le
cu
es,
a
p
r
o
g
r
es
s
io
n
o
v
er
s
o
m
e
o
th
er
m
eth
o
d
s
th
a
t
co
n
ce
n
tr
at
ed
o
n
s
in
g
le
cu
e
s
–
f
o
ca
l
p
r
o
x
i
m
it
y
[
5
]
,
m
a
ch
i
n
e
lear
n
in
g
al
g
o
r
ith
m
s
[
6
-
7
]
,
an
d
2
D
p
r
o
p
o
r
tio
n
al
d
i
m
en
s
io
n
s
(
h
eig
h
t
an
d
w
id
t
h
r
atio
)
an
d
tex
tu
r
e
ch
an
g
es
[
8
]
an
d
g
eo
m
etr
ic
m
ea
s
u
r
e
m
en
ts
.
A
s
i
g
n
i
f
ica
n
t
d
ev
elo
p
m
en
t
in
Mu
lti
-
f
r
a
m
e
Me
th
o
d
is
th
e
s
u
cc
es
s
f
u
l
h
y
b
r
id
izatio
n
o
f
MT
D
(
Mo
d
if
ie
d
T
im
e
Di
f
f
er
en
ce
)
a
n
d
C
I
D
(
C
o
m
p
u
ted
I
m
a
g
e
Dep
th
)
ac
h
ie
v
ed
b
y
I
i
n
u
m
a
et
al.
,
[
9
]
th
at
u
s
es
tr
ian
g
u
latio
n
s
u
cc
e
s
s
f
u
ll
y
to
g
e
n
er
ate
clo
s
e
to
r
ea
l
3
D
im
a
g
es.
R
ec
en
t
ap
p
r
o
ac
h
es
[
1
8
]
-
[
2
0
]
th
at
s
h
o
w
ed
co
m
m
e
n
d
ab
le
r
esu
lt
s
w
h
en
u
til
ized
in
a
d
i
f
f
er
en
t
ap
p
lica
tio
n
b
u
t
ca
n
al
s
o
b
e
d
ep
lo
y
ed
to
s
o
lv
e
f
o
r
2
D
to
3
D
r
e
n
d
er
in
g
.
S
ter
eo
s
co
p
ic
f
o
o
tag
e
a
n
d
d
ep
t
h
a
s
s
es
s
m
en
t
f
r
o
m
m
o
tio
n
p
r
o
v
id
e
t
h
e
r
eq
u
ir
ed
d
ata
f
o
r
tr
ian
g
u
latio
n
[
1
0
–
1
3
]
.
An
al
y
s
i
s
o
f
Mo
tio
n
p
ar
allax
o
f
f
er
s
th
e
te
m
p
o
r
al
co
r
r
esp
o
n
d
en
ce
b
r
o
u
g
h
t
ab
o
u
t
b
y
t
h
e
e
x
is
tin
g
d
i
s
p
ar
it
y
.
T
h
e
m
o
tio
n
v
ec
to
r
(
MV
)
th
u
s
e
x
t
r
ac
ted
is
co
n
v
er
ted
to
d
is
p
ar
ity
v
ec
to
r
(
DV)
.
I
n
an
alter
n
ate
m
et
h
o
d
,
th
e
d
ep
th
m
ap
is
n
o
t
p
r
o
d
u
ce
d
,
in
s
tead
a
s
ter
e
o
s
co
p
ic
v
is
io
n
co
m
p
r
is
in
g
o
f
r
ig
h
t
-
e
y
e
an
d
th
e
lef
t
e
y
e
v
i
s
io
n
.
Ho
w
ev
er
,
s
u
c
h
m
et
h
o
d
s
r
eq
u
ir
e
th
e
d
ep
lo
y
m
e
n
t
o
f
m
o
tio
n
ca
m
er
a.
As
s
tated
ea
r
lier
,
a
co
m
b
in
at
io
n
o
f
m
o
n
o
cu
lar
an
d
b
in
o
cu
lar
s
tr
ate
g
ie
s
ap
p
ea
r
to
h
av
e
th
e
p
o
ten
tia
l
to
d
eliv
er
th
e
d
esire
d
q
u
alit
y
ef
f
i
cien
tl
y
.
1
.
1
Sin
g
le
-
f
ra
m
e
M
et
ho
ds
P
ar
k
et
al.
,
[
5
]
laid
em
p
h
as
is
o
n
th
e
clar
it
y
co
n
te
n
t
o
f
d
if
f
er
en
t
ar
ea
s
m
ea
s
u
r
ed
b
y
t
h
e
d
is
t
an
ce
f
r
o
m
th
e
f
o
ca
l
p
lan
e
to
esti
m
ate
t
h
e
r
elati
v
e
d
ep
th
o
f
ele
m
e
n
t
s
in
th
e
2
D
i
m
a
g
e.
T
h
e
i
m
ag
e
is
d
iv
id
ed
i
n
to
r
ec
tan
g
u
lar
b
lo
ck
s
t
h
at
f
o
r
m
a
m
atr
i
x
a
n
d
th
e
b
lu
r
r
i
n
ess
o
f
ea
ch
b
lo
ck
i
s
co
n
s
id
er
ed
to
es
ti
m
ate
th
e
d
ep
th
o
f
th
at
p
ar
t
o
f
th
e
i
m
a
g
e
f
r
o
m
t
h
e
v
ie
w
er
.
T
h
e
lim
itat
io
n
o
f
s
u
ch
esti
m
atio
n
i
s
clea
r
l
y
th
e
n
u
m
b
er
o
f
d
is
ce
r
n
ib
le
b
lo
ck
s
th
a
t
ca
n
b
e
ac
co
m
m
o
d
ated
lead
in
g
,
i
n
t
u
r
n
to
i
n
co
n
s
is
ten
c
y
i
n
r
ea
d
-
o
f
f
v
al
u
es
a
m
o
n
g
s
t
n
o
n
-
ad
j
ac
en
t
p
atch
es
o
f
s
a
m
e
s
h
ad
es,
a
n
d
t
h
e
ac
cu
r
ac
y
o
f
th
e
m
ea
s
u
r
i
n
g
d
ev
ice
to
m
ea
s
u
r
e
g
r
ad
u
all
y
ch
an
g
i
n
g
h
u
es
th
at
m
a
y
v
ar
y
b
y
i
n
f
in
i
tesi
m
all
y
s
m
all
a
m
o
u
n
ts
.
T
h
is
m
et
h
o
d
i
s
also
ca
lled
th
e
C
I
D
(
C
o
m
p
u
ted
I
m
a
g
e
Dep
th
)
.
T
h
e
m
ac
h
i
n
e
lear
n
in
g
m
et
h
o
d
o
lo
g
y
w
as
u
s
ed
b
y
Ho
ie
m
[
7
]
w
h
o
m
ap
p
ed
th
e
i
m
a
g
e
to
as
s
ess
t
h
e
d
ep
th
a
n
d
th
en
ad
d
ed
th
e
ex
tr
ac
ted
p
ar
a
m
eter
to
th
e
2
D
i
m
a
g
e
to
o
b
tain
th
e
3
D
v
is
u
al.
I
n
m
ac
h
in
e
lear
n
i
n
g
m
et
h
o
d
o
lo
g
y
,
t
h
e
o
b
j
ec
ts
an
d
d
is
tan
ce
s
ar
e
a
s
s
i
g
n
ed
p
r
o
g
r
ess
i
v
el
y
t
h
r
o
u
g
h
tr
ai
n
ed
cl
ass
es,
a
n
d
p
o
s
e
th
e
p
o
ten
tial
s
k
ip
p
in
g
o
f
ca
s
es
i
n
th
e
in
itial
tr
ain
in
g
p
h
ase
ca
u
s
i
n
g
p
er
ce
p
tib
le
in
ac
cu
r
ac
y
in
t
h
e
f
i
n
al
r
en
d
itio
n
,
as
t
h
e
p
r
o
g
r
ess
iv
e
s
ta
g
es
d
ep
en
d
h
ea
v
il
y
o
n
t
h
e
p
r
ec
ed
in
g
f
ee
d
.
T
h
u
s
,
m
is
s
ed
ca
s
es
ca
n
l
ea
d
to
u
n
ac
ce
p
tab
le
d
is
r
u
p
tio
n
f
r
o
m
t
h
e
d
esire
d
o
u
tco
m
e.
T
h
e
g
eo
m
etr
ic
m
et
h
o
d
u
s
ed
b
y
T
s
ai
[
1
4
]
em
p
lo
y
s
t
h
e
‘
v
a
n
i
s
h
in
g
p
o
in
t
’
co
n
ce
p
t
to
ad
d
th
e
th
ir
d
d
i
m
e
n
s
io
n
f
r
o
m
s
i
n
g
le
f
r
a
m
e
s
.
J
u
n
g
et
al.
,
[
8
]
u
s
ed
t
h
e
i
n
f
o
r
m
ati
o
n
ac
cr
u
ed
f
r
o
m
th
e
ed
g
es
an
d
tr
ied
t
o
co
n
s
o
lid
ate
th
e
p
r
o
ce
s
s
w
it
h
th
e
h
e
lp
o
f
k
n
o
w
n
d
ata
(
h
u
e
a
g
ain
s
t
d
ep
th
)
to
esti
m
ate
th
e
r
eq
u
ir
ed
q
u
an
tit
y
(
d
ep
th
)
-
p
er
h
ap
s
a
f
ir
s
t
s
tep
to
w
ar
d
s
p
r
ed
ictiv
e
ex
tr
ac
tio
n
.
T
h
e
ab
o
v
e
m
et
h
o
d
s
m
a
y
b
e
s
u
cc
e
s
s
f
u
ll
y
d
ep
lo
y
ed
in
ce
r
ta
in
s
p
ec
i
f
ic
ca
s
es
in
d
i
v
id
u
all
y
an
d
li
m
i
ted
to
s
h
o
r
ter
d
ep
th
esti
m
atio
n
i
n
i
m
ag
e
s
u
s
ed
.
L
o
s
s
o
f
o
th
er
p
ar
a
m
etr
ic
v
alu
e
s
t
h
an
t
h
e
o
n
es
u
n
d
er
co
n
s
id
er
atio
n
m
a
k
es
t
h
e
m
eth
o
d
s
u
s
in
g
s
i
n
g
le
p
ar
am
eter
s
p
r
o
n
e
to
o
v
er
lo
o
k
in
g
o
t
h
er
s
ig
n
i
f
ica
n
t
p
r
o
p
er
ties
th
at
ad
d
to
th
e
d
ep
th
p
er
ce
p
tio
n
.
C
o
n
s
eq
u
e
n
t
l
y
,
th
e
g
e
n
er
ated
3
D
i
m
a
g
es a
r
e
m
o
s
t li
k
el
y
to
s
u
f
f
er
f
r
o
m
lo
s
s
o
f
te
m
p
o
r
al
co
h
er
en
ce
an
d
s
p
ik
i
n
g
.
1
.
2
M
ulti
-
f
r
a
m
e
M
et
ho
ds
T
h
e
w
o
r
k
o
n
2
D
to
3
D
co
n
v
er
s
io
n
u
s
in
g
Mu
l
ti
-
f
r
a
m
es
h
a
s
t
o
co
n
ten
d
w
it
h
t
h
e
is
s
u
e
o
f
d
e
p
th
w
it
h
i
n
an
o
b
j
ec
t
in
t
h
e
f
r
a
m
e
s
a
n
d
r
ela
tio
n
al
d
ep
th
in
f
o
r
m
at
io
n
o
f
all
o
b
j
ec
ts
i
n
t
h
e
f
r
a
m
e
s
.
A
m
et
h
o
d
o
lo
g
y
o
f
g
r
o
u
p
in
g
o
f
p
ix
el
s
w
it
h
s
a
m
e
h
u
e
an
d
s
p
atia
l
o
r
ien
tatio
n
i
s
t
h
e
b
asic
n
ee
d
f
o
r
o
b
tain
in
g
d
e
s
ir
ed
u
n
if
o
r
m
it
y
i
n
d
ep
th
esti
m
a
tio
n
.
So
m
e
ea
r
lie
r
w
o
r
k
s
h
av
e
u
s
ed
m
o
tio
n
p
ar
allax
as
t
h
e
p
r
i
m
ar
y
cu
e
to
all
o
w
f
o
r
g
r
o
u
p
i
n
g
o
f
p
ix
els
w
it
h
in
t
h
e
f
r
a
m
e
s
.
Ho
w
e
v
er
,
s
o
m
e
p
i
x
els
ca
r
r
y
d
i
f
f
er
en
t
s
el
f
MV
’
s
lead
i
n
g
to
er
r
o
r
-
p
r
o
n
e
d
ep
th
in
f
o
r
m
atio
n
.
T
h
u
s
th
e
co
n
te
n
tio
n
s
ee
m
s
to
r
ev
o
lv
e
ar
o
u
n
d
th
e
m
et
h
o
d
o
lo
g
y
n
ee
d
ed
to
g
r
o
u
p
p
ix
els
d
ep
th
v
alu
e
s
to
p
r
ed
ict
th
e
d
ep
th
o
f
in
d
iv
id
u
al
o
b
j
ec
ts
an
d
f
o
r
r
ela
tio
n
al
d
ep
th
esti
m
atio
n
.
T
h
is
p
ap
er
atten
d
s
to
th
e
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
I
C
T
I
SS
N:
2252
-
8776
A
d
d
r
ess
in
g
La
ten
cy
I
s
s
u
es in
2
D
to
3
D
C
o
n
ve
r
s
io
n
:
Dep
lo
y
in
g
A
va
ila
b
le
S
yn
th
etic
…
(
N
.
L.
Ma
n
a
s
a
)
143
is
s
u
e
o
f
d
ep
th
p
r
ed
ictio
n
t
h
r
o
u
g
h
a
n
o
v
el
s
c
h
e
m
e
o
f
r
e
f
er
e
n
tial
d
ep
th
a
s
s
i
g
n
m
e
n
t
p
r
ese
n
ted
in
th
e
f
o
llo
w
i
n
g
s
ec
tio
n
.
Su
c
h
in
itializatio
n
v
a
lu
e
h
elp
s
ac
ce
ler
ate
t
h
e
f
o
llo
w
-
u
p
iter
ati
v
e
e
s
ti
m
atio
n
p
r
o
ce
s
s
.
E
x
p
er
i
m
e
n
tal
r
esu
lt
m
ea
s
u
r
e
m
e
n
t
an
d
f
ee
d
b
ac
k
f
r
o
m
co
n
s
u
m
er
s
o
f
t
h
e
v
i
d
eo
f
ee
d
s
u
p
p
o
r
ts
th
e
co
n
ten
ti
o
n
th
at
th
e
o
u
tp
u
t
is
v
is
u
all
y
p
leas
in
g
a
n
d
co
m
p
ar
ab
le
to
s
y
n
t
h
etic
v
id
eo
-
g
a
m
e
ex
p
er
ien
ce
,
w
h
ic
h
i
s
it
s
elf
r
at
ed
a
s
v
er
y
clo
s
e
to
‘
r
ea
l’
.
Sect
io
n
2
p
r
esen
ts
th
e
p
r
o
p
o
s
ed
s
y
s
te
m
.
Sectio
n
3
p
r
esen
t
s
th
e
o
u
tco
m
e
o
f
th
e
e
x
p
er
i
m
en
tal
r
ig
o
r
an
d
d
is
cu
s
s
es
t
h
e
co
m
p
lex
it
y
an
d
ac
cu
r
ac
y
as
co
m
p
ar
ed
to
o
th
e
r
co
n
ten
tio
n
s
r
eg
ar
d
i
n
g
d
ep
th
esti
m
atio
n
.
Fi
n
all
y
s
ec
tio
n
4
an
d
5
s
u
m
m
ar
iz
e
th
e
p
ap
er
an
d
s
u
g
g
e
s
ts
p
o
s
s
ib
le
ar
ea
s
o
f
f
u
r
th
er
w
o
r
k
.
2.
P
RO
P
O
SE
D
SYS
T
E
M
C
o
m
m
er
cial
v
iab
ilit
y
a
n
d
ti
m
e
co
n
s
u
m
ed
to
ad
d
th
e
d
ep
th
p
er
ce
p
tio
n
th
r
o
u
g
h
iter
ati
v
e
p
r
ed
ictio
n
s
ar
e
n
o
t
p
r
ac
tical
w
h
e
n
s
tr
ea
m
in
g
li
v
e
e
v
e
n
ts
.
Mo
s
t
o
f
t
h
e
‘
p
ix
el
g
r
o
u
p
in
g
’
m
et
h
o
d
o
lo
g
ies
s
ee
m
to
s
er
v
e
a
t
b
est
ac
ad
em
ic
i
n
ter
est
s
.
T
h
o
u
g
h
s
u
ch
m
et
h
o
d
o
lo
g
ies
o
f
3
D
r
en
d
er
in
g
ar
e
p
o
ten
tiall
y
v
er
y
clo
s
e
to
‘
r
ea
l
’
p
er
ce
p
tio
n
o
f
o
b
s
er
v
atio
n
s
to
th
e
h
u
m
a
n
e
y
e,
t
h
e
ti
m
e
la
g
b
et
w
ee
n
i
m
a
g
e
ca
p
tu
r
e
an
d
f
i
n
al
o
u
tp
u
t
d
eli
v
er
ed
f
o
r
co
n
s
u
m
p
tio
n
i
s
r
ar
el
y
w
it
h
in
ac
ce
p
tab
le
li
m
its
.
T
h
e
s
ta
g
es
an
d
s
eq
u
en
ce
(
s
tar
ti
n
g
f
r
o
m
lef
t:
lef
t
g
r
ad
ien
t,
lef
t
b
o
tto
m
,
b
o
tto
m
-
u
p
,
r
ig
h
t
b
o
tto
m
,
r
i
g
h
t)
o
f
h
u
e
g
r
ad
in
g
to
ass
es
s
d
ep
th
t
h
at
co
n
v
er
ts
2
D
to
3
D
u
s
u
all
y
f
o
llo
w
s
t
h
e
s
eq
u
e
n
ce
d
ep
icted
in
F
i
g
u
r
e
1
.
Fig
u
r
e
1
.
[
1
5
]
T
h
e
p
r
o
p
o
s
ed
s
y
s
te
m
atte
m
p
t
s
to
i
m
p
r
o
v
e
th
e
esti
m
atio
n
p
r
o
ce
s
s
b
y
u
s
in
g
e
x
i
s
ti
n
g
r
ef
er
e
n
tial
d
ep
t
h
cu
e
i
n
s
tead
o
f
tr
y
in
g
to
esti
m
ate
th
e
s
a
m
e
f
r
o
m
th
e
f
r
a
m
es
th
at
h
a
v
e
to
b
e
co
n
v
er
ted
.
S
u
ch
r
ef
er
e
n
ce
h
u
e
g
r
ad
ien
t
h
elp
s
in
ac
ce
ler
ati
n
g
t
h
e
esti
m
at
io
n
p
r
o
ce
d
u
r
al
s
p
ee
d
b
y
p
r
o
v
id
in
g
th
e
d
i
f
f
icu
lt f
ir
s
t e
s
ti
m
atio
n
s
tep
.
Use o
f
e
x
is
tin
g
g
o
o
d
q
u
alit
y
3
D
s
ter
eo
s
co
p
ic
v
id
eo
s
ca
n
b
e
u
s
ed
to
p
r
o
v
id
e
th
e
r
e
f
er
en
tia
l
d
ep
th
cu
e
f
o
r
th
e
i
m
a
g
es
t
h
at
n
ee
d
to
b
e
p
r
o
ce
s
s
ed
to
d
eliv
er
3
D
f
o
r
liv
e
s
tr
ea
m
i
n
g
.
A
p
p
r
o
p
r
i
ate
3
D
v
id
eo
g
a
m
es
m
a
y
b
e
u
s
ed
as
r
ef
er
e
n
tial
i
n
p
u
t
t
h
at
co
r
r
esp
o
n
d
s
m
o
s
t
clo
s
el
y
to
th
e
i
m
ag
e
s
u
n
d
er
co
n
s
id
er
atio
n
.
I
n
t
h
is
w
o
r
k
liv
e
s
tr
ea
m
i
n
g
o
f
s
o
cc
er
g
a
m
es i
s
a
tte
m
p
ted
.
T
h
e
m
o
s
t
ac
ce
p
tab
le
q
u
ali
t
y
v
id
eo
g
a
m
e
s
p
r
o
v
id
e
th
e
m
o
s
t a
p
p
r
o
p
r
iate
d
ep
th
p
er
c
ep
tio
n
cu
e
th
r
o
u
g
h
th
e
tex
t
u
r
e
g
r
ad
ien
t
u
s
ed
to
d
ev
elo
p
s
u
ch
g
a
m
es.
T
h
e
r
est
o
f
th
e
p
r
o
ce
s
s
u
tili
ze
s
ap
p
r
o
p
r
iate
iter
ativ
e
alg
o
r
ith
m
s
b
ased
o
n
a
s
o
u
n
d
in
itial
v
al
u
e.
T
h
u
s
,
t
h
is
p
r
o
ce
s
s
u
tili
ze
s
th
e
e
x
is
tin
g
h
ig
h
-
q
u
al
it
y
i
n
v
id
eo
g
a
m
e
s
(
th
at
r
ese
m
b
le
r
ea
l
‘
f
ee
l
’
v
er
y
clo
s
el
y
)
a
n
d
ca
n
p
r
o
v
id
e
a
p
p
r
o
p
r
iate
r
ef
er
en
tial
h
u
e
g
r
a
d
ien
t
c
u
es
f
o
r
d
ep
th
esti
m
atio
n
.
T
h
e
r
est
o
f
th
e
p
r
o
ce
s
s
is
a
u
to
m
a
tic
as
i
n
o
th
er
m
et
h
o
d
s
,
to
o
.
T
h
is
p
r
o
ce
s
s
i
m
p
r
o
v
es
t
h
e
co
n
v
er
s
io
n
p
r
o
ce
s
s
m
ai
n
l
y
b
y
r
ed
u
cin
g
t
h
e
p
r
o
ce
s
s
i
n
g
ti
m
e
th
r
o
u
g
h
s
p
atial
(
h
eig
h
t
a
n
d
w
i
d
th
)
ass
i
g
n
m
en
t
o
f
d
ep
th
g
r
ad
ien
t
i
n
f
er
r
ed
f
r
o
m
th
e
‘
m
a
n
u
f
ac
t
u
r
ed
’
r
ef
er
e
n
ti
al
v
id
eo
g
a
m
e
s
.
A
s
t
h
is
p
r
o
ce
s
s
u
s
es
e
x
is
t
in
g
s
ter
eo
s
co
p
ic
3
D
f
r
a
m
es
a
s
r
ef
er
en
ce
cu
es,
t
h
e
p
r
o
p
o
s
ed
s
y
s
te
m
i
s
d
o
m
ai
n
-
s
p
ec
if
ic.
He
n
c
e,
th
e
alg
o
r
it
h
m
i
s
n
o
t
g
en
er
ic.
T
h
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
w
i
ll
h
a
v
e
to
b
e
m
o
d
u
lated
b
ased
u
p
o
n
t
h
e
co
n
t
ex
tu
a
l
d
ata.
I
t
al
s
o
i
m
p
lies
t
h
at
ea
c
h
f
r
a
m
e
m
a
y
n
ee
d
a
co
r
r
esp
o
n
d
in
g
d
ata
f
r
a
m
e
f
o
r
r
ef
er
e
n
ce
.
T
h
u
s
f
o
r
ea
ch
g
e
n
er
atio
n
,
t
h
e
p
r
o
ce
s
s
w
o
u
ld
r
ef
er
to
a
d
atab
ase
o
f
s
i
m
ilar
3
D
v
id
eo
g
a
m
es;
ass
u
r
i
n
g
th
at
th
e
r
e
n
d
er
in
g
is
o
f
g
o
o
d
q
u
alit
y
an
d
th
e
o
u
tp
u
t
i
s
v
is
u
all
y
p
le
asin
g
.
T
h
is
i
s
an
i
m
p
r
o
v
e
m
e
n
t
o
v
er
e
x
is
t
in
g
g
en
er
ic
m
et
h
o
d
o
lo
g
ies
th
at
m
a
y
s
u
f
f
er
f
r
o
m
late
n
c
y
is
s
u
e
s
t
o
d
eliv
er
ac
ce
p
tab
le
q
u
alit
y
s
ter
eo
s
co
p
ic
v
i
s
u
a
l,
w
h
e
th
e
r
s
tatic
(
i
m
a
g
e
s
)
o
r
d
y
n
a
m
ic
(
li
v
e
v
id
eo
s
tr
ea
m
s
)
.
3.
M
E
T
H
O
DO
L
O
G
Y
As i
s
e
v
id
en
t
f
r
o
m
F
i
g
u
r
e
2
,
t
h
e
t
w
o
m
a
in
co
m
p
o
n
e
n
ts
t
h
at
h
elp
cr
ea
te
t
h
e
d
ep
th
d
i
m
en
s
io
n
f
r
o
m
th
e
2
D
i
m
a
g
es
f
o
r
p
r
o
p
er
g
r
ad
ien
t
h
u
e
ar
e:
1
)
th
e
S
-
R
GB
D
(
s
y
n
th
e
tic
-
R
GB
an
d
Dep
th
)
d
at
ab
ase
f
r
o
m
s
i
m
i
lar
i
m
a
g
es
el
s
e
w
h
er
e
ca
n
b
e
u
s
ed
as
a
r
ef
er
e
n
ce
,
u
s
i
n
g
th
e
o
n
es
th
at
m
o
s
t
c
lo
s
el
y
m
a
tch
e
s
w
i
th
t
h
e
i
n
p
u
t
i
m
a
g
e
s
ec
to
r
s
;
2
)
Ob
j
ec
t
m
as
k
i
n
g
t
h
at
w
o
u
ld
h
e
lp
d
e
-
li
n
ea
te
ed
g
e
p
r
o
f
ile
s
an
d
cr
ea
te
s
h
ar
p
er
d
ep
th
p
er
ce
p
tio
n
,
w
h
ic
h
o
t
h
er
w
is
e
w
o
u
ld
cr
ea
te
b
lu
r
r
in
e
s
s
at
t
h
e
ed
g
es
d
u
e
t
o
o
v
er
lap
p
in
g
d
ep
th
v
al
u
es.
T
h
e
ac
c
u
m
u
la
tio
n
o
f
S
-
R
GB
D
i
m
a
g
er
y
co
u
ld
b
e
a
d
au
n
ti
n
g
tas
k
,
as
m
a
n
y
d
i
f
f
er
en
t
co
llectio
n
s
w
o
u
ld
b
e
r
eq
u
ir
ed
th
is
is
s
u
e
ca
n
b
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8776
IJ
-
I
C
T
Vo
l.
7
,
No
.
3
,
Dec
em
b
er
20
1
8
:
1
41
–
1
45
144
h
an
d
led
b
y
ca
te
g
o
r
izin
g
i
m
a
g
es
b
ased
o
n
s
itu
a
tio
n
,
s
ce
n
er
y
-
b
ased
v
alu
e,
e
v
en
t
-
b
ased
i
m
a
g
es/
v
id
eo
s
an
d
o
th
er
s
u
c
h
co
m
b
in
at
io
n
s
i
n
ad
d
itio
n
to
v
id
eo
g
a
m
es
t
h
at
b
es
t
in
co
r
p
o
r
ate
d
if
f
er
en
t
i
m
a
g
er
y
w
i
th
q
u
a
lit
y
d
ep
th
an
d
h
u
e
c
u
es.
Ma
s
k
i
n
g
i
n
v
o
l
v
es
s
ep
ar
ati
n
g
d
if
f
er
en
t
o
b
j
ec
t
b
o
u
n
d
ar
ies
t
h
at
i
s
ac
h
ie
v
ed
b
y
p
r
ep
r
o
ce
s
s
in
g
t
h
e
in
p
u
t
i
m
ag
e
s
b
ased
o
n
ap
p
ea
r
an
ce
,
h
u
e,
a
n
d
m
o
tio
n
.
Fo
r
clo
s
e
-
u
p
s
,
lo
ca
l
f
ea
t
u
r
es
s
u
ch
as
m
att
in
g
-
b
ased
ap
p
r
o
ac
h
ar
e
u
s
ed
t
h
at
in
v
o
l
v
es
i
n
itiatio
n
u
s
i
n
g
tr
aj
ec
to
r
y
s
eg
m
e
n
tat
io
n
[
1
6
]
,
[
1
7
]
f
o
r
n
o
n
-
c
lo
s
e
-
u
p
i
m
a
g
es
th
e
GM
M
(
Gau
s
s
ia
n
Mi
x
t
u
r
e
Mo
d
el)
p
r
o
ce
s
s
in
g
d
eliv
er
s
t
h
e
r
eq
u
ir
ed
co
n
ten
t o
f
f
in
e
q
u
a
l
it
y
.
Fig
u
r
e
2
.
T
h
e
co
n
ce
p
tu
al
p
r
o
c
ess
s
eq
u
e
n
ce
i
s
d
ep
icted
in
Fi
g
u
r
e
2
T
h
e
o
v
er
ar
ch
in
g
co
n
ce
p
t
is
t
h
at
o
f
Gr
ad
ien
t
-
b
ased
co
n
v
er
s
i
o
n
.
T
h
e
g
r
ad
ien
t
in
th
e
h
u
e
co
r
r
esp
o
n
d
s
to
d
ep
th
o
f
th
e
o
b
j
ec
ts
an
d
ca
n
b
e
u
s
ed
to
ass
i
g
n
co
r
r
esp
o
n
d
i
n
g
v
al
u
es.
I
n
o
r
d
er
to
ac
h
iev
e
ac
cu
r
ate
ass
i
g
n
m
e
n
t,
t
h
e
i
n
p
u
t
i
m
a
g
e
is
d
if
f
er
en
tiated
in
n
x
n
b
lo
ck
s
a
n
d
th
e
n
co
m
p
ar
ed
w
it
h
s
i
m
ilar
i
m
a
g
e
s
th
a
t
ass
i
g
n
th
e
d
ep
th
v
al
u
es
ac
co
r
d
in
g
to
th
e
h
u
e
i
n
th
e
r
e
f
er
en
ce
i
m
a
g
e
(
S
-
R
GB
D
d
atab
ase)
.
T
h
e
m
as
k
ed
i
m
a
g
es
w
it
h
c
lear
d
if
f
er
en
tiatio
n
o
f
o
b
j
ec
ts
ad
d
to
th
e
clar
it
y
,
b
o
th
in
h
u
e
a
n
d
d
ep
th
.
T
h
e
p
ictu
r
e
q
u
alit
y
a
n
d
co
m
p
u
tatio
n
a
l
d
if
f
ic
u
ltie
s
f
o
r
m
t
h
e
b
asis
o
f
t
h
e
tr
ad
e
-
o
f
f
p
o
i
n
ts
w
h
er
e
t
h
e
co
n
v
er
s
io
n
p
r
o
ce
s
s
i
s
r
eq
u
ir
ed
to
ac
h
iev
e
a
b
alan
ce
b
et
w
ee
n
t
h
ese
t
w
o
p
ar
a
m
eter
s
.
T
h
e
co
m
p
u
tatio
n
a
l
d
if
f
ic
u
lt
y
ar
i
s
es
f
r
o
m
t
h
e
n
u
m
b
er
o
f
b
lo
ck
s
cr
ea
ted
to
ac
h
iev
e
a
v
is
u
all
y
p
leas
in
g
3
D
i
m
a
g
e.
T
h
e
d
u
r
atio
n
r
eq
u
ir
ed
f
o
r
m
o
r
e
n
u
m
b
er
o
f
b
lo
ck
s
i
s
O(
e
5E
-
5n
)
,
w
h
er
e,
n
is
th
e
n
u
m
b
er
o
f
b
lo
ck
s
.
Fi
g
u
r
e
3
b
elo
w
ill
u
s
tr
ate
s
th
e
co
m
p
u
tatio
n
al
d
u
r
atio
n
.
Fig
u
r
e
3
.
C
o
m
p
u
tat
io
n
al
co
m
p
lex
it
y
[
1
5
]
4.
DIS
CU
SS
I
O
N
T
h
e
b
asic
b
en
ch
m
ar
k
ca
te
g
o
r
ies
f
o
r
3
D
p
er
ce
p
tio
n
in
i
m
a
g
es
ar
e
b
ased
o
n
th
r
ee
ca
teg
o
r
ies:
v
is
u
al
p
leasan
t
n
es
s
,
d
ep
th
q
u
alit
y
,
an
d
p
ictu
r
e
q
u
alit
y
.
V
is
u
a
l
p
leasan
t
n
es
s
/d
is
co
m
f
o
r
t
is
th
e
p
h
y
s
io
lo
g
ica
l
d
is
co
m
f
itu
r
e
s
u
c
h
as
v
is
u
al
s
tr
ain
,
h
ea
d
ac
h
e
o
n
co
n
ti
n
u
o
u
s
v
ie
w
i
n
g
an
d
o
t
h
er
s
y
m
p
to
m
s
.
T
h
e
d
ep
th
q
u
alit
y
i
s
m
ea
s
u
r
ed
as
a
p
er
ce
p
tio
n
o
f
en
v
i
s
ag
ed
d
ep
th
.
T
h
e
tr
an
s
m
is
s
io
n
p
ar
a
m
e
ter
s
an
d
t
h
e
en
co
d
in
g
f
r
a
m
e
w
o
r
k
af
f
ec
t
t
h
e
p
ictu
r
e
q
u
a
lit
y
.
Si
n
ce
d
ep
th
q
u
alit
y
an
d
v
is
u
al
p
l
ea
s
an
t
n
es
s
ar
e
q
u
alita
tiv
e
p
ar
a
m
eter
s
,
as
s
es
s
m
en
t
s
h
o
u
ld
in
v
o
l
v
e
s
h
o
w
i
n
g
d
if
f
er
en
t
3
D
i
m
ag
e
s
g
en
er
ated
b
y
d
if
f
er
e
n
t
tec
h
n
iq
u
es
to
th
e
s
a
m
e
s
e
t
o
f
v
ie
w
er
s
/r
e
s
p
o
n
d
en
ts
to
g
et
a
f
air
er
g
r
ad
in
g
.
S
u
c
h
i
m
ag
e
s
/
v
id
eo
s
ca
n
b
e
s
h
o
w
n
f
o
r
s
h
o
r
t d
u
r
atio
n
s
i
n
a
r
an
d
o
m
s
eq
u
en
ce
,
a
n
d
allo
w
t
h
e
v
i
e
w
er
s
to
r
ate
th
e
m
o
n
asc
en
d
in
g
o
r
d
er
o
f
q
u
alit
y
,
m
ar
k
ed
b
y
d
is
ti
n
ct
n
u
m
b
er
s
/
g
r
ad
es
(
b
ad
,
p
o
o
r
,
f
air
,
g
o
o
d
,
ex
ce
llen
t)
f
o
r
d
ep
th
q
u
alit
y
a
n
d
o
n
a
s
i
m
i
lar
g
r
ad
in
g
t
h
at
s
tar
ts
at
v
er
y
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
I
C
T
I
SS
N:
2252
-
8776
A
d
d
r
ess
in
g
La
ten
cy
I
s
s
u
es in
2
D
to
3
D
C
o
n
ve
r
s
io
n
:
Dep
lo
y
in
g
A
va
ila
b
le
S
yn
th
etic
…
(
N
.
L.
Ma
n
a
s
a
)
145
u
n
co
m
f
o
r
tab
le
an
d
asce
n
d
s
to
v
er
y
co
m
f
o
r
tab
le
f
o
r
v
is
u
al
co
m
f
o
r
t.
I
n
a
r
an
d
o
m
s
t
u
d
y
c
ar
r
ied
o
u
t
th
u
s
,
th
e
i
m
a
g
es p
r
o
d
u
ce
d
b
y
t
h
e
p
r
esen
ted
tech
n
iq
u
e
s
co
r
ed
v
er
y
h
i
g
h
in
t
h
e
q
u
ali
t
y
a
s
s
es
s
m
en
t o
n
b
o
th
co
u
n
t
s
.
5.
CO
NCLU
SI
O
N
T
h
is
p
ap
er
h
as
in
tr
o
d
u
ce
d
a
n
in
n
o
v
ati
v
e
m
et
h
o
d
f
o
r
2
D
t
o
3
D
co
n
v
er
s
io
n
o
f
i
m
a
g
es/
v
id
eo
s
.
T
h
e
s
tr
en
g
th
o
f
th
e
p
r
o
p
o
s
al
lies
in
f
a
s
ter
co
n
v
er
s
io
n
r
ate.
T
h
e
q
u
alit
y
o
f
th
e
r
en
d
er
ed
3
D
im
ag
e
i
s
also
r
ated
h
ig
h
l
y
b
y
v
ie
w
er
s
.
T
h
e
m
ai
n
p
r
o
ce
s
s
es
in
v
o
l
v
ed
ar
e
th
o
s
e
o
f
S
-
R
GB
D
d
atab
ase
co
llated
f
r
o
m
s
i
m
ilar
v
id
eo
s
o
r
f
r
o
m
v
id
eo
g
a
m
e
s
an
d
ed
g
e
in
f
o
r
m
atio
n
.
T
h
e
o
v
er
ar
ch
in
g
s
ch
e
m
a
is
t
h
e
s
q
u
ar
e
m
atr
i
x
b
lo
ck
s
eg
m
e
n
tatio
n
th
at
is
e
v
en
t
u
all
y
m
o
d
i
f
ied
f
o
l
lo
w
i
n
g
d
ep
th
h
u
e
g
r
ad
atio
n
cu
es
an
d
co
m
p
ar
is
o
n
w
it
h
r
ef
er
e
n
tial
d
atab
ase.
T
h
e
m
ai
n
co
n
tr
ib
u
tio
n
o
f
th
e
p
ap
er
is
th
at
o
f
d
o
m
ai
n
ce
n
ter
ed
in
itial
p
r
ed
ictio
n
v
alu
e
s
.
Ma
tr
ix
b
ased
ed
g
e
in
f
o
r
m
atio
n
r
e
m
o
v
e
s
th
e
b
o
u
n
d
ar
y
d
is
co
n
t
in
u
itie
s
,
an
d
u
s
e
o
f
ex
is
tin
g
,
q
u
ali
t
y
i
m
a
g
es
,
th
at
ad
d
s
to
th
e
g
r
ad
ien
t
d
i
s
ce
r
n
m
e
n
t.
T
h
u
s
t
h
e
m
et
h
o
d
r
eso
lv
e
s
t
w
o
m
aj
o
r
is
s
u
es
–
th
at
o
f
late
n
c
y
an
d
i
n
v
e
s
t
m
en
t
r
eq
u
ir
ed
f
o
r
2
D
to
3
D
co
n
v
er
s
io
n
.
A
p
o
s
s
ib
le
li
m
itat
io
n
co
u
ld
b
e
th
e
d
i
v
er
s
e
a
n
d
q
u
alit
y
d
ata
b
ase
r
eq
u
ir
ed
to
i
m
p
le
m
en
t
t
h
e
p
r
o
ce
s
s
p
r
esen
ted
in
t
h
is
w
o
r
k
.
Seco
n
d
ly
,
t
h
i
s
s
y
s
te
m
is
n
o
t g
e
n
er
ic
-
a
s
ac
r
if
ice
f
o
r
q
u
alit
y
an
d
r
e
d
u
ctio
n
i
n
laten
c
y
.
RE
F
E
R
E
NC
E
S:
[1
]
C.
R.
M
a
d
a
n
,
“
Cre
a
ti
n
g
3
D v
i
su
a
li
z
a
ti
o
n
s
o
f
M
RI
d
a
ta:
A
b
ri
e
f
g
u
id
e
,
”
F1
0
0
0
Res
e
a
rc
h
,
v
o
l
.
4
,
A
u
g
.
2
0
1
5
.
[2
]
A
.
Zele
n
e
r,
“
S
u
rv
e
y
o
f
o
b
jec
t
c
las
si
f
ica
ti
o
n
in
3
d
ra
n
g
e
sc
a
n
s,”
T
e
c
h
.
Rep
.
,
2
0
1
5
.
[3
]
W
.
J.
T
a
m
a
n
d
L
.
Zh
a
n
g
,
“
3
D
-
T
V
Co
n
ten
t
G
e
n
e
ra
ti
o
n
:
2
D
-
to
-
3
D
Co
n
v
e
rsio
n
,
”
in
2
0
0
6
IEE
E
In
ter
n
a
t
io
n
a
l
Co
n
fer
e
n
c
e
o
n
M
u
lt
im
e
d
i
a
a
n
d
E
x
p
o
,
2
0
0
6
,
p
p
.
1
8
6
9
–
1
8
7
2
.
[4
]
S
.
Ba
tt
iato
,
S
.
C
u
rti
,
M
.
L
.
Ca
sc
ia,
M
.
T
o
rto
ra
,
a
n
d
E.
S
c
o
r
d
a
to
,
“
De
p
th
ma
p
g
e
n
e
ra
ti
o
n
b
y
ima
g
e
c
la
ss
if
ica
ti
o
n
,
”
p
re
se
n
ted
a
t
t
h
e
T
h
re
e
-
Di
m
e
n
sio
n
a
l
Im
a
g
e
Ca
p
tu
re
a
n
d
A
p
p
li
c
a
ti
o
n
s V
I,
2
0
0
4
,
v
o
l.
5
3
0
2
,
p
p
.
9
5
–
1
0
5
.
[5
]
J.
P
a
rk
a
n
d
C.
Kim
,
“
Extra
c
ti
n
g
fo
c
u
se
d
o
b
jec
t
fro
m
l
o
w
d
e
p
th
-
of
-
fi
e
ld
ima
g
e
se
q
u
e
n
c
e
s
,
”
p
re
se
n
ted
a
t
t
h
e
Visu
a
l
Co
m
m
u
n
ica
ti
o
n
s a
n
d
Im
a
g
e
P
ro
c
e
ss
in
g
2
0
0
6
,
2
0
0
6
,
v
o
l.
6
0
7
7
,
p
.
6
0
7
7
1
O.
[6
]
P
.
V.
Ha
rm
a
n
,
J.
F
lac
k
,
S
.
F
o
x
,
a
n
d
M
.
Do
w
le
y
,
“
Ra
p
id
2
D
-
to
-
3
D
c
o
n
v
e
rs
io
n
,
”
p
re
se
n
ted
a
t
th
e
S
tere
o
sc
o
p
ic
Disp
lay
s a
n
d
V
irt
u
a
l
Re
a
li
ty
S
y
st
e
m
s IX
,
2
0
0
2
,
v
o
l
.
4
6
6
0
,
p
p
.
7
8
–
87.
[7
]
D.
Ho
iem
,
A
.
A
.
E
f
ro
s,
a
n
d
M
.
He
b
e
rt,
“
Au
to
m
a
ti
c
Ph
o
to
P
o
p
-
up
,
”
in
A
CM
S
IGG
R
A
P
H
2
0
0
5
P
a
p
e
rs,
Ne
w
Yo
rk
,
NY
,
USA
,
2
0
0
5
,
p
p
.
5
7
7
–
5
8
4
.
[8
]
Y.
J.
Ju
n
g
,
A
.
Ba
ik
,
J.
Kim
,
a
n
d
D.
P
a
rk
,
“
A
n
o
v
e
l
2
D
-
to
-
3
D
c
o
n
v
e
rs
io
n
tec
h
n
iq
u
e
b
a
se
d
o
n
re
la
ti
v
e
h
e
ig
h
t
-
d
e
p
t
h
c
u
e
,
”
p
re
se
n
ted
a
t
t
h
e
S
tere
o
sc
o
p
i
c
Disp
lay
s a
n
d
A
p
p
li
c
a
ti
o
n
s XX
,
2
0
0
9
,
v
o
l.
7
2
3
7
,
p
.
7
2
3
7
1
U.
[9
]
T
.
Iin
u
m
a
,
H.
M
u
ra
ta,
S
.
Ya
m
a
sh
it
a
,
a
n
d
K.
O
y
a
m
a
d
a
,
“
5
4
.
2
:
Na
tu
ra
l
S
ter
e
o
De
p
th
Cre
a
ti
o
n
M
e
th
o
d
o
l
o
g
y
fo
r
a
Rea
l
-
ti
me
2
D
-
to
-
3
D Im
a
g
e
C
o
n
v
e
rs
io
n
,
”
S
ID S
y
m
p
.
Dig
.
T
e
c
h
.
P
a
p
.
,
v
o
l
.
3
1
,
n
o
.
1
,
p
p
.
1
2
1
2
–
1
2
1
5
,
M
a
y
2
0
0
0
.
[1
0
]
J.
y
Ch
a
n
g
,
C.
c
Ch
e
n
g
,
S
.
y
Ch
ien
,
a
n
d
L
.
g
Ch
e
n
,
“
Rela
ti
v
e
De
p
th
L
a
y
e
r
Extra
c
ti
o
n
fo
r
M
o
n
o
sc
o
p
ic
Vi
d
e
o
b
y
Us
e
o
f
M
u
lt
i
d
ime
n
si
o
n
a
l
Fi
l
ter
,
”
in
2
0
0
6
IE
EE
I
n
tern
a
ti
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
M
u
l
ti
m
e
d
ia an
d
Ex
p
o
,
2
0
0
6
,
p
p
.
2
2
1
–
2
2
4
.
[1
1
]
Y.
L
.
Ch
a
n
g
,
C.
Y.
F
a
n
g
,
L
.
F
.
Din
g
,
S
.
Y.
Ch
e
n
,
a
n
d
L
.
G
.
C
h
e
n
,
“
De
p
th
M
a
p
Ge
n
e
ra
ti
o
n
fo
r
2
D
-
to
-
3
D Co
n
v
e
rs
io
n
b
y
S
h
o
rt
-
T
e
rm
M
o
ti
o
n
Assiste
d
Co
lo
r
S
e
g
me
n
ta
t
io
n
,
”
in
2
0
0
7
I
EE
E
In
tern
a
ti
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
M
u
l
ti
m
e
d
ia
a
n
d
Ex
p
o
,
2
0
0
7
,
p
p
.
1
9
5
8
–
1
9
6
1
.
[1
2
]
C.
-
H.
C
h
o
i,
B.
-
H.
Kw
o
n
,
a
n
d
M
.
-
R.
C
h
o
i
,
“
A
re
a
l
-
ti
m
e
f
ield
-
se
q
u
e
n
ti
a
l
ste
re
o
sc
o
p
ic
im
a
g
e
c
o
n
v
e
rter,”
IEE
E
T
ra
n
s
.
Co
n
su
m.
El
e
c
tro
n
.
v
o
l.
5
0
,
n
o
.
3
,
p
p
.
9
0
3
–
9
1
0
,
A
u
g
.
2
0
0
4
.
[1
3
]
C.
-
C
.
Ch
e
n
g
,
C.
-
T
.
L
i,
P
.
-
S
.
Hu
a
n
g
,
T
.
-
K.
L
in
,
Y.
M
.
T
sa
i,
a
n
d
L
.
G
.
Ch
e
n
,
“
A
b
l
o
c
k
-
b
a
se
d
2
D
-
to
-
3
D
c
o
n
v
e
rs
io
n
sy
ste
m
wit
h
b
il
a
ter
a
l
fi
lt
e
r
,
”
in
2
0
0
9
Dig
e
st
o
f
T
e
c
h
n
ica
l
P
a
p
e
rs
I
n
tern
a
ti
o
n
a
l
Co
n
f
e
re
n
c
e
o
n
C
o
n
su
m
e
r
El
e
c
tro
n
ics
,
2
0
0
9
,
p
p
.
1
–
2.
[1
4
]
Y.
M
.
T
sa
i,
Y.
L
.
Ch
a
n
g
,
a
n
d
L
.
G
.
Ch
e
n
,
“
Bl
o
c
k
-
b
a
se
d
Va
n
i
sh
in
g
L
i
n
e
a
n
d
Va
n
ish
i
n
g
Po
i
n
t
De
tec
ti
o
n
fo
r 3
D S
c
e
n
e
Rec
o
n
stru
c
ti
o
n
,
”
in
2
0
0
6
I
n
tern
a
t
io
n
a
l
S
y
m
p
o
siu
m
o
n
In
telli
g
e
n
t
S
ig
n
a
l
P
r
o
c
e
ss
in
g
a
n
d
Co
m
m
u
n
ica
ti
o
n
s,
2
0
0
6
,
p
p
.
586
–
5
8
9
.
[1
5
]
C
.
C.
C
h
e
n
g
,
C.
T
.
L
i,
a
n
d
L
.
G
.
Ch
e
n
,
“
A
n
o
v
e
l
2
D
-
to
-
3
D
c
o
n
v
e
rsio
n
sy
ste
m
u
sin
g
e
d
g
e
in
f
o
rm
a
ti
o
n
,
”
IEE
E
T
ra
n
s.
Co
n
su
m.
El
e
c
tro
n
.
v
o
l.
5
6
,
n
o
.
3
,
p
p
.
1
7
3
9
–
1
7
4
5
,
A
u
g
.
2
0
1
0
.
[1
6
]
A
.
L
e
v
in
,
D.
L
isc
h
in
sk
i,
a
n
d
Y.
W
e
iss,
“
A
Clo
se
d
-
F
o
rm
S
o
lu
ti
o
n
to
Na
t
u
ra
l
Im
a
g
e
M
a
tt
in
g
,
”
IEE
E
T
ra
n
s.
P
a
tt
e
r
n
An
a
l
.
M
a
c
h
.
In
tell
.
,
v
o
l.
3
0
,
n
o
.
2
,
p
p
.
2
2
8
–
2
4
2
,
F
e
b
.
2
0
0
8
.
[1
7
]
P
.
Oc
h
s,
J.
M
a
li
k
,
a
n
d
T
.
Bro
x
,
“
S
e
g
m
e
n
tatio
n
o
f
M
o
v
in
g
Ob
jec
ts
b
y
L
o
n
g
T
e
r
m
V
id
e
o
An
a
ly
sis,”
IEE
E
T
ra
n
s.
Pa
tt
e
rn
A
n
a
l.
M
a
c
h
.
In
tell
., v
o
l
.
3
6
,
n
o
.
6
,
p
p
.
1
1
8
7
–
1
2
0
0
,
Ju
n
.
2
0
1
4
.
[1
8
]
L
e
e
,
Jo
n
g
w
o
n
,
H
y
u
n
ju
L
e
e
,
Do
n
g
g
y
u
n
Yu
,
a
n
d
Ho
e
k
y
u
n
g
Ju
n
g
,
"
Bo
d
y
in
f
o
r
m
a
ti
o
n
a
n
a
l
y
sis
b
a
se
d
p
e
rso
n
a
l
e
x
e
r
c
ise
m
a
n
a
g
e
m
e
n
t
s
y
ste
m
.
"
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
Co
m
p
u
ter
E
n
g
i
n
e
e
rin
g
(
IJ
ECE
)
8
,
n
o
.
1
,
2
0
1
8
.
[1
9
]
Z
a
k
a
ri
a
B
o
u
lo
u
a
rd
,
Ami
n
e
El
Ha
d
d
a
d
i,
e
t
a
l,
“
Ba
t
-
Clu
ste
r:
A
Ba
t
A
lg
o
rit
h
m
-
Ba
se
d
A
u
to
m
a
ted
G
ra
p
h
Clu
ste
ri
n
g
A
p
p
ro
a
c
h
”
,
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
C
o
mp
u
ter
En
g
i
n
e
e
rin
g
(
IJ
ECE
)
8
,
n
o
.
1
,
2
0
1
8
.
[2
0
]
M
a
d
h
u
Ch
a
n
d
ra
,
“
F
ra
m
e
wo
rk
f
o
r
Co
n
tex
t
u
a
l
Ou
tl
ier
Id
e
n
ti
f
ica
ti
o
n
u
sin
g
M
u
lt
iv
a
riate
A
n
a
l
y
sis
a
p
p
ro
a
c
h
a
n
d
Un
su
p
e
rv
ise
d
L
e
a
rn
in
g
”
,
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
C
o
mp
u
ter
E
n
g
in
e
e
rin
g
(
IJ
ECE
)
8
,
n
o
.
1
,
2
0
1
8
.
Evaluation Warning : The document was created with Spire.PDF for Python.