I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Art
if
icia
l In
t
ellig
ence
(
I
J
-
AI
)
Vo
l.
14
,
No
.
6
,
Dec
em
b
er
2
0
2
5
,
p
p
.
5
3
4
4
~
5
3
5
4
I
SS
N:
2
2
5
2
-
8
9
3
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijai.v
14
.i
6
.
p
p
5
3
4
4
-
5
3
5
4
5344
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
Enha
nced obje
ct
tracking
wi
th
a
rti
ficial bee
colo
ny
,
mo
tion
mo
deling
,
a
nd de
ep learning
Ra
m
da
ne
T
a
g
lo
ut
1,
2
,
B
ila
l Sa
o
ud
3,
4
1
LI
M
La
b
o
r
a
t
o
r
y
,
F
a
c
u
l
t
y
o
f
E
x
a
c
t
S
c
i
e
n
c
e
s
,
U
n
i
v
e
r
si
t
y
o
f
B
o
u
i
r
a
,
B
o
u
i
r
a
,
A
l
g
e
r
i
a
2
D
e
p
a
r
t
me
n
t
o
f
I
n
f
o
r
mat
i
o
n
T
e
c
h
n
o
l
o
g
y
,
F
a
c
u
l
t
y
o
f
E
x
a
c
t
S
c
i
e
n
c
e
s
,
U
n
i
v
e
r
si
t
y
o
f
B
o
u
i
r
a
,
B
o
u
i
r
a
,
A
l
g
e
r
i
a
3
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
,
M
o
h
a
me
d
El
B
a
c
h
i
r
E
l
I
b
r
a
h
i
mi
U
n
i
v
e
r
si
t
y
o
f
B
o
r
d
j
B
o
u
A
r
r
e
r
i
d
j
,
B
o
r
d
j
B
o
u
A
r
r
e
r
i
d
j
,
A
l
g
e
r
i
a
4
LI
S
EA
La
b
o
r
a
t
o
r
y
,
F
a
c
u
l
t
y
o
f
A
p
p
l
i
e
d
S
c
i
e
n
c
e
s,
U
n
i
v
e
r
si
t
y
o
f
B
o
u
i
r
a
,
B
o
u
i
r
a
,
A
l
g
e
r
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
an
21
,
2
0
2
5
R
ev
is
ed
Oct
28
,
2
0
2
5
Acc
ep
ted
No
v
8
,
2
0
2
5
As
a
fu
n
d
a
m
e
n
tal
a
sp
e
c
t
o
f
c
o
m
p
u
ter
v
isi
o
n
,
v
is
u
a
l
o
b
jec
t
trac
k
i
n
g
su
p
p
o
rts
a
wid
e
a
rra
y
o
f
a
p
p
l
ica
ti
o
n
s,
n
o
ta
b
ly
in
tran
s
p
o
rt
in
f
ra
stru
c
tu
re
a
n
d
a
d
v
a
n
c
e
d
in
d
u
strial
a
u
t
o
m
a
ti
o
n
.
Alt
h
o
u
g
h
c
o
rre
latio
n
fil
ter
-
b
a
se
d
trac
k
e
rs
d
e
m
o
n
stra
te
ro
b
u
st
p
e
rfo
rm
a
n
c
e
,
th
e
y
fa
c
e
p
e
rsiste
n
t
li
m
it
a
ti
o
n
s
in
c
l
u
d
i
n
g
sc
a
le
c
h
a
n
g
e
s,
o
b
jec
t
o
c
c
lu
sio
n
,
b
o
u
n
d
a
ry
a
rti
fa
c
ts,
a
n
d
c
o
m
p
lex
b
a
c
k
g
r
o
u
n
d
in
t
e
rfe
re
n
c
e
.
To
a
d
d
re
ss
th
e
se
issu
e
s,
we
h
a
v
e
in
tro
d
u
c
e
d
a
n
a
p
p
r
o
a
c
h
th
a
t
c
o
m
b
in
e
s
a
rti
ficia
l
b
e
e
c
o
lo
n
y
(ABC)
o
p
t
imiz
a
ti
o
n
,
d
e
e
p
n
e
u
ra
l
a
r
c
h
it
e
c
tu
re
s,
a
n
d
Ka
lma
n
fil
terin
g
tec
h
n
i
q
u
e
s.
Ou
r
m
e
th
o
d
o
lo
g
y
b
e
g
i
n
s
with
re
li
a
b
il
it
y
a
ss
e
ss
m
e
n
t
o
f
th
e
trac
k
in
g
p
i
p
e
li
n
e
,
p
r
o
c
e
e
d
in
g
t
o
c
o
m
p
u
te
targ
e
t
c
o
n
fi
d
e
n
c
e
m
e
a
su
re
s
a
t
th
e
p
re
d
icte
d
p
o
siti
o
n
,
fo
ll
o
we
d
b
y
a
n
a
d
a
p
ti
v
e
u
p
d
a
te
m
e
c
h
a
n
ism
.
Th
e
p
ro
p
o
se
d
s
y
ste
m
lev
e
ra
g
e
s
ABC
o
p
ti
m
iza
ti
o
n
f
o
r
d
y
n
a
m
ic
sc
a
le
a
d
a
p
tatio
n
wh
il
e
e
m
p
lo
y
i
n
g
Ka
lma
n
fil
terin
g
to
m
o
d
e
l
i
n
ter
-
fra
m
e
targ
e
t
m
o
ti
o
n
d
y
n
a
m
ics
.
Co
m
p
re
h
e
n
siv
e
e
v
a
l
u
a
ti
o
n
a
c
ro
ss
m
u
lt
ip
le
b
e
n
c
h
m
a
rk
d
a
tas
e
ts
d
e
m
o
n
stra
tes
o
u
r
m
e
th
o
d
'
s
e
ffica
c
y
,
p
re
c
i
sio
n
,
a
n
d
re
sili
e
n
c
e
,
a
c
h
iev
i
n
g
e
n
h
a
n
c
e
d
p
e
rfo
rm
a
n
c
e
re
lativ
e
t
o
e
x
isti
n
g
st
a
te
-
of
-
th
e
-
a
rt
a
p
p
ro
a
c
h
e
s.
K
ey
w
o
r
d
s
:
Ar
tific
ial
b
ee
co
lo
n
y
C
o
n
f
id
en
ce
in
f
o
r
m
atio
n
C
o
r
r
elatio
n
f
ilter
Kalm
an
f
ilter
Ob
ject
tr
ac
k
in
g
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
:
R
am
d
an
e
T
ag
lo
u
t
L
I
M
L
ab
o
r
at
o
r
y
,
Facu
lty
o
f
E
x
ac
t Scie
n
ce
s
,
Un
iv
er
s
ity
o
f
B
o
u
ir
a
B
o
u
ir
a,
Alg
er
ia
E
m
ail:
r
.
tag
lo
u
t@
u
n
iv
-
b
o
u
ir
a
.
d
z
1.
I
NT
RO
D
UCT
I
O
N
C
o
m
p
u
ter
v
is
io
n
p
lay
s
an
im
p
o
r
tan
t
r
o
le
to
d
a
y
.
I
ts
im
p
o
r
tan
ce
lay
s
in
th
e
p
r
o
b
lem
s
wh
ich
ca
n
s
o
lv
e.
I
t
ca
n
b
e
s
ee
n
as
a
to
o
l
th
at
all
o
ws
th
e
d
ig
ital
wo
r
ld
o
f
a
n
im
ag
e
to
in
ter
ac
t
with
its
p
h
y
s
ical
wo
r
ld
.
C
o
m
p
u
ter
v
is
io
n
h
as
b
ee
n
u
s
ed
t
o
s
o
lv
e
m
an
y
p
r
o
b
lem
s
lik
e
s
elf
-
d
r
iv
i
n
g
ca
r
s
[
1
]
,
f
ac
ial
r
ec
o
g
n
itio
n
[
2
]
,
au
g
m
en
ted
an
d
m
ix
ed
r
ea
lity
[
3
]
,
h
ea
lth
ca
r
e
an
d
h
ea
lth
tech
n
o
lo
g
y
[
4
]
,
an
d
in
ter
n
et
o
f
th
in
g
s
(
I
o
T
)
[
5
]
.
Fu
r
th
er
m
o
r
e,
co
m
p
u
ter
v
is
io
n
is
a
v
ast
d
o
m
ain
o
f
r
esear
ch
with
m
a
n
y
ax
e
s
.
Am
o
n
g
th
ese
ax
es
o
f
r
esear
ch
,
we
f
in
d
v
is
u
al
o
b
ject
tr
ac
k
in
g
.
T
h
e
m
ajo
r
g
o
al
o
f
th
is
ax
e
o
f
r
esear
ch
is
to
esti
m
ate
th
e
lo
ca
tio
n
o
f
tar
g
et
in
all
f
r
am
es
b
ased
o
n
th
e
in
itial
f
r
am
e
tar
g
et.
I
n
o
th
er
w
o
r
d
s
,
af
ter
i
d
en
tify
in
g
a
tar
g
et
in
a
v
id
e
o
f
r
am
e
,
it
is
ad
v
an
ta
g
eo
u
s
to
f
o
llo
w
its
p
o
s
itio
n
in
th
e
n
ex
t
f
r
a
m
es.
E
v
er
y
f
r
am
e
wh
er
e
th
e
tar
g
et
is
co
r
r
ec
tly
f
o
llo
we
d
p
r
o
d
u
c
es
ad
d
itio
n
al
d
etails
ab
o
u
t
th
e
tar
g
et
id
en
tity
an
d
m
o
tio
n
.
No
wad
a
y
s
,
v
is
u
al
o
b
ject
tr
ac
k
in
g
is
a
h
o
t
s
u
b
ject
o
f
r
esear
ch
,
wh
er
e
m
an
y
wo
r
k
s
h
a
v
e
b
ee
n
p
u
b
li
s
h
ed
[
6
]
.
T
h
ese
w
o
r
k
s
aim
to
im
p
r
o
v
e
t
h
e
ex
is
tin
g
m
eth
o
d
s
an
d
o
v
e
r
co
m
e
ch
allen
g
in
g
s
ce
n
ar
io
s
s
u
c
h
d
if
f
icu
lt
s
itu
atio
n
s
in
clu
d
in
g
a
lter
atio
n
s
in
t
h
e
tar
g
et
a
p
p
ea
r
an
ce
a
n
d
tr
ac
k
in
g
tar
g
ets
th
r
o
u
g
h
i
n
tr
icate
m
o
v
e
m
en
ts
.
I
n
ad
d
itio
n
,
m
eth
o
d
s
f
o
r
v
is
u
al
o
b
ject
tr
ac
k
in
g
ar
e
t
y
p
ically
ca
teg
o
r
ized
in
to
m
an
y
class
es.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
E
n
h
a
n
ce
d
o
b
ject
tr
a
ck
in
g
w
ith
a
r
tifi
cia
l b
ee
co
lo
n
y,
mo
ti
o
n
mo
d
elin
g
…
(
R
a
md
a
n
e
Ta
g
lo
u
t)
5345
T
h
ey
ar
e
ca
p
ab
le
o
f
b
ein
g
ca
teg
o
r
ized
as
m
u
ltip
le
-
o
b
je
ct
v
s
.
s
in
g
le
-
o
b
ject
tr
ac
k
in
g
m
eth
o
d
s
,
d
is
cr
im
in
ativ
e
v
s
.
g
en
er
ativ
e
m
eth
o
d
s
,
o
f
f
lin
e
m
eth
o
d
s
v
s
.
o
n
lin
e
lear
n
i
n
g
m
et
h
o
d
s
,
m
et
h
o
d
s
b
ased
o
n
n
o
n
-
awa
r
e
o
r
co
n
tex
t
-
awa
r
e.
Sin
g
l
e
o
b
ject
tr
ac
k
in
g
m
eth
o
d
s
aim
to
tr
ac
k
o
n
e
o
b
ject
at
ea
ch
s
eq
u
en
ce
in
th
e
o
t
h
er
h
an
d
,
m
u
lti
-
o
b
ject
tr
ac
k
in
g
m
eth
o
d
s
ca
n
tr
ac
k
m
u
ltip
le
tar
g
ets
s
im
u
ltan
eo
u
s
ly
[
7
]
,
[
8
]
.
Fo
r
g
en
er
ativ
e
m
eth
o
d
s
,
th
e
tr
ac
k
in
g
o
p
e
r
at
io
n
is
co
m
p
leted
b
y
f
in
d
in
g
th
e
m
o
s
t
s
u
itab
le
win
d
o
w
[
9
]
,
[
1
0
]
.
Ho
wev
e
r
,
d
is
cr
im
in
ativ
e
m
eth
o
d
s
s
ep
ar
a
te
th
e
tar
g
et
o
b
ject
p
atch
f
r
o
m
its
b
ac
k
g
r
o
u
n
d
[
1
1
]
.
Dis
cr
im
in
ativ
e
co
r
r
elatio
n
f
ilt
er
(
C
F)
-
b
ased
tr
ac
k
er
s
co
n
tin
u
e
to
p
r
esen
t
s
ig
n
if
ica
n
t
ch
all
en
g
es
d
u
e
to
a
v
ar
iety
o
f
f
ac
to
r
s
.
B
u
ild
in
g
a
r
o
b
u
s
t
tr
ac
k
er
r
e
q
u
ir
es
t
h
e
d
ev
elo
p
m
en
t
o
f
h
ig
h
l
y
d
is
cr
im
in
ativ
e
f
ea
tu
r
es
an
d
a
s
tr
o
n
g
class
if
ier
.
Mo
s
t
cu
r
r
en
t
DC
F
-
b
ased
m
eth
o
d
s
s
tr
u
g
g
le
t
o
m
ain
tain
lo
n
g
-
ter
m
co
n
s
is
ten
cy
,
o
f
ten
lo
s
in
g
cr
itical
tar
g
et
in
f
o
r
m
ati
o
n
f
r
o
m
th
e
in
itial
f
r
am
e,
wh
i
ch
lead
s
to
d
r
if
t
an
d
r
e
d
u
ce
d
tr
ac
k
in
g
r
eliab
ilit
y
o
v
er
tim
e.
Ad
d
itio
n
ally
,
m
a
n
y
tr
ac
k
er
s
lac
k
r
o
b
u
s
t
m
ec
h
a
n
is
m
s
f
o
r
h
a
n
d
lin
g
ch
allen
g
in
g
s
ce
n
ar
io
s
s
u
ch
as
o
cc
lu
s
io
n
,
d
e
f
o
r
m
atio
n
,
an
d
r
a
p
id
m
o
tio
n
,
r
esu
ltin
g
i
n
d
ec
r
ea
s
ed
p
er
f
o
r
m
an
ce
u
n
d
er
r
ea
l
-
w
o
r
ld
co
n
d
itio
n
s
.
An
o
th
er
s
ig
n
if
ican
t
s
h
o
r
tco
m
in
g
is
th
e
ab
s
en
ce
o
f
ef
f
ec
ti
v
e
r
eliab
ilit
y
m
o
n
ito
r
i
n
g
d
u
r
i
n
g
m
o
d
el
tr
ain
in
g
,
wh
ic
h
ca
n
allo
w
n
o
i
s
y
o
r
in
v
alid
d
ata
t
o
d
e
g
r
ad
e
m
o
d
el
ac
c
u
r
ac
y
.
Fu
r
th
er
m
o
r
e,
th
e
s
ca
le
f
ilter
’
s
f
ix
ed
,
p
r
e
d
ef
in
e
d
s
ca
le
r
an
g
e
o
r
r
ed
u
n
d
an
t
s
ca
le
ass
u
m
p
tio
n
s
lim
it
its
ab
ilit
y
to
ad
ap
t
to
d
y
n
am
ic
ch
a
n
g
es
in
tar
g
et
s
ize.
T
h
ese
lim
itatio
n
s
h
ig
h
lig
h
t
th
e
n
ee
d
f
o
r
in
n
o
v
ativ
e
s
o
lu
tio
n
s
th
at
ca
n
en
s
u
r
e
in
f
o
r
m
atio
n
r
eten
tio
n
,
r
o
b
u
s
t tr
ajec
to
r
y
m
o
d
elin
g
,
r
eliab
le
tr
ai
n
in
g
p
r
o
ce
s
s
es,
an
d
ad
ap
tiv
e
s
ca
le
m
an
a
g
em
en
t.
T
o
o
v
e
r
co
m
e
th
ese
is
s
u
es,
we
h
av
e
p
r
o
p
o
s
ed
a
n
ew
m
eth
o
d
o
f
v
is
u
al
o
b
ject
tr
ac
k
i
n
g
m
eth
o
d
u
tili
zin
g
a
CF
f
r
am
ewo
r
k
,
e
n
h
an
ce
d
b
y
in
te
g
r
atin
g
a
d
e
ep
lear
n
i
n
g
n
etwo
r
k
.
Ou
r
m
eth
o
d
is
s
h
o
wn
t
o
o
u
tp
er
f
o
r
m
s
ev
er
al
lead
i
n
g
s
tate
-
of
-
th
e
-
a
r
t a
p
p
r
o
ac
h
es in
ex
p
er
im
en
tal
ev
alu
atio
n
s
.
T
h
e
m
ain
co
n
tr
ib
u
tio
n
s
o
f
o
u
r
wo
r
k
ar
e
as f
o
llo
ws:
–
Usi
n
g
a
r
eliab
le
o
n
lin
e
tr
ai
n
in
g
m
eth
o
d
th
at
ca
n
k
ee
p
im
p
o
r
t
an
t in
f
o
r
m
atio
n
f
r
o
m
th
e
f
ir
s
t f
r
am
e.
–
Utilizin
g
th
e
Kalm
an
f
ilter
in
g
ap
p
r
o
ac
h
to
g
ath
er
tar
g
et
tr
ajec
to
r
y
in
f
o
r
m
atio
n
s
o
th
at
th
e
tr
ac
k
er
ca
n
co
n
d
u
ct
r
o
b
u
s
t tr
ac
k
in
g
e
v
en
wh
en
th
e
o
b
ject
is
o
cc
lu
d
ed
,
d
ef
o
r
m
ab
le,
o
r
m
o
v
in
g
q
u
ic
k
ly
.
–
Usi
n
g
a
m
ec
h
an
is
m
f
o
r
m
o
n
it
o
r
in
g
th
e
p
r
o
ce
s
s
r
eliab
ilit
y
d
u
r
in
g
th
e
m
o
d
el
tr
ain
in
g
g
u
ar
an
tees
th
at
th
e
tar
g
et
in
f
o
r
m
atio
n
is
v
alid
.
–
W
e
p
r
o
p
o
s
e
a
m
eth
o
d
f
o
r
ass
ess
in
g
s
ca
le
v
ar
iatio
n
in
tar
g
e
t
tr
ac
k
in
g
alg
o
r
ith
m
s
th
at
tak
e
s
ad
v
an
tag
e
o
f
an
o
p
tim
izatio
n
s
tr
ateg
y
,
ai
m
in
g
to
a
d
d
r
ess
r
ed
u
n
d
a
n
cy
an
d
f
i
x
ed
s
ca
le
is
s
u
es
co
m
m
o
n
ly
f
o
u
n
d
i
n
cu
r
r
en
t te
ch
n
iq
u
es.
T
h
e
p
ap
e
r
will
b
e
s
tr
u
ctu
r
ed
as
f
o
llo
ws
:
th
e
s
ec
o
n
d
s
ec
tio
n
will
co
v
er
o
u
r
p
r
o
p
o
s
ed
m
eth
o
d
.
W
e
s
h
all
o
u
tlin
e
th
e
m
eth
o
d
in
th
e
th
ir
d
p
ar
t.
R
esu
lts
an
d
d
is
cu
s
s
io
n
will
b
e
illu
s
tr
ated
in
th
e
f
o
u
r
th
s
ec
tio
n
.
T
h
e
f
if
th
s
ec
tio
n
will d
escr
ib
e
o
u
r
co
n
clu
s
io
n
an
d
f
u
tu
r
e
d
i
r
ec
tio
n
s
.
2.
P
RO
P
O
SE
D
M
E
T
H
O
D
T
h
is
p
ar
t o
f
th
e
p
ap
er
p
r
esen
ts
b
o
th
th
e
d
esig
n
p
r
in
ci
p
les an
d
th
e
o
p
er
atio
n
al
p
r
o
ce
s
s
o
f
th
e
p
r
o
p
o
s
ed
o
b
ject
tr
ac
k
in
g
a
p
p
r
o
ac
h
.
Fi
r
s
t
o
f
all,
t
h
e
ef
f
icien
t
o
n
lin
e
tr
ain
in
g
s
tr
ateg
y
,
wh
ich
k
ee
p
s
th
e
v
alu
ab
le
in
f
o
r
m
atio
n
o
f
th
e
f
i
r
s
t
f
r
am
e
tar
g
et
in
tact
th
r
o
u
g
h
o
u
t
th
e
t
r
a
in
in
g
,
will
b
e
illu
s
tr
ated
.
T
h
en
,
th
e
s
im
u
latio
n
o
f
o
b
ject
m
o
tio
n
b
ased
o
n
Kal
m
an
f
ilter
(
KF)
i
n
o
r
d
er
t
o
g
et
in
f
o
r
m
atio
n
ab
o
u
t
its
tr
aje
cto
r
y
is
g
o
i
n
g
to
b
e
p
r
esen
ted
.
Fin
ally
,
th
e
ar
tific
ial
b
ee
co
lo
n
y
o
p
tim
izatio
n
(
AB
C
)
wi
ll
b
e
u
s
ed
to
en
h
an
ce
th
e
p
r
ec
i
s
io
n
o
f
o
u
r
tr
ac
k
er
.
T
h
e
p
r
o
p
o
s
ed
tr
ac
k
in
g
m
eth
o
d
in
v
o
lv
es
in
itializin
g
th
e
CF
with
tar
g
et
in
f
o
r
m
at
io
n
f
r
o
m
th
e
in
itial
f
r
am
e,
u
s
in
g
b
o
th
th
e
KF
an
d
th
e
C
F
to
est
im
ate
th
e
tar
g
et
p
o
s
itio
n
in
co
n
s
ec
u
tiv
e
f
r
a
m
es,
an
d
ass
ess
in
g
co
n
s
is
ten
cy
th
r
o
u
g
h
a
r
eliab
il
ity
an
aly
s
is
m
o
d
u
le.
Af
te
r
ac
h
iev
in
g
ac
c
u
r
ate
lo
ca
lizatio
n
,
th
e
AB
C
m
eth
o
d
is
ap
p
lied
to
esti
m
ate
th
e
tar
g
et
s
ca
le.
Nex
t,
th
e
f
ir
s
t
f
r
am
e
tar
g
et
in
f
o
r
m
atio
n
a
n
d
t
h
e
cu
r
r
en
t
ac
c
u
r
ate
p
o
s
itio
n
ar
e
u
s
ed
to
tr
ain
th
e
m
o
d
el
s
im
u
ltan
eo
u
s
ly
.
T
h
e
C
F
is
u
p
d
ated
to
in
clu
d
e
"r
,
"
wh
ich
r
e
p
r
esen
ts
th
e
r
eliab
ilit
y
o
f
th
e
tr
ac
k
in
g
r
esu
lt,
an
d
th
e
th
r
esh
o
ld
"T
h
r
"
is
u
s
ed
to
m
ea
s
u
r
e
tr
ac
k
in
g
r
o
b
u
s
tn
ess
.
C
N
N
is
em
p
lo
y
ed
f
o
r
ex
tr
ac
tin
g
v
is
u
al
in
f
o
r
m
atio
n
f
r
o
m
th
e
im
ag
es.
T
h
e
AB
C
m
eth
o
d
is
th
e
n
u
s
ed
f
o
r
tar
g
et
s
ca
le
esti
m
atio
n
.
T
h
e
f
r
am
ewo
r
k
o
f
o
u
r
n
ew
o
b
ject
tr
ac
k
e
r
is
r
ep
r
e
s
en
ted
in
Fig
u
r
e
1
.
2
.
1
.
T
ra
j
ec
t
o
ry
esti
m
a
t
io
n b
a
s
ed
o
n
K
a
lm
a
n f
ilte
r
KF
ca
n
p
r
o
v
id
e
an
esti
m
atio
n
o
f
s
o
m
e
u
n
k
n
o
wn
v
ar
iab
le
s
b
ased
o
n
s
o
m
e
m
ea
s
u
r
em
e
n
ts
tak
en
o
v
er
tim
e.
KF
h
as
b
ee
n
v
er
y
u
s
ef
u
l
in
d
if
f
er
en
t
a
p
p
licati
o
n
s
.
Am
o
n
g
th
ese
ap
p
licatio
n
s
,
we
f
in
d
v
i
d
eo
tr
ac
k
in
g
.
KF
u
s
es
s
im
p
le
f
o
r
m
s
an
d
d
o
es
n
o
t
n
ee
d
a
l
o
t
o
f
co
m
p
u
tatio
n
al
p
o
wer
[
1
2
]
.
KF
ca
n
b
e
s
ee
n
as
an
id
ea
l
f
ilter
,
b
ec
a
u
s
e
it
m
in
i
m
izes
th
e
d
if
f
e
r
en
ce
b
etwe
en
th
e
r
ea
l
an
d
th
e
esti
m
ated
s
tate.
B
ec
au
s
e
o
f
s
im
p
licity
an
d
ef
f
icien
cy
o
f
K
F
to
r
e
p
r
esen
t
tar
g
et
m
o
tio
n
,
a
co
n
s
tan
t
v
elo
city
o
f
m
o
d
el
h
as
b
ee
n
u
s
ed
.
T
h
is
p
r
o
ce
d
u
r
e
ca
n
b
e
d
i
v
id
ed
in
t
o
two
s
tag
es,
wh
ich
ar
e
p
r
e
d
ictio
n
an
d
co
r
r
ec
tio
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
er
20
25
:
5
3
4
4
-
5
3
5
4
5346
Fig
u
r
e
1
.
T
h
e
f
r
a
m
ewo
r
k
o
f
th
e
p
r
o
p
o
s
ed
o
b
ject
tr
ac
k
er
2
.
2
.
Co
nv
o
lutio
n t
a
rg
e
t
lo
ca
t
io
n
Stan
d
ar
d
tr
ac
k
er
s
b
ased
o
n
co
r
r
elatio
n
f
ilter
in
g
[
1
3
]
h
av
e
a
f
u
n
d
am
e
n
tal
id
ea
,
wh
ich
is
to
lear
n
a
d
is
cr
im
in
an
t
class
if
ier
at
f
ir
s
t
,
th
en
s
ea
r
ch
f
o
r
th
e
lar
g
est
v
alu
e
o
f
th
e
r
elev
an
t
r
esp
o
n
s
e
g
r
ap
h
to
d
eter
m
i
n
e
th
e
esti
m
ated
lo
ca
tio
n
o
f
th
e
t
ar
g
et
o
b
ject.
Fu
r
th
er
m
o
r
e,
th
e
s
e
tr
ac
k
er
s
b
eg
in
b
y
tr
ai
n
in
g
t
h
e
f
ilter
with
lab
els
an
d
s
am
p
les.
T
h
e
tar
g
et
is
th
en
lo
ca
ted
i
n
th
e
s
ea
r
c
h
p
atch
u
s
in
g
th
e
f
ilter
.
Fin
ally
,
t
h
e
n
ew
o
b
ject
p
o
s
itio
n
will b
e
u
s
ed
to
u
p
d
ate
th
e
f
ilte
r
.
T
h
e
m
o
d
el
o
f
th
is
p
r
o
ce
s
s
is
as
(
1
)
:
∗
=
∑
‖
.
,
−
(
,
)
‖
2
+
‖
‖
2
2
(
1
)
W
h
er
e
s
tan
d
s
f
o
r
s
am
p
les,
f
o
r
th
e
lear
n
ed
CF
,
f
o
r
th
e
r
e
g
r
ess
io
n
r
esp
o
n
s
e
with
a
Ga
u
s
s
ian
s
h
ap
e,
is
a
p
ar
am
eter
o
f
r
eg
u
lar
izatio
n
to
o
v
er
co
m
e
th
e
o
v
er
-
f
itti
n
g
p
r
o
b
lem
.
T
h
e
s
e
n
s
it
i
v
i
t
y
o
f
d
e
e
p
a
r
c
h
i
te
c
t
u
r
e
s
t
o
a
p
p
e
a
r
a
n
c
e
c
h
a
n
g
e
s
n
e
c
e
s
s
it
a
t
es
a
r
o
b
u
s
t
f
e
a
t
u
r
e
e
x
t
r
a
c
t
i
o
n
s
t
r
a
t
e
g
y
.
C
o
n
s
e
q
u
e
n
tl
y
,
w
e
e
m
p
l
o
y
t
h
e
p
r
e
-
t
r
a
i
n
e
d
V
GG
N
e
t
[
1
4
]
t
o
g
e
n
e
r
a
t
e
a
m
u
l
t
i
-
l
a
y
e
r
f
e
at
u
r
e
r
e
p
r
e
s
e
n
t
a
ti
o
n
f
r
o
m
l
a
y
e
r
s
c
o
n
v
3
-
4
,
c
o
n
v
4
-
4
,
a
n
d
c
o
n
v
5
-
4
.
T
h
e
r
e
s
u
l
t
i
n
g
f
e
at
u
r
e
v
e
c
t
o
r
,
w
i
t
h
d
i
m
e
n
s
i
o
n
s
(
w
i
d
t
h
)
,
(
h
e
i
g
h
t
)
,
a
n
d
(
c
h
a
n
n
e
l
s
)
.
I
t
i
s
i
n
t
e
g
r
a
t
e
d
i
n
t
o
o
u
r
t
r
a
c
k
i
n
g
f
r
a
m
e
w
o
r
k
.
T
h
e
f
i
n
a
l
o
u
t
p
u
t
i
s
p
r
o
d
u
c
e
d
b
y
c
o
m
p
u
t
i
n
g
a
n
i
n
n
e
r
p
r
o
d
u
c
t
t
h
r
o
u
g
h
t
h
e
a
p
p
l
i
c
a
ti
o
n
o
f
a
l
i
n
e
a
r
k
e
r
n
e
l
i
n
Hi
l
b
e
r
t
s
p
a
c
e
,
s
u
c
h
a
s
(
2
)
:
.
,
=
∑
,
,
=
1
.
,
,
(
2
)
T
h
e
m
o
d
el
em
p
l
o
y
s
an
×
CF
.
T
r
ain
in
g
d
ata
is
g
en
er
ated
b
y
cr
ea
tin
g
cir
c
u
lar
s
h
if
ts
o
f
a
b
ase
s
am
p
le
,
p
r
o
d
u
cin
g
a
s
et
o
f
s
am
p
les
,
in
d
ex
ed
o
v
er
(
,
)
∈
[
0
,
1
,
.
.
.
,
−
1
]
×
[
0
,
1
,
.
.
.
,
−
1
]
.
E
ac
h
cir
cu
lar
s
am
p
le
is
ass
ig
n
ed
a
two
-
d
im
e
n
s
io
n
al
Gau
s
s
ian
f
u
n
ctio
n
as
(
3
)
.
(
,
)
=
−
(
−
/
2
)
2
+
(
−
/
2
)
2
2
2
(
3
)
T
h
e
v
a
r
iab
les
an
d
ar
e
u
s
ed
t
o
d
escr
ib
e
th
e
d
im
en
s
io
n
s
o
f
th
e
co
n
v
o
lu
tio
n
al
f
ea
t
u
r
e
m
ap
,
wh
ile
d
en
o
tes
th
e
wid
th
o
f
th
e
k
er
n
el.
Ad
d
itio
n
ally
,
b
o
th
an
d
t
h
e
co
r
r
elatio
n
f
ilter
s
h
av
e
id
en
tical
s
izes.
T
h
e
s
o
lu
tio
n
o
f
th
e
f
ilter
ca
n
b
e
o
b
tain
e
d
in
th
e
f
r
e
q
u
en
cy
d
o
m
ain
,
f
o
llo
win
g
th
e
s
am
e
p
r
o
ce
d
u
r
e
as
d
ep
icted
in
[
1
5
]
,
b
y
ap
p
l
y
in
g
t
h
e
f
ast
Fo
u
r
ier
tr
an
s
f
o
r
m
(
FF
T
)
to
ea
ch
ch
a
n
n
el.
T
h
e
c
o
r
r
es
p
o
n
d
in
g
f
ilter
in
th
e
f
r
eq
u
e
n
cy
d
o
m
ain
ca
n
b
e
r
e
p
r
esen
ted
as
(
4
)
:
=
⊙
∑
⊙
=
1
+
(4
)
C
o
m
p
lex
co
n
j
u
g
ates
ar
e
s
h
o
wn
b
y
t
h
e
b
ar
s
a
b
o
v
e
th
e
v
ar
iab
les.
T
h
e
o
p
er
ato
r
⊙
r
ep
r
esen
ts
a
Had
am
ar
d
(
elem
en
t
-
wis
e)
p
r
o
d
u
ct
an
d
∈
1
,
.
.
,
is
th
e
n
u
m
b
e
r
o
f
ch
an
n
els.
T
h
e
s
u
b
s
eq
u
en
t
f
r
am
e
will
b
e
u
tili
ze
d
f
o
r
o
b
tain
in
g
th
e
R
OI
co
n
v
o
lu
tio
n
al
f
ea
tu
r
es.
W
h
er
e
th
e
f
i
r
s
t
lay
er
is
s
p
ec
if
ie
d
a
s
∈
,
,
.
T
h
e
f
o
llo
win
g
f
o
r
m
u
la
(
5
)
s
h
o
w
s
th
e
ca
lcu
latio
n
o
f
f
ir
s
t la
y
er
r
esp
o
n
s
e
m
ap
:
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
E
n
h
a
n
ce
d
o
b
ject
tr
a
ck
in
g
w
ith
a
r
tifi
cia
l b
ee
co
lo
n
y,
mo
ti
o
n
mo
d
elin
g
…
(
R
a
md
a
n
e
Ta
g
lo
u
t)
5347
=
−
1
(
∑
⊙
=
1
)
(
5
)
B
y
id
en
tify
in
g
th
e
m
a
x
im
u
m
v
alu
e
with
in
th
e
r
esp
o
n
s
e
m
ap
,
we
co
u
ld
f
in
d
th
e
p
r
ed
icted
p
o
s
itio
n
o
f
th
e
tar
g
et,
wh
ich
is
p
r
o
v
i
d
ed
b
y
th
e
f
i
r
s
t
lay
er
.
W
e
tak
e
a
cr
itical
lo
o
k
at
th
e
r
esp
o
n
s
e
m
ap
s
p
r
o
d
u
ce
d
b
y
a
v
ar
iety
o
f
d
if
f
er
en
t
f
ea
tu
r
e
la
y
er
s
.
T
o
b
e
m
o
r
e
s
p
ec
if
ic,
w
e
em
p
lo
y
weig
h
ts
to
co
m
b
in
e
th
e
r
esp
o
n
s
e
m
ap
s
m
ad
e
u
p
o
f
co
n
v
o
lu
tio
n
al
lay
er
s
.
T
o
estab
lis
h
th
e
d
ef
in
it
e
tar
g
et
lo
ca
tio
n
,
we
co
m
m
e
n
ce
b
y
u
tili
zin
g
th
e
co
o
r
d
in
ates o
f
th
e
h
i
g
h
est v
alu
e
in
s
id
e
th
e
f
ac
to
r
e
d
co
m
p
o
s
ite
r
esp
o
n
s
e
m
ap
as (
6
)
to
(
8
)
:
=
∑
3
=
1
(
6
)
(
,
)
=
,
(
,
)
(
7
)
=
(
,
)
(
8
)
T
h
e
f
u
s
io
n
weig
h
t
o
f
ea
ch
lay
er
is
r
ep
r
esen
ted
b
y
.
W
e
ass
u
m
e
th
at
th
e
weig
h
ts
o
f
th
e
c
o
n
v
o
lu
ti
o
n
la
y
er
s
ar
e
1
,
0
.
5
,
a
n
d
0
.
2
f
r
o
m
c
o
ar
s
e
to
f
in
e,
r
esp
ec
tiv
ely
.
is
th
e
tar
g
et
lo
ca
tio
n
ca
lc
u
lated
b
y
o
u
r
tr
ac
k
er
.
2
.
3
.
Appea
ra
nce
m
o
del upd
a
t
e
As
d
escr
ib
ed
in
[
1
6
]
,
wh
en
tr
ac
k
in
g
in
a
c
o
m
p
lex
e
n
v
ir
o
n
m
e
n
t,
th
e
o
b
ject
b
ein
g
tr
ac
k
ed
’s
in
f
o
r
m
atio
n
r
e
m
ain
s
r
elativ
el
y
co
n
s
is
ten
t
b
etwe
en
c
o
n
s
ec
u
tiv
e
f
r
am
es
an
d
in
clu
d
es
a
s
i
g
n
if
ican
t
a
m
o
u
n
t
o
f
r
ed
u
n
d
an
t
i
n
f
o
r
m
atio
n
.
T
h
e
ch
an
g
e
a
n
d
r
ed
u
n
d
an
t
in
f
o
r
m
ati
o
n
lead
to
s
lo
w
d
o
wn
th
e
tr
ac
k
in
g
s
p
ee
d
a
n
d
also
wea
k
en
th
e
tr
ac
k
er
’s
p
er
f
o
r
m
an
ce
.
T
o
o
v
e
r
co
m
e
th
is
p
r
o
b
lem
,
we
h
av
e
u
s
ed
a
s
tr
ateg
y
b
ased
o
n
s
p
ar
s
e
u
p
d
atin
g
in
co
n
j
u
n
ctio
n
with
a
r
eliab
ilit
y
an
al
y
s
is
o
f
tar
g
et
in
f
o
r
m
atio
n
.
I
n
o
r
d
er
to
esti
m
ate
th
e
r
eliab
ilit
y
,
we
h
av
e
ex
p
lo
ited
th
e
(
9
)
a
n
d
(
1
0
)
:
=
1
−
(
,
0
.
6
)
(
9
)
=
(
.
,
1
)
(
1
0
)
W
h
e
r
e
a
n
d
r
e
p
r
e
s
e
n
t
t
h
e
r
esp
o
n
s
e
m
a
p
m
a
x
i
m
u
m
a
n
d
a
v
er
a
g
e
v
a
l
u
e
s
,
r
e
s
p
e
ct
i
v
e
l
y
.
T
h
i
s
p
r
o
c
e
s
s
a
l
l
o
ws
u
s
t
o
o
b
t
a
i
n
m
o
r
e
r
o
b
u
s
t
a
n
d
a
c
c
u
r
a
te
f
i
lt
e
r
s
.
CF
h
as
b
e
e
n
u
p
d
a
t
e
d
b
y
a
m
o
v
i
n
g
a
v
e
r
a
g
e
a
p
p
r
o
a
c
h
.
T
h
i
s
u
p
d
a
t
e
i
m
p
r
o
v
e
s
p
e
r
f
o
r
m
a
n
c
e
o
f
t
r
a
c
k
e
r
b
y
a
v
o
i
d
i
n
g
s
i
g
n
i
f
i
c
an
t
c
h
a
n
g
e
s
i
n
t
h
e
m
o
d
e
l
.
T
h
e
u
p
d
a
t
i
n
g
i
s
as
(
1
1
)
:
̂
=
(
1
−
)
̂
−
1
+
̂
(
1
1
)
T
h
e
r
elev
an
t le
a
r
n
in
g
r
ate
is
id
en
tifie
d
b
y
.
2
.
4
.
Sca
le
esti
m
a
t
io
n
Dete
r
m
in
in
g
t
h
e
in
te
n
d
ed
s
ize
o
f
th
e
o
b
ject
in
th
e
im
a
g
e
is
cr
itical.
Am
o
n
g
th
e
ap
p
licatio
n
s
,
wh
er
e
th
e
esti
m
atio
n
s
ca
le
is
im
p
o
r
tan
t,
we
f
in
d
r
o
b
o
tics
an
d
s
u
r
v
eillan
ce
.
I
n
o
r
d
e
r
to
esti
m
ate
th
e
tar
g
et
p
o
s
itio
n
,
th
e
CF
f
o
r
lo
ca
tin
g
th
e
tar
g
et
is
tr
ain
ed
b
y
ex
tr
ac
tin
g
d
ep
t
h
in
f
o
r
m
atio
n
f
r
o
m
th
e
im
ag
e
o
r
f
r
am
e.
Ho
we
v
er
,
th
e
s
ca
le
o
f
th
e
tar
g
et
is
m
ain
t
ain
ed
as
th
e
s
am
e
as
it
is
in
t
h
e
p
r
ev
io
u
s
f
r
am
e.
T
h
e
tr
ain
in
g
o
f
a
s
ca
le
f
ilter
is
th
er
ef
o
r
e
r
eq
u
ir
e
d
in
o
r
d
er
to
o
b
tain
an
ac
c
u
r
ate
esti
m
atio
n
o
f
th
e
p
r
esen
t ta
r
g
et
s
ca
lin
g
.
Fo
r
ea
ch
f
r
a
m
e,
s
ev
er
al
s
ca
le
s
ar
e
u
s
ed
to
ev
al
u
ate
wh
ich
i
m
ag
es
ar
e
m
o
s
t
lik
ely
to
b
e
t
ar
g
et,
with
th
e
m
id
d
le
o
f
ea
ch
p
ictu
r
e
b
ei
n
g
u
s
ed
as
a
r
ef
er
en
ce
p
o
in
t
f
o
r
all
s
ca
les.
Af
ter
o
b
tain
in
g
th
e
HOG
f
ea
tu
r
es
o
f
ea
ch
s
am
p
le,
a
f
ea
tu
r
e
v
ec
to
r
is
co
n
s
tr
u
cted
b
y
s
er
ially
in
ter
co
n
n
ec
tin
g
f
ea
tu
r
es
a
cq
u
ir
ed
f
r
o
m
th
e
r
etr
iev
ed
s
am
p
les.
Fu
r
th
er
m
o
r
e,
ea
ch
h
as
d
-
d
im
en
s
io
n
al
f
ea
tu
r
es.
L
astl
y
,
tr
ain
in
g
s
am
p
les
o
f
d
if
f
er
e
n
t
s
ca
les
in
th
e
lay
er
wer
e
p
ass
e
d
th
r
o
u
g
h
t
o
tr
ain
t
h
e
s
ca
le
f
ilter
.
T
h
e
(
1
2
)
h
as
b
ee
n
u
s
ed
t
o
m
in
im
ize
th
e
er
r
o
r
o
f
co
r
r
elatio
n
r
esp
o
n
s
e
m
ea
s
u
r
ed
ag
ain
s
t th
e
d
esire
d
c
o
r
r
elatio
n
o
u
tp
u
t
:
=
‖
−
∑
∗
=
1
‖
2
+
∑
‖
‖
2
=
1
(
1
2
)
W
h
er
e
is
n
o
r
m
ally
th
e
p
r
ed
i
cted
Gau
s
s
ian
r
esp
o
n
s
e.
,
,
an
d
h
av
e
th
e
s
am
e
d
im
en
s
io
n
an
d
s
ize.
T
h
e
r
eg
u
lar
izatio
n
co
ef
f
icien
t
is
d
en
o
ted
b
y
.
∗
s
tan
d
s
f
o
r
cir
cu
lar
co
r
r
elatio
n
.
T
h
e
(
12
)
c
an
b
e
s
ee
n
as
a
lin
ea
r
least
s
q
u
ar
es
p
r
o
b
lem
.
T
o
s
o
l
v
e
it,
it
s
h
o
u
ld
b
e
tr
a
n
s
f
o
r
m
ed
to
t
h
e
Fo
u
r
ier
d
o
m
ain
.
T
h
e
f
ilter
t
h
at
m
in
im
izes
(
12
)
is
as
(
1
3
)
:
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
er
20
25
:
5
3
4
4
-
5
3
5
4
5348
=
∑
̅
+
=
1
=
(
1
3
)
I
n
th
is
s
itu
atio
n
,
th
e
co
r
r
esp
o
n
d
in
g
v
alu
e
d
is
cr
ete
Fo
u
r
ier
tr
an
s
f
o
r
m
is
r
ep
r
esen
ted
b
y
ca
p
ital
letter
s
.
s
tan
d
s
f
o
r
c
o
m
p
lex
c
o
n
ju
g
atio
n
.
T
o
s
o
lv
e
th
e
p
r
o
b
lem
o
f
r
ed
u
n
d
a
n
t
o
r
f
ix
ed
-
s
ca
le
tr
ac
k
i
n
g
,
we
h
a
v
e
u
s
ed
an
esti
m
atio
n
s
ca
le
ap
p
r
o
ac
h
b
ased
o
n
s
war
m
-
o
p
tim
izatio
n
.
W
e
h
av
e
u
s
ed
th
e
AB
C
alg
o
r
ith
m
[
1
7
]
,
wh
ich
is
o
n
e
o
f
th
e
m
o
s
t
p
o
wer
f
u
l
s
war
m
-
o
p
tim
izatio
n
alg
o
r
ith
m
s
.
AB
C
m
im
ics
th
e
h
o
n
ey
b
ee
s
war
m
’s
clev
er
f
o
r
ag
in
g
b
eh
av
io
r
.
T
h
e
s
ea
r
ch
m
ec
h
an
is
m
o
f
th
e
AB
C
alg
o
r
ith
m
is
in
s
p
ir
ed
b
y
th
e
f
o
r
a
g
in
g
p
atter
n
s
o
b
s
er
v
e
d
in
h
o
n
ey
b
ee
co
lo
n
ies,
u
tili
zin
g
th
r
ee
d
is
tin
ct
r
o
les:
em
p
l
o
y
ed
b
ee
s
,
o
n
lo
o
k
e
r
b
ee
s
,
a
n
d
s
co
u
t
b
ee
s
.
T
h
e
p
r
o
ce
s
s
b
eg
in
s
with
s
co
u
t
b
e
es
d
is
co
v
er
in
g
in
itial
f
o
o
d
s
o
u
r
ce
s
(
p
o
ten
tial
s
o
lu
tio
n
s
)
.
E
m
p
lo
y
ed
b
ee
s
th
e
n
ex
p
lo
it th
ese
s
o
u
r
ce
s
,
s
h
ar
in
g
in
f
o
r
m
atio
n
with
o
n
lo
o
k
er
b
ee
s
,
wh
ich
s
elec
tiv
ely
f
o
cu
s
o
n
p
r
o
m
is
in
g
s
o
lu
tio
n
s
b
ased
o
n
th
eir
f
it
n
ess
(
n
ec
tar
q
u
ality
)
.
I
f
a
f
o
o
d
s
o
u
r
ce
is
e
x
h
au
s
ted
(
n
o
f
u
r
th
e
r
im
p
r
o
v
e
m
en
t)
,
th
e
em
p
lo
y
ed
b
ee
ab
an
d
o
n
s
it
an
d
b
ec
o
m
es
a
s
co
u
t
to
s
ea
r
ch
f
o
r
a
n
ew
s
o
lu
tio
n
.
C
r
u
cially
,
th
e
n
u
m
b
e
r
o
f
em
p
lo
y
ed
b
ee
s
eq
u
als
th
e
n
u
m
b
er
o
f
f
o
o
d
s
o
u
r
ce
s
,
m
ain
tain
in
g
a
b
ala
n
ce
d
p
o
p
u
latio
n
.
T
h
e
A
lg
o
r
it
h
m
1
iter
ates
th
r
o
u
g
h
th
e
f
o
llo
win
g
p
h
ases
:
Alg
o
r
ith
m
1
.
AB
C
alg
o
r
ith
m
I
n
p
u
t:
(
p
o
p
.
s
ize)
,
Ma
x
C
y
cles,
Ma
x
T
im
e
Ou
tp
u
t:
b
est s
o
lu
tio
n
I
n
itialize
s
o
lu
tio
n
s
wh
ile
(
cy
cles
<
Ma
x
C
y
cles
an
d
tim
e
<
Ma
x
T
im
e)
E
m
p
lo
y
ed
b
ee
s
: g
en
er
ate
n
e
w
s
o
lu
tio
n
s
n
ea
r
ex
is
tin
g
o
n
es
On
lo
o
k
er
b
ee
s
: selec
t
s
o
lu
tio
n
s
p
r
o
p
o
r
tio
n
al
to
f
itn
ess
; e
x
p
lo
r
e
n
ea
r
b
y
Sco
u
t
b
ee
s
: r
ep
lace
ab
an
d
o
n
ed
s
o
lu
tio
n
s
with
r
a
n
d
o
m
o
n
es
Up
d
ate
g
lo
b
al
b
est s
o
lu
tio
n
cy
cles
←
cy
cles
+1
en
d
wh
ile
r
etu
r
n
b
est s
o
lu
tio
n
T
h
e
AB
C
m
et
h
o
d
p
r
o
v
i
d
es
a
r
an
d
o
m
d
is
t
r
i
b
u
ti
o
n
o
f
f
o
o
d
s
o
u
r
ce
l
o
c
ati
o
n
s
o
f
s
o
l
u
ti
o
n
s
,
w
h
e
r
e
is
t
h
e
n
u
m
b
er
o
f
o
n
lo
o
k
e
r
b
e
es
o
r
e
m
p
lo
y
ed
b
ee
s
.
E
a
c
h
i
n
d
i
v
i
d
u
al
s
o
lu
ti
o
n
t
ak
es
t
h
e
s
t
r
u
ct
u
r
e
o
f
a
v
ec
t
o
r
d
e
f
i
n
e
d
i
n
d
-
d
im
en
s
io
n
al
s
p
a
c
e.
d
e
n
o
t
es
t
h
e
to
tal
n
u
m
b
er
o
f
p
a
r
a
m
e
te
r
s
s
u
b
jec
t
t
o
o
p
ti
m
iz
ati
o
n
.
A
f
t
er
wa
r
d
s
,
ca
l
cu
lat
e
th
e
v
a
lu
e
o
f
f
i
tn
ess
f
u
n
cti
o
n
f
o
r
p
o
s
s
i
b
le
s
o
l
u
ti
o
n
.
I
n
o
u
r
c
ase
,
t
h
e
f
u
n
ct
io
n
(
1
4
)
o
f
f
it
n
ess
is
:
=
−
1
{
∑
̅
−
1
=
1
−
1
+
}
(1
4
)
I
n
th
e
AB
C
alg
o
r
ith
m
,
a
s
ca
le
p
o
o
l
is
as
s
ig
n
ed
as
th
e
v
ec
to
r
f
o
r
ea
c
h
f
r
am
e
o
f
th
e
v
id
eo
.
Mu
ltip
le
tar
g
ets
o
f
v
ar
io
u
s
s
c
ales
ar
e
ch
o
s
en
f
r
o
m
th
is
p
o
o
l.
At
f
r
am
e
,
AB
C
alg
o
r
ith
m
in
itializes
th
e
v
ec
to
r
with
r
an
d
o
m
v
alu
es
,
w
h
er
e
th
e
f
itn
ess
ev
alu
atio
n
is
p
e
r
f
o
r
m
ed
u
s
in
g
th
e
m
ath
e
m
atica
l
ex
p
r
ess
io
n
d
ef
i
n
ed
a
s
(
14
)
.
−
1
an
d
−
1
ar
e
th
e
f
ilter
co
ef
f
i
cien
ts
at
f
r
am
e
−
1
.
T
h
e
d
if
f
er
e
n
t step
s
o
f
o
u
r
n
ew
o
b
ject
tr
ac
k
e
r
ar
e
s
u
m
m
ar
ized
in
A
lg
o
r
ith
m
2
.
Alg
o
r
ith
m
2
.
T
h
e
f
u
n
d
a
m
en
tal
p
r
o
ce
s
s
o
f
o
u
r
alg
o
r
ith
m
1
: I
n
p
u
t: p
o
s
(
1
)
th
e
s
tar
tin
g
l
o
ca
tio
n
o
f
th
e
o
b
ject
in
th
e
in
iti
al
f
r
am
e,
VGGN
et
-
1
9
p
r
e
-
tr
ai
n
ed
m
o
d
els;
2
: O
u
tp
u
t: Pr
ed
icted
lo
ca
ti
o
n
p
o
s
t; u
p
d
ated
c
o
r
r
elatio
n
f
ilter
s
; u
p
d
ated
s
ca
le
m
o
d
el;
3
: I
n
itiate
f
ilter
m
o
d
el
u
s
in
g
(
1
)
;
4
: I
n
itiate
KF m
o
d
els;
5
: X
0
ex
tr
ac
ted
C
NN
f
ea
tu
r
es
at
in
itial f
r
am
e;
6
: f
o
r
t=
2
,
3
,
….
D
o
7
: E
x
tr
ac
ted
C
NN
f
ea
tu
r
es a
t f
r
am
e
t a
cc
o
r
d
in
g
to
p
o
s
(
t
-
1
)
.
8
: E
q
u
atio
n
(
5
)
to
ca
lcu
late
th
e
r
esp
o
n
s
e
m
ap
s
o
f
f
ilter
;
9
:
U
s
i
n
g
t
h
e
c
a
l
c
u
l
a
t
e
d
p
o
s
(
t
)
a
s
r
e
c
o
r
d
i
n
g
,
p
r
e
d
i
c
t
t
h
e
t
a
r
g
e
t
l
o
c
a
t
i
o
n
p
o
s
k
a
l
m
a
n
(
t
)
u
s
i
n
g
t
h
e
K
a
l
m
a
n
F
i
l
t
e
r
(
K
F
)
;
1
0
: E
x
tr
ac
ted
C
NN
f
ea
tu
r
es o
f
f
r
am
e
t a
cc
o
r
d
in
g
to
p
o
s
k
alm
an
(
t)
,
c
alcu
late
th
e
r
esp
o
n
s
e
m
ap
s
r
eg
ar
d
in
g
(
5
)
;
1
1
:
C
h
o
o
s
e
th
e
m
ax
im
u
m
r
esp
o
n
s
e
m
ap
b
etwe
en
th
e
two
a
n
d
c
o
m
p
u
te
th
e
r
eliab
ilit
y
in
f
o
r
m
atio
n
u
s
in
g
(
8
)
an
d
(
5
)
;
1
2
: if
r
>T
h
r
th
e
n
1
3
: U
s
e
(
8
)
to
d
eter
m
in
e
th
e
ta
r
g
et
lo
ca
tio
n
p
o
s
(
t)
at
th
e
c
u
r
r
en
t f
r
am
e;
14:
Fin
d
th
e
b
est s
ca
le
th
at
m
ax
im
izes (
1
4
)
b
ased
o
n
AB
C
al
g
o
r
ith
m
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
E
n
h
a
n
ce
d
o
b
ject
tr
a
ck
in
g
w
ith
a
r
tifi
cia
l b
ee
co
lo
n
y,
mo
ti
o
n
mo
d
elin
g
…
(
R
a
md
a
n
e
Ta
g
lo
u
t)
5349
15:
Up
d
ate
s
ca
le
m
o
d
el
16:
T
h
e
o
n
lin
e
tr
ain
i
n
g
m
o
d
u
l
e
f
ee
d
s
it with
th
e
ex
tr
ac
ted
f
e
atu
r
es X0
an
d
Xt
an
d
u
p
d
ates
C
F;
17:
else
1
8
: Ch
o
o
s
e
p
o
s
(
t)
th
at
h
as th
e
m
ax
im
u
m
r
esp
o
n
s
e
b
etwe
en
t
h
e
two
.
19:
en
d
if
2
0
: e
n
d
f
o
r
3.
M
E
T
H
O
D
3
.
1
.
T
he
ev
a
lua
t
io
n
o
f
perf
o
rm
a
nce
I
n
th
is
s
ec
tio
n
,
we
will
f
ir
s
t
e
x
am
in
e
th
e
d
etails
o
f
o
u
r
tr
ac
k
er
im
p
lem
en
tatio
n
an
d
th
e
c
o
n
f
ig
u
r
atio
n
o
f
its
p
ar
a
m
eter
s
.
T
h
e
tr
ac
k
er
we
s
u
g
g
ested
was
d
ev
elo
p
ed
in
MA
T
L
AB
2
0
1
8
b
an
d
ev
al
u
ated
o
n
a
d
esk
to
p
co
m
p
u
ter
with
an
I
n
tel
i7
-
4
7
0
0
HQ
C
PU
@
2
.
4
0
GHz
x
8
an
d
8
GB
R
AM
.
I
n
o
r
d
er
to
im
p
l
em
en
t
VGGN
et
-
1
9
[
1
5
]
co
n
v
o
lu
tio
n
al
f
ea
tu
r
e
ex
t
r
ac
tio
n
,
we
u
tili
ze
Ma
tC
o
n
v
Ne
t,
a
p
o
p
u
lar
MA
T
L
AB
d
ee
p
le
ar
n
in
g
f
r
am
ewo
r
k
.
Seco
n
d
ly
,
we
will
ex
h
ib
it
th
e
tr
ac
k
in
g
p
er
f
o
r
m
an
ce
o
f
o
u
r
p
r
o
p
o
s
ed
tr
ac
k
er
i
n
co
m
p
ar
is
o
n
to
v
a
r
io
u
s
s
tate
-
of
-
th
e
-
ar
t
tr
ac
k
er
s
.
W
e
an
aly
ze
th
e
p
e
r
f
o
r
m
an
ce
o
f
o
u
r
s
y
s
tem
u
s
in
g
t
h
e
OT
B
-
2
0
1
5
a
n
d
OT
B
-
2
0
1
3
d
atasets
,
th
at
in
clu
d
e
attr
ib
u
tes
-
b
ased
ev
alu
atio
n
s
,
q
u
an
tit
ativ
e,
an
d
q
u
alitativ
e.
T
o
b
e
m
o
r
e
p
r
ec
is
e,
th
e
tr
ac
k
er
s
ettin
g
s
ar
e
s
et
to
th
e
f
o
llo
win
g
:
T
h
e
r
eg
u
lar
izatio
n
ter
m
p
a
r
a
m
eter
is
s
et
to
=0
.
0
1
,
th
e
p
ad
d
in
g
v
alu
e
f
o
r
th
e
r
e
g
io
n
h
as
b
ee
n
cu
s
to
m
ized
to
1
.
8
,
t
h
e
lear
n
in
g
r
ate
is
=0
.
0
1
,
a
n
d
with
a
v
alu
e
o
f
0
.
1
th
e
f
ilter
k
er
n
e
l
wid
th
is
a
d
ju
s
ted
.
W
ith
weig
h
ts
o
f
1
,
0
.
5
,
an
d
0
.
2
5
,
c
o
r
r
esp
o
n
d
in
g
l
y
,
we
h
a
v
e
u
tili
ze
d
th
e
f
ea
tu
r
es
o
f
co
n
v
5
-
4
,
co
n
v
4
-
4
,
an
d
co
n
v
3
-
4
o
f
VGGN
et
-
19.
Fo
r
t
r
an
s
latio
n
,
KF
ass
ig
n
s
a
v
alu
e
o
f
=
[
25
,
10
,
10
]
,
=
25
to
th
e
co
v
ar
ian
ce
s
o
f
m
o
tio
n
an
d
m
ea
s
u
r
ed
n
o
is
e
.
Fin
ally
,
th
e
co
n
tr
o
l p
a
r
am
ete
r
s
th
e
AB
C
s
ca
le
ar
e
g
iv
en
in
T
ab
le
1
.
T
ab
le
1
.
AB
C
alg
o
r
ith
m
p
a
r
a
m
eter
s
P
a
r
a
me
t
e
r
V
a
l
u
e
M
a
x
i
m
u
m
n
u
m
b
e
r
o
f
i
t
e
r
a
t
i
o
n
s
30
P
o
p
u
l
a
t
i
o
n
s
i
z
e
30
Li
mi
t
v
a
l
u
e
20
A
c
c
e
l
e
r
a
t
i
o
n
c
o
e
f
f
i
c
i
e
n
t
1
D
i
me
n
si
o
n
1
3
.
2
.
E
v
a
lua
t
i
o
n c
rit
er
ia
W
e
u
t
i
li
z
e
t
h
e
OP
E
c
r
i
t
e
r
i
a
[
1
8
]
,
[
1
9
]
t
o
e
v
a
l
u
a
t
e
a
l
l
t
r
a
c
k
e
r
s
o
n
t
h
e
O
T
B
d
a
t
a
s
e
t
,
w
h
i
c
h
i
n
c
l
u
d
e
s
t
w
o
m
e
t
r
i
c
s
:
s
u
c
c
es
s
r
a
t
e
a
n
d
P
r
e
cis
i
o
n
.
P
r
ec
is
i
o
n
is
d
e
f
i
n
e
d
a
s
‖
−
‖
,
w
h
e
r
e
a
n
d
d
e
n
o
t
e
t
h
e
c
e
n
t
e
r
s
o
f
t
h
e
p
r
e
d
i
c
t
e
d
a
n
d
g
r
o
u
n
d
-
t
r
u
t
h
b
o
u
n
d
i
n
g
b
o
x
e
s
,
r
e
s
p
e
c
t
i
v
el
y
.
W
h
e
n
c
o
m
p
a
r
i
n
g
t
h
e
p
e
r
f
o
r
m
a
n
c
e
o
f
d
i
f
f
e
r
e
n
t
t
r
a
c
k
e
r
s
,
a
2
0
-
p
i
x
e
l
d
is
t
a
n
ce
th
r
e
s
h
o
l
d
is
t
y
p
i
ca
l
l
y
u
ti
l
iz
e
d
.
T
h
e
d
e
g
r
e
e
o
f
c
r
o
s
s
o
v
e
r
b
e
twe
e
n
t
h
e
tw
o
b
o
x
e
s
d
e
t
e
r
m
i
n
e
s
t
h
e
s
u
c
c
ess
r
a
t
e
.
A
t
r
a
c
k
i
n
g
i
n
s
t
a
n
ce
i
s
d
e
e
m
e
d
s
u
c
c
e
s
s
f
u
l
w
h
e
n
t
h
e
i
n
t
e
r
s
e
ct
i
o
n
o
v
e
r
u
n
i
o
n
(
I
o
U
)
v
a
l
u
e
r
e
a
c
h
e
s
o
r
e
x
c
e
e
d
s
a
p
r
ed
e
f
i
n
e
d
t
h
r
e
s
h
o
l
d
,
t
y
p
i
c
al
l
y
s
e
t
a
t
0
.
5
.
T
h
e
I
o
U
m
e
t
r
i
c
i
s
c
al
c
u
la
t
e
d
u
s
i
n
g
t
h
e
(
1
5
)
:
=
(
⋂
)
(
⋃
)
(
1
5
)
T
h
e
ar
ea
u
n
d
e
r
th
e
cu
r
v
e
(
AU
C
)
,
co
m
p
u
te
d
f
r
o
m
s
u
cc
ess
p
l
o
ts
ac
r
o
s
s
v
ar
y
in
g
I
o
U
th
r
esh
o
ld
s
,
s
er
v
es
as
a
s
tan
d
ar
d
m
ea
s
u
r
e
o
f
o
v
e
r
all
tr
ac
k
in
g
q
u
ality
.
Su
cc
ess
at
th
e
f
r
am
e
lev
el
is
d
eter
m
i
n
ed
b
y
th
e
o
v
er
lap
r
atio
(
I
o
U)
o
b
tain
ed
f
r
o
m
co
m
p
ar
in
g
t
h
e
p
r
e
d
icted
r
eg
io
n
to
th
e
g
r
o
u
n
d
tr
u
th
,
wh
er
e
tr
ac
k
in
g
is
co
n
s
id
er
ed
co
r
r
ec
t
if
th
is
v
al
u
e
s
u
r
p
ass
es
a
s
p
ec
i
f
ied
t
h
r
e
s
h
o
ld
.
T
h
e
s
u
cc
ess
r
ate
is
th
er
ef
o
r
e
th
e
r
atio
o
f
co
r
r
ec
tly
tr
ac
k
e
d
f
r
a
m
es to
th
e
to
tal
s
eq
u
en
ce
len
g
t
h
.
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
4
.
1
.
Abla
t
io
n
e
x
perim
ent
s
Ab
latio
n
ex
p
er
im
en
ts
o
n
th
e
OT
B
d
ataset
ar
e
p
er
f
o
r
m
ed
to
ev
alu
ate
th
e
p
er
f
o
r
m
an
ce
o
f
e
ac
h
m
o
d
u
le
d
escr
ib
ed
in
th
is
ar
ticle.
T
h
e
DC
F
tr
ac
k
er
is
u
s
ed
as
a
b
aselin
e,
b
u
t
f
ea
tu
r
es
ar
e
ex
tr
ac
ted
u
s
in
g
a
co
n
v
o
lu
tio
n
n
etwo
r
k
.
Fo
r
th
e
p
u
r
p
o
s
e
o
f
t
h
e
ev
alu
atio
n
o
f
d
if
f
er
e
n
t
co
m
p
o
n
e
n
ts
ef
f
icien
cy
,
we
b
u
ilt
th
r
ee
in
d
e
p
en
d
e
n
t
tr
ac
k
er
s
b
y
in
teg
r
atin
g
th
e
b
a
s
elin
e
tr
ac
k
er
with
th
e
AB
C
s
ca
le
an
d
ea
c
h
co
m
p
o
n
en
t:
B
aselin
e
with
AB
C
s
ca
le+
R
A
th
e
b
aselin
e
an
d
th
e
r
eliab
ilit
y
an
aly
s
is
m
o
d
u
le
a
r
e
co
m
b
in
e
d
,
th
e
b
aselin
e
with
AB
C
s
ca
le+
K
F
is
p
r
o
d
u
ce
d
b
y
co
m
b
i
n
in
g
th
e
b
aselin
e
with
th
e
KF
a
n
d
t
h
e
l
ast
o
n
e
is
b
aselin
e
with
AB
C
s
ca
le+
OT
s
ig
n
if
ies
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
er
20
25
:
5
3
4
4
-
5
3
5
4
5350
th
at
th
e
tar
g
et
in
f
o
r
m
atio
n
f
r
o
m
b
o
th
th
e
in
itial
f
r
am
e
an
d
th
e
p
r
esen
t
f
r
am
e
is
u
s
ed
to
tr
ain
th
e
u
p
d
ate
d
f
ilter
.
Mo
r
eo
v
er
,
we
ev
alu
ate
th
e
p
er
f
o
r
m
an
ce
o
f
AB
C
s
ca
le
tech
n
iq
u
e
ag
ain
s
t th
e
tr
a
d
itio
n
al
o
n
e
.
Fig
u
r
es
2
an
d
3
p
r
esen
t
a
s
u
m
m
ar
y
o
f
th
e
o
v
e
r
all
ex
p
er
im
en
tal
o
u
tco
m
es.
Acc
u
r
ac
y
m
e
asu
r
em
en
ts
wer
e
ca
r
r
ied
o
u
t
e
x
p
er
im
e
n
tally
,
as
s
ee
n
in
th
e
lef
t
f
ig
u
r
e.
T
h
e
tr
ac
k
in
g
ac
cu
r
ac
y
at
a
d
is
tan
ce
er
r
o
r
th
r
esh
o
ld
o
f
2
0
is
r
ep
r
esen
ted
b
y
t
h
e
n
u
m
b
er
in
t
h
e
leg
en
d
.
T
h
e
r
ig
h
t
g
r
ap
h
illu
s
tr
ates
th
e
o
v
er
all
s
u
cc
ess
r
ate
o
f
ea
ch
tr
ac
k
er
; it
is
r
ep
r
esen
ted
b
y
th
e
n
u
m
b
e
r
in
th
e
le
g
en
d
(
AUC
).
Fo
r
Fig
u
r
e
2
,
th
e
leg
en
d
s
s
h
o
w
th
e
AUC
an
d
DP
s
co
r
e
at
2
0
p
ix
els
f
o
r
ea
c
h
ab
latio
n
,
b
ased
o
n
o
n
e
-
p
ass
ev
alu
atio
n
o
f
p
r
ec
is
io
n
a
n
d
s
u
cc
ess
p
lo
ts
o
n
th
e
5
0
s
eq
u
en
ce
s
o
f
th
e
OT
B
-
2
0
1
3
d
ataset.
Fig
u
r
e
3
f
o
llo
ws
th
e
s
am
e
ev
alu
atio
n
p
r
o
ce
d
u
r
e
f
o
r
all
1
0
0
s
eq
u
e
n
ce
s
o
f
th
e
OT
B
-
2
0
1
5
d
ataset,
with
th
e
leg
en
d
s
d
is
p
lay
in
g
b
o
th
t
h
e
DP
s
co
r
e
at
2
0
p
ix
els
an
d
th
e
AUC
f
o
r
ea
ch
a
b
latio
n
.
On
OT
B
2
0
1
3
a
n
d
O
T
B
2
0
1
5
,
B
aselin
e
ac
h
iev
es
m
ax
im
u
m
p
r
ec
is
io
n
p
lo
t
p
er
f
o
r
m
a
n
ce
o
f
8
6
.
7
%
an
d
8
5
.
1
%,
r
esp
ec
tiv
ely
.
Ov
er
th
e
o
t
h
er
f
o
u
r
tr
ac
k
er
s
,
OT
B
2
0
1
3
h
as
a
p
r
ec
i
s
io
n
p
er
f
o
r
m
an
ce
i
n
cr
ea
s
e
o
f
0
.
9
%;
1
%;
3
.
3
%;
an
d
7
%;
wh
er
ea
s
OT
B
2
0
1
5
h
as
a
p
r
ec
is
io
n
p
er
f
o
r
m
an
ce
im
p
r
o
v
em
en
t
o
f
1
.
6
%;
2
.
2
%;
3
.
8
%;
an
d
5
.
9
%.
I
n
th
e
s
am
e
way
,
B
aselin
e
with
AB
C
s
ca
le+
All
ac
h
iev
es
p
r
ec
i
s
io
n
s
co
r
es
in
th
e
s
u
cc
ess
p
lo
ts
6
6
.
5
%
an
d
6
2
.
5
%
o
n
OT
B
2
0
1
3
an
d
OT
B
2
0
1
5
,
r
esp
ec
tiv
ely
.
OT
B
2
0
1
3
s
h
o
ws
s
u
cc
ess
in
cr
ea
s
es
o
f
1
.
1
%,
3
.
9
%,
4
.
7
%
,
an
d
1
0
%,
alth
o
u
g
h
OT
B
2
0
1
5
s
h
o
ws
s
u
cc
ess
g
ain
s
o
f
1
.
5
%,
2
.
0
%,
2
.
3
%
,
an
d
6
.
3
% in
c
o
m
p
a
r
is
o
n
to
th
e
o
th
er
f
o
u
r
t
r
ac
k
er
s
.
Fig
u
r
e
2
.
Pre
cisi
o
n
a
n
d
s
u
cc
es
s
p
lo
ts
o
f
th
e
ab
latio
n
s
tu
d
y
o
n
t
h
e
OT
B
-
2
0
1
3
d
atase
t
Fig
u
r
e
3
.
Pre
cisi
o
n
a
n
d
s
u
cc
es
s
p
lo
ts
o
f
th
e
ab
latio
n
s
tu
d
y
o
n
t
h
e
OT
B
-
2
0
1
5
d
atase
t
KF
s
h
o
ws
th
e
s
m
alles
t
im
p
r
o
v
em
en
t
in
b
en
ch
m
ar
k
tr
ac
k
e
r
p
er
f
o
r
m
a
n
ce
r
elativ
e
to
th
e
o
th
er
two
m
o
d
u
les.
T
h
e
p
o
o
r
p
r
ec
is
io
n
o
f
th
e
b
en
c
h
m
ar
k
tr
ac
k
e
r
lead
s
to
m
o
r
e
m
ea
s
u
r
em
e
n
t
er
r
o
r
in
th
e
K
F
p
r
o
ce
s
s
,
p
r
o
d
u
cin
g
m
ed
io
cr
e
p
r
e
d
ictio
n
r
esu
lts
.
T
h
e
r
ef
o
r
e
,
s
in
g
le
KF
m
o
d
u
le
en
h
a
n
ce
s
th
e
b
en
ch
m
ar
k
tr
ac
k
er
ef
f
icien
cy
s
lig
h
tly
co
m
p
a
r
ed
to
th
e
o
th
er
th
r
ee
m
o
d
u
les.
Ad
d
itio
n
ally
,
th
e
OT
m
o
d
u
le
m
ak
es
an
im
p
o
r
tan
t
co
n
tr
ib
u
tio
n
to
im
p
r
o
v
in
g
h
o
w
well
th
e
p
er
f
o
r
m
a
n
ce
tr
ac
k
er
wo
r
k
s
.
T
h
e
m
o
d
u
le
c
o
n
s
is
ten
tly
r
etain
s
th
e
in
itial f
r
am
e
tar
g
et
in
f
o
r
m
atio
n
.
Acc
o
r
d
in
g
to
th
e
r
esu
lts
o
f
th
i
s
s
tu
d
y
,
a
n
e
f
f
ec
tiv
e
t
r
ac
k
in
g
f
r
am
ewo
r
k
r
eq
u
ir
es
ac
cu
r
ate
in
f
o
r
m
atio
n
ab
o
u
t
o
b
jects.
T
h
e
R
A
m
o
d
u
l
e
h
as
th
e
m
o
s
t
im
p
ac
t
o
n
im
p
r
o
v
in
g
th
e
ef
f
icien
cy
o
f
th
e
b
aselin
e
tr
ac
k
er
.
Sin
ce
it
is
ab
le
to
m
ain
tain
p
r
ec
is
e
o
b
ject
in
f
o
r
m
atio
n
wh
ile
av
o
id
in
g
tr
ac
k
in
g
d
r
if
t
an
d
lo
s
s
in
v
ar
io
u
s
s
ce
n
ar
io
s
(
e.
g
.
,
o
cc
lu
s
io
n
an
d
ex
ten
d
ed
s
eq
u
en
ce
s
)
.
Fig
u
r
es
4
an
d
5
s
h
o
w
th
at
th
e
p
ap
er
s
u
g
g
ested
tr
ac
k
er
t
ec
h
n
iq
u
e
o
n
th
e
OT
B
-
1
0
0
d
ata
s
et
h
as
co
m
p
ar
ab
le
b
en
ef
its
: D
P r
ate
p
er
f
o
r
m
an
ce
was h
ig
h
er
with
t
h
e
tr
ac
k
er
we
s
u
g
g
ested
co
m
p
ar
ed
to
o
n
e
with
o
u
t
th
e
AB
C
s
ca
le,
wh
ich
was
8
6
.
7
%
v
s
.
8
5
.
3
8
%
in
OT
B
-
2
0
1
3
an
d
8
5
.
1
%
v
s
.
8
4
.
5
3
%
in
OT
B
-
2
0
1
5
.
I
n
ter
m
s
o
f
co
v
er
ag
e
s
u
cc
ess
r
ate,
th
e
p
er
f
o
r
m
a
n
ce
was
co
m
p
ar
ab
le
to
6
6
.
5
8
%
an
d
6
3
.
2
%,
6
2
.
5
5
%
an
d
6
2
.
1
%
in
OT
B
-
2
0
1
3
an
d
OT
B
-
2
0
1
5
r
e
s
p
ec
tiv
ely
.
I
n
Fig
u
r
e
4
,
th
e
p
r
ec
is
io
n
an
d
s
u
cc
ess
p
lo
ts
,
an
n
o
tated
with
DP
s
co
r
es
at
2
0
p
ix
els
an
d
AUC
v
alu
es
in
th
eir
leg
en
d
s
,
wer
e
o
b
tain
ed
f
r
o
m
ex
p
er
im
e
n
ts
co
n
d
u
cted
o
n
th
e
OT
B
-
2
0
1
3
b
en
c
h
m
ar
k
.
T
h
e
an
aly
s
is
was
ca
r
r
ied
o
u
t
u
s
in
g
th
e
s
tan
d
ar
d
o
n
e
-
p
ass
ev
alu
atio
n
p
r
o
to
c
o
l
ac
r
o
s
s
all
5
0
s
eq
u
en
ce
s
f
o
r
ea
c
h
ab
la
tio
n
ex
p
er
i
m
en
ts
with
an
d
wit
h
o
u
t
AB
C
ap
p
r
o
ac
h
.
I
n
Fig
u
r
e
5
,
th
e
e
v
alu
atio
n
o
n
OT
B
-
2
0
1
5
was
co
n
d
u
cted
u
s
in
g
o
n
e
-
p
ass
ev
alu
atio
n
ac
r
o
s
s
1
0
0
s
eq
u
en
ce
s
,
f
r
o
m
wh
ic
h
b
o
th
s
u
cc
ess
an
d
p
r
ec
is
io
n
p
l
o
ts
wer
e
d
e
r
iv
ed
.
L
eg
en
d
s
ac
c
o
m
p
an
y
in
g
th
ese
p
lo
ts
d
is
p
lay
th
e
DP
s
co
r
e
at
2
0
p
ix
els
alo
n
g
with
th
e
AUC v
alu
es f
o
r
ab
latio
n
e
x
p
er
im
en
ts
th
at
eith
er
in
clu
d
e
o
r
ex
clu
d
e
th
e
AB
C
ap
p
r
o
ac
h
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
E
n
h
a
n
ce
d
o
b
ject
tr
a
ck
in
g
w
ith
a
r
tifi
cia
l b
ee
co
lo
n
y,
mo
ti
o
n
mo
d
elin
g
…
(
R
a
md
a
n
e
Ta
g
lo
u
t)
5351
Fig
u
r
e
4
.
Pre
cisi
o
n
a
n
d
s
u
cc
es
s
p
lo
ts
co
m
p
ar
in
g
th
e
tr
ac
k
er
with
an
d
with
o
u
t th
e
AB
C
m
o
d
u
le
o
n
th
e
OT
B
-
2
0
1
3
d
ataset
Fig
u
r
e
5
.
Pre
cisi
o
n
a
n
d
s
u
cc
es
s
p
lo
ts
co
m
p
ar
in
g
th
e
tr
ac
k
er
with
an
d
with
o
u
t th
e
AB
C
m
o
d
u
le
o
n
th
e
OT
B
-
2
0
1
5
d
ataset
T
h
e
OT
B
d
ataset
h
as
b
ee
n
m
eticu
lo
u
s
ly
an
n
o
tated
with
1
1
u
n
iq
u
e
attr
ib
u
tes,
ea
ch
o
f
wh
i
ch
p
o
s
es
a
d
if
f
er
en
t
d
if
f
icu
lty
.
T
h
ese
f
ea
t
u
r
es
in
clu
d
e:
o
cc
l
u
s
io
n
,
f
ast
m
o
tio
n
,
d
ef
o
r
m
atio
n
,
lo
w
r
eso
lu
tio
n
,
b
ac
k
g
r
o
u
n
d
clu
tter
,
s
ca
le
v
ar
iatio
n
,
in
-
p
l
an
e
r
o
tatio
n
,
o
u
t
-
of
-
v
iew,
illu
m
in
atio
n
v
a
r
iatio
n
,
o
u
t
-
of
-
p
lan
e
r
o
tatio
n
,
an
d
m
o
tio
n
b
lu
r
.
T
h
ese
attr
ib
u
te
-
b
ased
s
u
b
s
ets
ar
e
cr
u
cial
f
o
r
ass
ess
in
g
tr
ac
k
er
p
er
f
o
r
m
an
ce
an
d
m
a
k
in
g
im
p
r
o
v
em
e
n
ts
.
Fo
r
ea
ch
tr
ac
k
er
,
t
h
e
p
r
ec
is
io
n
an
d
AUC
v
alu
es
ar
e
s
h
o
wn
in
T
ab
les
2
an
d
3
f
o
r
th
e
attr
i
b
u
te
-
b
ased
s
u
b
s
ets.
B
a
s
elin
eAll
clea
r
ly
o
u
tp
er
f
o
r
m
s
th
e
ch
allen
g
e
ac
r
o
s
s
th
e
b
o
ar
d
,
an
d
th
is
h
o
ld
s
tr
u
e
f
o
r
all
attr
ib
u
te
s
u
b
s
ets.
A
s
we
ca
n
s
ee
,
B
a
s
eli
n
e+
All o
u
tp
er
f
o
r
m
s
all
o
th
er
tr
ac
k
er
s
in
ev
er
y
attr
ib
u
te
s
u
b
s
et.
W
e
r
ea
lized
th
a
t
o
u
r
p
r
o
p
o
s
ed
f
r
am
ewo
r
k
g
r
e
atly
im
p
r
o
v
ed
th
e
b
aselin
e
t
r
ac
k
er
p
er
f
o
r
m
an
ce
an
d
th
at
th
e
p
er
f
o
r
m
an
ce
in
cr
ea
s
e
ac
co
r
d
i
n
g
to
ea
ch
m
o
d
u
le
was
in
lin
e
with
t
h
e
o
v
er
all
r
esu
lts
.
I
n
ad
d
itio
n
,
th
i
s
d
em
o
n
s
tr
ates
th
at
r
eliab
le
tar
g
et
in
f
o
r
m
atio
n
is
cr
itical
f
o
r
th
e
tr
ac
k
in
g
p
r
o
ce
s
s
an
d
th
at
KF
also
s
u
p
p
lies
c
r
itical
s
u
p
p
lem
en
tal
in
f
o
r
m
atio
n
f
o
r
ef
f
ec
tiv
e
t
r
ac
k
in
g
o
p
e
r
atio
n
s
.
T
ab
le
2
.
Pre
cisi
o
n
r
esu
lts
o
f
O
T
B
1
0
0
ab
latio
n
e
v
alu
ate
d
o
n
1
1
d
is
tin
ct
attr
ib
u
tes
A
t
t
r
i
b
u
t
e
B
a
se
+
A
l
l
B
a
se
+
R
A
B
a
se
+
O
T
B
a
se
+
K
F
B
a
se
O
c
c
l
u
si
o
n
0
.
8
0
2
0
0
.
7
9
1
8
0
.
7
3
2
1
0
.
7
7
4
8
0
.
7
7
1
B
a
c
k
g
r
o
u
n
d
c
l
u
t
t
e
r
s
0
.
8
4
5
2
0
.
8
1
1
7
0
.
8
0
8
4
0
.
7
9
8
5
0
.
7
8
6
F
a
st
m
o
t
i
o
n
0
.
8
0
5
0
0
.
8
0
4
3
0
.
7
5
5
7
0
.
8
0
3
2
0
.
7
0
6
I
l
l
u
mi
n
a
t
i
o
n
v
a
r
i
a
t
i
o
n
0
.
8
3
6
4
0
.
8
0
5
2
0
.
8
0
6
2
0
.
7
9
7
5
0
.
7
8
6
D
e
f
o
r
ma
t
i
o
n
0
.
8
1
6
4
0
.
7
9
2
9
0
.
8
0
9
9
0
.
7
4
5
9
0
.
8
0
5
In
-
p
l
a
n
e
r
o
t
a
t
i
o
n
0
.
8
3
3
8
0
.
8
3
2
7
0
.
7
7
2
0
0
.
8
3
1
3
0
.
7
8
9
S
c
a
l
e
v
a
r
i
a
t
i
o
n
0
.
7
9
4
1
0
.
7
7
7
7
0
.
7
3
6
5
0
.
7
7
0
3
0
.
7
5
4
Lo
w
r
e
s
o
l
u
t
i
o
n
0
.
8
3
1
6
0
.
8
2
8
7
0
.
8
0
9
9
0
.
7
9
4
4
0
.
7
3
7
M
o
t
i
o
n
b
l
u
r
0
.
8
0
0
7
0
.
7
9
6
8
0
.
7
5
6
6
0
.
7
8
6
8
0
.
7
3
6
O
u
t
-
of
-
p
l
a
n
e
r
o
t
a
t
i
o
n
0
.
8
2
4
1
0
.
7
8
0
8
0
.
7
7
0
8
0
.
7
7
6
4
0
.
7
8
2
O
u
t
-
of
-
v
i
e
w
0
.
7
3
7
2
0
.
6
7
2
1
0
.
6
2
1
4
0
.
6
7
1
4
0
.
6
6
7
T
ab
le
3
.
R
esu
lts
o
f
th
e
OT
B
1
0
0
ab
latio
n
e
x
p
er
im
en
t: AU
C
v
alu
es a
cr
o
s
s
1
1
attr
ib
u
te
ca
teg
o
r
ies
A
t
t
r
i
b
u
t
e
B
a
se
+
A
l
l
B
a
se
+
R
A
B
a
se
+
O
T
B
a
se
+
K
F
B
a
se
O
c
c
l
u
si
o
n
0
.
6
1
1
1
0
.
6
0
7
5
0
.
5
9
9
7
0
.
6
0
9
4
0
.
5
3
9
B
a
c
k
g
r
o
u
n
d
c
l
u
t
t
e
r
s
0
.
6
2
4
5
0
.
6
0
7
8
0
.
6
0
3
0
0
.
6
0
4
1
0
.
5
8
3
F
a
st
m
o
t
i
o
n
0
.
5
8
2
9
0
.
5
9
5
5
0
.
5
9
3
6
0
.
5
9
1
0
0
.
5
3
9
I
l
l
u
mi
n
a
t
i
o
n
v
a
r
i
a
t
i
o
n
0
.
6
1
4
0
0
.
5
9
3
0
0
.
5
8
9
2
0
.
6
0
8
8
0
.
5
5
6
D
e
f
o
r
ma
t
i
o
n
0
.
5
9
9
3
0
.
6
0
4
3
0
.
6
0
3
2
0
.
5
7
2
0
0
.
5
3
1
In
-
p
l
a
n
e
r
o
t
a
t
i
o
n
0
.
6
0
1
7
0
.
5
8
5
2
0
.
5
7
3
1
0
.
5
7
9
9
0
.
5
6
8
S
c
a
l
e
v
a
r
i
a
t
i
o
n
0
.
5
8
6
8
0
.
5
7
4
5
0
.
5
6
9
1
0
.
5
7
1
1
0
.
5
3
2
Lo
w
r
e
s
o
l
u
t
i
o
n
0
.
5
2
3
3
0
.
5
1
8
8
0
.
5
1
5
1
0
.
5
0
1
9
0
.
4
9
6
M
o
t
i
o
n
b
l
u
r
0
.
6
4
7
0
0
.
6
3
3
3
0
.
6
1
7
8
0
.
6
2
1
2
0
.
5
4
8
O
u
t
-
of
-
p
l
a
n
e
r
o
t
a
t
i
o
n
0
.
6
0
7
8
0
.
5
9
1
1
0
.
5
8
2
4
0
.
5
9
1
8
0
.
5
5
1
O
u
t
-
of
-
v
i
e
w
0
.
6
1
4
4
0
.
6
0
9
2
0
.
5
8
5
4
0
.
5
9
8
8
0
.
4
8
2
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
er
20
25
:
5
3
4
4
-
5
3
5
4
5352
4
.
2
.
Co
m
pa
riso
n wit
h o
t
her
t
ra
ck
er
s
T
o
co
n
d
u
ct
a
m
o
r
e
th
o
r
o
u
g
h
an
aly
s
is
an
d
ev
alu
atio
n
o
f
th
e
p
r
o
p
o
s
ed
tr
ac
k
er
,
we
p
er
f
o
r
m
a
co
m
p
ar
ativ
e
s
tu
d
y
a
g
ain
s
t
s
ev
er
al
r
ep
r
esen
tativ
e
tr
ac
k
in
g
m
eth
o
d
s
,
in
clu
d
i
n
g
DSST
[
2
0
]
,
I
FC
T
[
2
1
]
,
s
elec
tiv
e
p
ar
t
-
b
ased
co
r
r
elatio
n
f
ilter
(
SP
C
F)
[
2
2
]
,
Z
h
ao
et
a
l.
[
2
3
]
,
k
er
n
elize
d
co
r
r
elatio
n
f
ilter
(
KC
F)
[
2
4
]
,
L
iu
et
a
l.
[
2
5
]
,
Au
to
tr
ac
k
[
2
6
]
,
Ad
_
SASTC
A
[
2
7
]
,
Sh
u
et
a
l.
[
2
8
]
,
lear
n
i
n
g
ad
a
p
tiv
e
d
is
cr
im
in
ativ
e
co
r
r
elatio
n
f
ilter
s
(
L
ADCF
)
_
HC
[
2
9
]
,
MCMC
F
[
3
0
]
.
Ad
d
itio
n
ally
,
we
p
er
f
o
r
m
ed
ex
p
er
im
en
ts
u
s
in
g
OT
B
2
0
1
3
an
d
OT
B
2
0
1
5
as
well.
T
h
e
p
r
o
p
o
s
ed
tr
ac
k
in
g
m
eth
o
d
attain
s
s
u
p
er
io
r
r
esu
l
ts
o
n
th
e
OT
B
2
0
1
3
b
en
ch
m
ar
k
,
r
ea
ch
in
g
a
p
r
ec
is
io
n
s
co
r
e
o
f
0
.
8
6
7
0
a
n
d
a
s
u
cc
ess
r
ate
o
f
0
.
6
6
5
8
,
as
illu
s
t
r
ated
in
Fig
u
r
e
6
.
I
n
co
m
p
a
r
is
o
n
,
SP
C
F
r
ep
o
r
ts
0
.
8
5
9
/
0
.
6
2
8
,
L
iu
et
a
l.
[
2
5
]
0
.
8
5
2
3
/
0
.
6
3
0
,
I
FC
T
0
.
8
3
5
/
0
.
6
3
7
,
Z
h
ao
et
a
l.
[
2
3
]
0
.
8
2
5
/
0
.
6
6
2
,
Au
to
t
r
ac
k
0
.
8
2
5
/0
.
6
1
9
,
DSST
0
.
7
4
7
/
0
.
5
5
7
,
a
n
d
KC
F
0
.
7
4
0
/
0
.
5
1
4
in
ter
m
s
o
f
p
r
ec
is
io
n
an
d
s
u
cc
ess
,
r
esp
ec
tiv
ely
.
T
h
e
r
ep
o
r
ted
r
esu
lts
v
alid
ate
th
e
ef
f
icien
cy
o
f
th
e
p
r
o
p
o
s
ed
tech
n
i
q
u
e
in
m
ain
tain
in
g
ac
cu
r
ate
an
d
r
o
b
u
s
t
tr
ac
k
in
g
ac
r
o
s
s
d
iv
er
s
e
s
ce
n
ar
io
s
.
E
x
p
er
im
en
tal
v
alid
atio
n
o
n
th
e
O
T
B
2
0
1
5
b
en
ch
m
a
r
k
co
n
f
ir
m
s
th
at
th
e
p
r
o
p
o
s
ed
t
r
ac
k
er
attain
s
s
tate
-
of
-
th
e
-
ar
t
p
er
f
o
r
m
an
ce
.
I
t
s
ec
u
r
es
th
e
h
ig
h
est
p
r
ec
is
io
n
(
0
.
8
5
1
0
)
an
d
a
co
m
p
etitiv
e
s
u
cc
ess
r
ate
(
0
.
6
2
5
5
)
am
o
n
g
s
ev
en
ev
alu
ated
tr
ac
k
er
s
,
th
e
r
eb
y
v
alid
atin
g
its
d
esig
n
o
b
jectiv
es
f
o
r
ac
h
iev
i
n
g
h
ig
h
ac
cu
r
ac
y
an
d
r
o
b
u
s
tn
ess
in
co
m
p
lex
v
is
u
al
tr
ac
k
in
g
en
v
ir
o
n
m
en
ts
,
as
s
h
o
wn
in
Fig
u
r
e
7
.
Fig
u
r
e
6
.
Su
cc
ess
(
AUC)
an
d
p
r
ec
is
io
n
(
2
0
-
p
ix
el
t
h
r
esh
o
ld
)
co
m
p
ar
is
o
n
o
f
th
e
p
r
o
p
o
s
ed
tr
ac
k
er
with
s
ev
en
o
th
er
s
o
n
th
e
OT
B
2
0
1
3
d
ataset
Fig
u
r
e
7
.
Per
f
o
r
m
an
c
e
co
m
p
ar
is
o
n
o
f
th
e
p
r
o
p
o
s
ed
tr
ac
k
e
r
w
ith
o
th
er
s
o
n
th
e
OT
B
2
0
1
5
d
at
aset,
b
ased
o
n
s
u
cc
ess
r
ates a
n
d
p
r
ec
is
io
n
5.
CO
NCLU
SI
O
N
T
h
is
r
esear
ch
in
tr
o
d
u
ce
s
a
r
o
b
u
s
t
v
is
u
al
tr
ac
k
in
g
ap
p
r
o
ac
h
co
n
s
is
tin
g
o
f
f
o
u
r
m
ain
c
o
m
p
o
n
e
n
ts
:
an
ex
am
in
atio
n
o
f
r
eliab
ilit
y
m
o
d
u
le,
ac
c
u
r
ate
o
n
lin
e
tr
ain
i
n
g
an
d
u
p
d
ate
m
o
d
u
le,
KF m
o
d
u
le,
an
d
AB
C
s
ca
le
m
eth
o
d
.
T
h
e
r
eliab
ilit
y
an
al
y
s
is
m
o
d
u
le
ass
es
s
es
th
e
tr
a
ck
in
g
p
r
o
ce
s
s
,
d
ec
id
in
g
wh
e
th
er
to
in
itiate
th
e
tr
ain
in
g
u
p
d
ate
p
h
ase
to
p
r
e
v
en
t
n
o
is
e
in
tr
o
d
u
ctio
n
.
T
h
e
r
eliab
le
o
n
lin
e
tr
ain
in
g
u
p
d
ate
m
o
d
u
le
m
e
r
g
es
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
E
n
h
a
n
ce
d
o
b
ject
tr
a
ck
in
g
w
ith
a
r
tifi
cia
l b
ee
co
lo
n
y,
mo
ti
o
n
mo
d
elin
g
…
(
R
a
md
a
n
e
Ta
g
lo
u
t)
5353
in
f
o
r
m
atio
n
f
r
o
m
th
e
in
itial
an
d
cu
r
r
e
n
t
f
r
am
es,
p
r
eser
v
in
g
tr
u
s
two
r
th
y
tar
g
et
in
f
o
r
m
atio
n
.
T
h
e
KF
m
o
d
u
le
s
im
u
lates
in
ter
-
f
r
am
e
m
o
tio
n
,
o
f
f
er
in
g
v
alu
a
b
le
s
u
p
p
le
m
en
tal
d
ata.
T
h
e
p
r
o
p
o
s
ed
tech
n
iq
u
e
e
n
h
an
ce
s
tr
ac
k
in
g
p
er
f
o
r
m
a
n
ce
in
ch
alle
n
g
in
g
s
ce
n
ar
io
s
,
ad
d
r
ess
in
g
ap
p
ea
r
an
ce
ch
an
g
es,
tr
ac
k
in
g
d
r
i
f
t,
an
d
o
cc
lu
s
io
n
s
.
T
h
e
AB
C
s
ca
le
ap
p
r
o
ac
h
,
b
ased
o
n
th
e
AB
C
o
p
tim
izatio
n
a
lg
o
r
ith
m
,
m
itig
ates
s
ca
le
v
ar
iatio
n
is
s
u
es
in
tar
g
et
tr
ac
k
in
g
,
en
h
a
n
cin
g
ac
cu
r
ac
y
an
d
r
esil
ien
ce
.
W
e
ass
ess
e
d
o
u
r
p
r
o
p
o
s
ed
tr
ac
k
in
g
f
r
a
m
ewo
r
k
u
s
in
g
t
h
e
s
tan
d
ar
d
b
e
n
ch
m
ar
k
d
atasets
OT
B
2
0
1
3
an
d
OT
B
2
0
1
5
.
T
h
e
r
esu
lts
s
h
o
w
th
at
o
u
r
tr
ac
k
in
g
m
eth
o
d
ac
h
ie
v
ed
to
p
p
er
f
o
r
m
a
n
ce
,
s
ec
u
r
in
g
f
ir
s
t
p
lace
o
n
b
o
th
OT
B
2
0
1
3
an
d
OT
B
2
0
1
5
d
atasets
.
T
h
e
tr
ac
k
in
g
r
esu
lts
co
n
f
ir
m
th
at
o
u
r
ap
p
r
o
ac
h
d
eliv
er
s
s
tate
-
of
-
th
e
-
ar
t
p
er
f
o
r
m
a
n
ce
lev
el
s
.
Ho
wev
er
,
we
o
b
s
er
v
ed
a
li
m
itatio
n
wh
er
e
o
u
r
tr
ac
k
er
ex
p
e
r
ien
ce
s
d
if
f
icu
lti
es
in
m
ain
tain
in
g
ac
cu
r
ate
p
o
s
itio
n
in
g
wh
en
tar
g
et
o
b
jects
u
n
d
er
g
o
in
-
p
lan
e
r
o
tatio
n
.
T
h
is
wea
k
n
ess
p
r
esen
ts
an
o
p
p
o
r
tu
n
it
y
f
o
r
im
p
r
o
v
e
m
en
t in
f
u
t
u
r
e
d
e
v
elo
p
m
e
n
t p
h
ases
.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
No
f
u
n
d
in
g
was p
r
o
v
id
e
d
to
s
u
p
p
o
r
t th
e
co
n
d
u
ct
o
f
th
is
r
ese
ar
ch
.
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
R
am
d
an
e
T
ag
lo
u
t
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
B
ilal Sao
u
d
✓
✓
✓
✓
✓
✓
✓
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
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
T
h
e
au
th
o
r
s
af
f
ir
m
th
at
t
h
ey
ar
e
n
o
t
awa
r
e
o
f
an
y
f
i
n
an
cia
l
o
r
p
er
s
o
n
al
a
f
f
iliatio
n
s
th
at
co
u
ld
h
av
e
in
ap
p
r
o
p
r
iately
in
f
l
u
en
ce
d
th
e
r
esear
ch
an
d
f
in
d
i
n
g
s
p
r
esen
t
ed
in
th
is
m
an
u
s
cr
ip
t.
DATA AV
AI
L
AB
I
L
I
T
Y
T
h
e
d
atasets
em
p
lo
y
ed
t
o
g
e
n
er
ate
th
e
r
esu
lts
o
f
th
is
wo
r
k
ar
e
av
ailab
le
f
r
o
m
t
h
e
co
r
r
esp
o
n
d
in
g
au
th
o
r
u
p
o
n
r
ea
s
o
n
a
b
le
r
eq
u
est.
RE
F
E
R
E
NC
E
S
[
1
]
M
.
S
.
H
o
ssa
i
n
e
t
a
l
.
,
“
A
u
t
o
mat
i
c
n
a
v
i
g
a
t
i
o
n
a
n
d
se
l
f
-
d
r
i
v
i
n
g
t
e
c
h
n
o
l
o
g
y
i
n
a
g
r
i
c
u
l
t
u
r
a
l
mac
h
i
n
e
r
y
:
a
s
t
a
t
e
-
of
-
t
h
e
-
a
r
t
sy
s
t
e
m
a
t
i
c
r
e
v
i
e
w
,
”
I
EEE
Ac
c
e
ss
,
v
o
l
.
1
3
,
p
p
.
9
4
3
7
0
‒
9
4
4
0
1
,
2
0
2
5
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
5
.
3
5
7
3
3
2
4
.
[
2
]
K
.
R
e
z
a
e
e
,
“
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
n
d
f
a
c
i
a
l
r
e
c
o
g
n
i
t
i
o
n
f
o
r
d
o
w
n
sy
n
d
r
o
m
e
d
e
t
e
c
t
i
o
n
:
a
c
o
m
p
r
e
h
e
n
si
v
e
r
e
v
i
e
w
,
”
C
o
m
p
u
t
e
rs
i
n
H
u
m
a
n
B
e
h
a
v
i
o
r R
e
p
o
r
t
s
,
v
o
l
.
1
7
,
2
0
2
5
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
h
b
r
.
2
0
2
5
.
1
0
0
6
0
0
.
[
3
]
T.
-
L.
L
e
,
H
.
N
.
-
X
u
a
n
,
a
n
d
S
.
-
A
.
K
i
m,
“
H
a
n
d
d
r
a
w
i
n
g
-
b
a
s
e
d
d
a
y
l
i
g
h
t
a
n
a
l
y
s
i
s u
s
i
n
g
d
e
e
p
l
e
a
r
n
i
n
g
a
n
d
a
u
g
m
e
n
t
e
d
r
e
a
l
i
t
y
,
”
Re
s
u
l
t
s
i
n
En
g
i
n
e
e
r
i
n
g
,
v
o
l
.
2
2
,
2
0
2
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
r
i
n
e
n
g
.
2
0
2
4
.
1
0
2
1
1
9
.
[
4
]
A
.
W
a
t
s
o
n
a
n
d
V
.
W
.
-
O
’
C
o
n
n
o
r
,
“
Th
e
p
r
o
m
i
se
o
f
a
r
t
i
f
i
c
i
a
l
i
n
t
e
l
l
i
g
e
n
c
e
i
n
h
e
a
l
t
h
:
p
o
r
t
r
a
y
a
l
s
o
f
e
m
e
r
g
i
n
g
h
e
a
l
t
h
c
a
r
e
t
e
c
h
n
o
l
o
g
i
e
s,
”
S
o
c
i
o
l
o
g
y
o
f
H
e
a
l
t
h
&
I
l
l
n
e
ss
,
v
o
l
.
4
7
,
n
o
.
1
,
2
0
2
5
,
d
o
i
:
1
0
.
1
1
1
1
/
1
4
6
7
-
9
5
6
6
.
1
3
8
4
0
.
[
5
]
G
.
S
u
r
e
sh
,
P
.
R
e
d
d
y
,
M
.
P
r
a
v
e
e
n
,
P
.
A
r
u
n
,
a
n
d
K
.
J
a
y
a
n
t
h
,
“
R
e
a
l
t
i
m
e
r
o
o
m
a
u
t
o
ma
t
i
o
n
u
si
n
g
I
o
T
d
e
v
i
c
e
,
”
T
u
rk
i
sh
J
o
u
rn
a
l
o
f
C
o
m
p
u
t
e
r
a
n
d
M
a
t
h
e
m
a
t
i
c
s E
d
u
c
a
t
i
o
n
(
T
U
RC
O
MA
T
)
,
v
o
l
.
1
4
,
p
p
.
4
2
2
‒
4
3
4
,
2
0
2
3
,
d
o
i
:
1
0
.
1
7
7
6
2
/
t
u
r
c
o
ma
t
.
v
1
4
i
2
.
1
3
6
7
2
.
[
6
]
P
.
K
a
d
a
m,
G
.
F
a
n
g
,
a
n
d
J
.
J.
Z
o
u
,
“
O
b
j
e
c
t
t
r
a
c
k
i
n
g
u
si
n
g
c
o
m
p
u
t
e
r
v
i
si
o
n
:
a
r
e
v
i
e
w
,
”
C
o
m
p
u
t
e
rs
,
v
o
l
.
1
3
,
n
o
.
6
,
2
0
2
4
,
d
o
i
:
1
0
.
3
3
9
0
/
c
o
mp
u
t
e
r
s1
3
0
6
0
1
3
6
.
[
7
]
A
.
Y
i
l
maz
,
O
.
Ja
v
e
d
,
a
n
d
M
.
S
h
a
h
,
“
O
b
j
e
c
t
t
r
a
c
k
i
n
g
:
a
s
u
r
v
e
y
,
”
AC
M
C
o
m
p
u
t
i
n
g
S
u
rve
y
s
,
v
o
l
.
3
8
,
n
o
.
4
,
p
p
.
1
‒
4
5
,
2
0
0
6
,
d
o
i
:
1
0
.
1
1
4
5
/
1
1
7
7
3
5
2
.
1
1
7
7
3
5
5
.
[
8
]
A
.
S
me
u
l
d
e
r
s
,
D
.
C
h
u
,
R
.
C
u
c
c
h
i
a
r
a
,
S
.
C
a
l
d
e
r
a
r
a
,
A
.
D
e
h
g
h
a
n
,
a
n
d
M
.
S
h
a
h
,
“
V
i
s
u
a
l
t
r
a
c
k
i
n
g
:
a
n
e
x
p
e
r
i
me
n
t
a
l
s
u
r
v
e
y
,
”
I
EE
E
T
ra
n
s
a
c
t
i
o
n
s
o
n
P
a
t
t
e
r
n
A
n
a
l
y
si
s
a
n
d
Ma
c
h
i
n
e
I
n
t
e
l
l
i
g
e
n
c
e
,
v
o
l
.
3
6
,
n
o
.
7
,
p
p
.
1
4
4
2
‒
1
4
6
8
,
2
0
1
3
,
d
o
i
:
1
0
.
1
1
0
9
/
T
P
A
M
I
.
2
0
1
3
.
5
0
.
[
9
]
M
.
F
i
a
z
,
A
.
M
a
h
mo
o
d
,
S
.
Ja
v
e
d
,
a
n
d
S
.
K
.
J
u
n
g
,
“
H
a
n
d
c
r
a
f
t
e
d
a
n
d
d
e
e
p
t
r
a
c
k
e
r
s:
r
e
c
e
n
t
v
i
s
u
a
l
o
b
j
e
c
t
t
r
a
c
k
i
n
g
a
p
p
r
o
a
c
h
e
s
a
n
d
t
r
e
n
d
s,
”
A
C
M
C
o
m
p
u
t
i
n
g
S
u
r
v
e
y
s
,
v
o
l
.
5
2
,
n
o
.
2
,
p
p
.
1
‒
4
4
,
2
0
1
9
,
d
o
i
:
1
0
.
1
1
4
5
/
3
3
0
9
6
6
5
.
[
1
0
]
B
.
L
i
u
,
L.
Y
a
n
g
,
J
.
H
u
a
n
g
,
P
.
M
e
e
r
,
L.
G
o
n
g
,
a
n
d
C
.
K
u
l
i
k
o
w
s
k
i
,
“
R
o
b
u
s
t
a
n
d
f
a
st
c
o
l
l
a
b
o
r
a
t
i
v
e
t
r
a
c
k
i
n
g
w
i
t
h
t
w
o
s
t
a
g
e
s
p
a
r
se
o
p
t
i
m
i
z
a
t
i
o
n
,
”
i
n
Eu
ro
p
e
a
n
C
o
n
f
.
C
o
m
p
u
t
e
r
Vi
s
i
o
n
(
E
C
C
V)
,
2
0
1
0
,
v
o
l
.
6
3
1
4
,
p
p
.
6
2
4
‒
637
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
6
4
2
-
1
5
5
6
1
-
1
_
4
4
.
[
1
1
]
Q
.
M
i
a
o
,
C
.
X
u
,
F
.
L
i
,
W
.
Z
u
o
,
a
n
d
Z.
M
e
n
g
,
“
D
e
l
a
y
e
d
r
e
c
t
i
f
i
c
a
t
i
o
n
o
f
d
i
scri
mi
n
a
t
i
v
e
c
o
r
r
e
l
a
t
i
o
n
f
i
l
t
e
r
s
f
o
r
v
i
s
u
a
l
t
r
a
c
k
i
n
g
,
”
T
h
e
Vi
su
a
l
C
o
m
p
u
t
e
r
,
v
o
l
.
3
9
,
n
o
.
6
,
p
p
.
1
2
3
7
‒
1
2
5
0
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
0
7
/
s0
0
3
7
1
-
0
2
2
-
0
2
7
5
0
-
5.
Evaluation Warning : The document was created with Spire.PDF for Python.