I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Ro
bo
t
ics a
nd
Aut
o
m
a
t
io
n
(
I
J
RA)
Vo
l.
13
,
No
.
4
,
Dec
em
b
er
20
24
,
p
p
.
4
8
5
~
4
9
4
I
SS
N:
2722
-
2
5
8
6
,
DOI
:
1
0
.
1
1
5
9
1
/i
jr
a
.
v
13
i
4
.
pp
4
8
5
-
494
485
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
r
a
.
ia
esco
r
e.
co
m
Desig
n and
imple
menta
tion o
f
deep
neural ne
twork
h
a
rdwa
re
chip and
i
ts per
forma
nce ana
ly
sis
Aruna
P
a
nt,
Adesh
K
um
a
r
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
i
c
a
l
a
n
d
El
e
c
t
r
o
n
i
c
s
En
g
i
n
e
e
r
i
n
g
,
S
c
h
o
o
l
o
f
A
d
v
a
n
c
e
d
En
g
i
n
e
e
r
i
n
g
,
U
P
ES
,
D
e
h
r
a
d
u
n
,
I
n
d
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Ma
r
1
3
,
2
0
2
4
R
ev
is
ed
J
u
n
9
,
2
0
2
4
Acc
ep
ted
J
u
l 1
7
,
2
0
2
4
Th
e
a
rti
ficia
l
n
e
u
ra
l
n
e
tw
o
rk
(
AN
N)
with
a
sin
g
le
lay
e
r
h
a
s
a
li
m
it
e
d
c
a
p
a
c
it
y
to
p
r
o
c
e
ss
d
a
ta.
M
u
lt
i
p
le
n
e
u
ro
n
s
a
re
c
o
n
n
e
c
ted
in
t
h
e
h
u
m
a
n
b
ra
in
,
a
n
d
th
e
a
c
tu
a
l
c
a
p
a
c
it
y
o
f
th
e
b
ra
in
li
e
s
in
th
e
in
terc
o
n
n
e
c
ted
n
e
ss
o
f
m
u
lt
ip
l
e
n
e
u
ro
n
s.
As
a
s
p
e
c
ifi
e
d
g
e
n
e
ra
li
z
a
ti
o
n
o
f
AN
N
d
e
e
p
lea
rn
i
n
g
m
a
k
e
s
u
se
o
f
two
o
r
m
o
re
h
i
d
d
e
n
lay
e
rs,
wh
ic
h
imp
li
e
s
t
h
a
t
a
g
re
a
ter
n
u
m
b
e
r
o
f
n
e
u
r
o
n
s
a
re
re
q
u
ired
t
o
c
o
n
stru
c
t
th
e
m
o
d
e
l.
A
n
e
two
r
k
t
h
a
t
h
a
s
m
o
re
th
a
n
o
n
e
h
id
d
e
n
lay
e
r,
a
ls
o
k
n
o
wn
a
s
two
o
r
m
o
re
h
id
d
e
n
lay
e
rs,
is
re
fe
rre
d
t
o
a
s
a
d
e
e
p
n
e
u
ra
l
n
e
two
r
k
,
a
n
d
th
e
p
r
o
c
e
ss
o
f
train
in
g
s
u
c
h
n
e
two
r
k
s
is
re
fe
rre
d
to
a
s
d
e
e
p
lea
rn
in
g
.
Th
e
re
se
a
rc
h
a
rti
c
le
fo
c
u
se
s
o
n
th
e
d
e
sig
n
o
f
a
m
u
lt
il
a
y
e
r
o
r
d
e
e
p
n
e
u
ra
l
n
e
tw
o
rk
p
re
se
n
t
e
d
fo
r
th
e
targ
e
t
field
p
r
o
g
ra
m
m
a
b
le
g
a
te
a
rra
y
(F
P
G
A)
d
e
v
ice
sp
a
rtan
-
6
(
x
c
6
stx
4
-
2
t
9
g
1
4
4
)
F
P
G
A.
Th
e
sim
u
latio
n
is
c
a
rried
o
u
t
u
si
n
g
Xi
li
n
x
IS
E
a
n
d
M
o
d
e
lS
im
so
ftwa
re
.
T
h
e
re
a
re
two
h
i
d
d
e
n
lay
e
rs
in
wh
ic
h
(
2
×
1
)
m
u
lt
i
p
lex
e
r
b
lo
c
k
s
a
re
u
ti
li
z
e
d
fo
r
p
ro
c
e
ss
in
g
twe
n
t
y
n
e
u
ro
n
s
in
t
o
th
e
o
u
tp
u
t
o
f
ten
n
e
u
ro
n
s
in
th
e
first
h
i
d
d
e
n
l
a
y
e
r
a
n
d
d
e
m
u
lt
ip
lex
e
rs
(1
×
2
)
a
n
d
v
ice
v
e
rsa
.
Th
e
h
a
r
d
wa
re
u
ti
li
z
a
ti
o
n
is
e
stim
a
ted
o
n
F
P
G
A
to
c
o
m
p
u
te
th
e
p
e
rfo
rm
a
n
c
e
o
f
th
e
d
e
e
p
n
e
u
ra
l
h
a
rd
wa
re
c
h
ip
b
a
se
d
o
n
m
e
m
o
r
y
,
fli
p
fl
o
p
s,
d
e
lay
,
a
n
d
fre
q
u
e
n
c
y
p
a
ra
m
e
ters
.
Th
e
d
e
sig
n
is
sc
a
lab
le
a
n
d
a
p
p
l
ica
b
le
t
o
v
a
rio
u
s
F
P
G
A
d
e
v
ice
s,
wh
ic
h
m
a
k
e
s
th
e
wo
r
k
n
o
v
e
l
.
F
P
G
A
-
b
a
se
d
n
e
u
r
o
m
o
rp
h
i
c
h
a
rd
wa
re
a
c
c
e
lera
ti
o
n
p
latfo
rm
with
h
ig
h
sp
e
e
d
a
n
d
l
o
w
p
o
we
r
fo
r
d
isc
re
te
sp
ik
e
p
r
o
c
e
ss
in
g
o
n
h
a
rd
wa
re
with
g
re
a
t
re
a
l
-
ti
m
e
p
e
rfo
rm
a
n
c
e
.
K
ey
w
o
r
d
s
:
C
o
n
tr
o
l lo
g
ic
Dee
p
lear
n
in
g
n
eu
r
al
n
etwo
r
k
Hid
d
en
lay
er
s
Per
f
o
r
m
an
ce
p
ar
am
eter
s
Sp
ar
tan
-
6
FP
GA
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
:
Ad
esh
Ku
m
ar
Dep
ar
tm
en
t o
f
E
lectr
ical
an
d
E
lectr
o
n
ics E
n
g
in
ee
r
i
n
g
,
Sch
o
o
l o
f
Ad
v
an
ce
d
E
n
g
in
ee
r
in
g
,
Un
iv
er
s
ity
o
f
Petr
o
leu
m
an
d
E
n
er
g
y
Stu
d
ies
(
UPES)
Deh
r
ad
u
n
,
I
n
d
ia
E
m
ail:
ad
esh
m
an
av
@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
Dee
p
n
eu
r
al
n
etwo
r
k
(
DNN
)
h
as
a
lo
t
o
f
p
o
p
u
lar
ity
d
u
e
to
wh
ic
h
is
u
s
ed
f
o
r
m
o
s
t
o
f
th
e
ap
p
licatio
n
s
.
L
ik
e
d
etec
tio
n
a
p
p
licatio
n
s
,
s
p
ec
if
ic
h
ar
d
war
e
d
esig
n
o
f
m
u
ltil
ay
er
n
eu
r
al
n
etwo
r
k
s
is
r
eq
u
ir
ed
wh
en
a
g
en
er
al
co
m
p
u
ter
is
n
o
t
s
u
f
f
icien
t
as
it
h
as
lim
i
tatio
n
s
in
s
p
ee
d
,
co
s
t
o
f
th
e
ch
ip
,
an
d
co
n
s
u
m
p
tio
n
o
f
en
er
g
y
.
A
d
ee
p
lear
n
in
g
m
o
d
el
is
u
s
ed
h
er
e
wh
ich
h
as
two
h
id
d
e
n
lay
e
r
s
.
T
h
e
m
u
ltil
ay
e
r
n
eu
r
al
n
etwo
r
k
is
m
ad
e
u
p
o
f
n
u
m
e
r
o
u
s
lay
e
r
s
b
etwe
en
th
e
in
p
u
t
an
d
o
u
tp
u
t
lay
er
s
,
with
all
h
id
d
en
lev
els
b
ein
g
Fu
lly
co
n
n
ec
ted
.
Dee
p
lear
n
in
g
n
e
u
r
al
n
etwo
r
k
im
p
lem
en
tatio
n
o
n
f
ield
p
r
o
g
r
am
m
ab
le
g
ate
ar
r
a
y
s
(
FP
GAs
)
p
r
esen
ts
v
ar
io
u
s
ch
allen
g
es,
in
clu
d
i
n
g
s
to
r
ag
e,
r
eso
u
r
ce
r
e
q
u
ir
em
e
n
ts
,
an
d
m
e
m
o
r
y
u
tili
za
tio
n
[
1
]
.
Dee
p
n
e
u
r
al
n
etwo
r
k
s
ar
e
m
a
d
e
u
p
o
f
n
u
m
er
o
u
s
“
d
ee
p
”
an
d
h
id
d
e
n
la
y
er
s
,
an
d
ea
c
h
o
n
e
o
f
t
h
em
d
o
es
a
p
ar
ticu
lar
task
.
Dee
p
lear
n
in
g
is
lear
n
in
g
a
b
o
u
t d
if
f
er
en
t le
v
els o
f
d
ata
r
ep
r
esen
tatio
n
s
an
d
th
eir
u
n
d
er
ly
i
n
g
d
is
tr
ib
u
tio
n
o
f
th
e
d
ata
[
2
]
.
DNNs
ar
e
m
o
s
t
u
s
e
d
in
v
i
d
eo
e
d
itin
g
ap
p
licatio
n
s
,
s
u
ch
as
o
b
ject
d
etec
tio
n
an
d
r
ea
l
-
tim
e
s
p
ee
ch
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
13
,
No
.
4
,
Dec
em
b
er
20
24
:
48
5
-
4
9
4
486
tr
an
s
latio
n
.
DNNs
will
b
ec
o
m
e
in
cr
ea
s
in
g
ly
s
ig
n
if
ican
t
i
n
m
ed
icin
e
,
r
o
b
o
tics
,
a
n
d
f
i
n
an
ce
f
o
r
wea
th
er
f
o
r
ec
asti
n
g
an
d
d
etec
tio
n
[
3
]
.
DNNs
ar
e
n
o
w
th
e
b
asis
f
o
r
m
an
y
m
o
d
er
n
ar
tific
ial
in
tell
ig
en
ce
ap
p
licatio
n
s
[
4
]
.
Sin
ce
th
e
g
r
o
u
n
d
b
r
ea
k
in
g
u
s
e
o
f
DNNs
in
s
p
ee
ch
r
ec
o
g
n
itio
n
an
d
im
ag
e
r
ec
o
g
n
itio
n
,
th
e
n
u
m
b
e
r
o
f
p
lace
s
wh
er
e
DNNs
h
a
v
e
s
k
y
r
o
ck
eted
.
T
h
ese
d
ee
p
n
eu
r
al
n
e
two
r
k
s
ar
e
u
s
ed
in
a
wid
e
r
an
g
e
o
f
a
p
p
licatio
n
s
,
in
clu
d
in
g
s
elf
-
d
r
i
v
in
g
ca
r
s
,
ca
n
ce
r
d
etec
tio
n
,
an
d
co
m
p
licat
ed
g
am
ep
lay
[
5
]
.
I
n
m
o
s
t
f
ield
s
,
m
u
ltil
ay
er
n
eu
r
al
n
etwo
r
k
s
ca
n
b
e
d
o
n
e
m
u
ch
m
o
r
e
ac
cu
r
ately
th
a
n
h
u
m
a
n
s
.
Ov
er
all,
m
u
ltil
ay
er
ar
tific
ial
n
eu
r
al
n
etwo
r
k
s
(
ANNs
)
,
p
ar
ticu
lar
ly
d
ee
p
n
et
wo
r
k
s
,
p
r
o
v
ed
t
h
em
s
elv
es
b
etter
at
h
an
d
lin
g
co
m
p
licated
ta
s
k
s
s
u
ch
as
im
ag
e
id
en
tific
atio
n
,
s
p
ee
c
h
r
ec
o
g
n
it
io
n
,
lan
g
u
ag
e
tr
an
s
latio
n
,
an
d
o
th
er
s
,
m
ak
in
g
th
e
m
an
ess
en
tial
co
m
p
o
n
en
t
o
f
m
o
d
er
n
ar
tific
ial
in
tellig
en
ce
s
y
s
tem
s
[
6
]
.
DNNs
ar
e
u
s
ed
in
a
wid
e
v
ar
iety
o
f
ar
t
if
icial
in
tellig
en
ce
ap
p
licatio
n
s
,
in
clu
d
in
g
co
m
p
u
ter
v
is
io
n
,
s
p
ee
ch
r
ec
o
g
n
itio
n
,
r
o
b
o
tics
,
an
d
m
a
n
y
m
o
r
e
.
DNNs
ca
n
ac
h
iev
e
cu
ttin
g
-
ed
g
e
ac
cu
r
ac
y
o
n
a
wi
d
e
v
ar
iety
o
f
a
r
tific
ial
in
tellig
en
ce
task
s
[
7
]
.
T
h
e
c
o
m
p
u
tati
o
n
al
co
m
p
lex
ity
o
f
th
ese
n
etwo
r
k
s
is
r
ath
er
h
ig
h
.
B
etter
p
e
r
f
o
r
m
an
ce
,
h
ig
h
e
r
b
a
n
d
wid
th
,
an
d
d
eter
m
in
is
tic
lo
w
laten
cy
ar
e
ac
h
iev
ed
with
th
e
u
s
ag
e
o
f
FP
GA
an
d
HDL
in
th
e
d
esig
n
o
f
p
r
o
g
r
a
m
m
ab
le
p
r
o
ce
s
s
o
r
s
.
Par
allel
p
r
o
ce
s
s
in
g
an
d
o
p
er
atio
n
s
[
8
]
ar
e
th
e
f
o
u
n
d
atio
n
o
f
t
h
e
ar
c
h
itectu
r
es
o
f
ANN
to
p
r
o
ce
s
s
n
eu
r
o
n
al
in
f
o
r
m
atio
n
,
r
e
g
u
lar
ch
ip
s
ar
e
u
n
ab
le
to
s
u
s
tain
a
h
u
g
e
n
u
m
b
er
o
f
h
ig
h
ly
s
im
u
ltan
eo
u
s
p
r
o
ce
s
s
es
[
9
]
.
Sev
er
al
d
is
tin
ct
p
r
o
ce
s
s
o
r
s
th
at
en
ab
le
p
ar
allel
p
r
o
ce
s
s
in
g
ar
e
in
clu
d
e
d
in
th
e
a
r
tific
ial
in
tellig
en
ce
-
b
ased
h
a
r
d
war
e
c
h
ip
.
T
h
e
p
r
o
b
lem
s
tatem
en
t
o
f
th
e
wo
r
k
is
to
d
esig
n
th
e
d
ee
p
n
e
u
r
al
n
etwo
r
k
h
a
r
d
war
e
c
h
ip
a
n
d
esti
m
ate
its
p
er
f
o
r
m
an
ce
o
n
FP
GA.
T
h
e
ch
ip
is
d
esig
n
ed
f
o
r
2
0
n
eu
r
o
n
in
p
u
t
an
d
ea
ch
n
e
u
r
o
n
in
p
u
t
is
o
f
8
b
its
.
E
ac
h
20
-
n
e
u
r
o
n
in
p
u
t is m
u
ltip
lied
b
y
2
0
i
n
p
u
t w
eig
h
ts
an
d
ea
ch
weig
h
t is 8
-
b
it so
wh
en
th
ese
2
0
in
p
u
t w
eig
h
ts
ar
e
m
u
ltip
lied
b
y
2
0
n
e
u
r
o
n
in
p
u
t
s
in
th
e
m
u
ltip
lier
it
g
iv
es
a
1
6
-
b
it
o
u
tp
u
t.
T
h
e
m
u
ltip
le
d
e
m
u
ltip
lex
er
(
1
×2
)
is
u
s
ed
wh
ich
a
g
ain
c
o
n
v
e
r
ts
1
0
-
n
eu
r
o
n
o
u
tp
u
t
i
n
to
2
0
-
n
e
u
r
o
n
o
u
tp
u
t
w
h
ich
is
g
i
v
en
to
th
e
s
e
co
n
d
h
id
d
e
n
lay
er
.
A
co
n
tr
o
l
lo
g
ic
is
u
s
ed
i
n
th
is
n
eu
r
o
m
o
r
p
h
ic
h
ar
d
war
e
ch
i
p
d
esig
n
wh
ich
is
u
s
ed
to
f
ee
d
m
u
ltip
lier
o
u
tp
u
t
t
o
ea
ch
in
p
u
t
o
f
th
e
h
id
d
en
lay
er
.
T
h
e
h
id
d
en
lay
er
o
u
tp
u
t
is
th
en
g
iv
en
to
th
e
m
u
ltip
lex
e
r
wh
ich
h
as
2
0
in
p
u
ts
an
d
o
n
e
s
in
g
le
o
u
tp
u
t.
T
h
e
m
u
x
h
as
a
5
-
b
it
s
elec
tio
n
lin
e
wh
ich
ca
n
b
e
u
s
ed
to
s
elec
t
an
y
o
f
th
e
2
0
o
u
tp
u
ts
o
f
th
e
co
n
tr
o
l
lo
g
ic.
Har
d
war
e
u
tili
za
tio
n
ch
ar
ac
ter
is
tics
s
u
ch
as
s
lice
s
u
s
ed
,
lo
o
k
-
u
p
tab
les
(
L
UT
s
)
,
f
lip
-
f
lo
p
s
(
FF
s
)
,
in
p
u
t/o
u
tp
u
t
b
lo
ck
s
(
I
OB
s
)
,
m
em
o
r
y
,
an
d
to
tal
p
r
o
p
ag
atio
n
laten
cy
ar
e
u
s
ed
to
ca
lcu
late
th
e
d
ee
p
n
eu
r
al
h
a
r
d
war
e
ch
i
p
’
s
p
er
f
o
r
m
an
ce
.
2.
RE
L
AT
E
D
WO
RK
Alth
o
u
g
h
s
o
f
twar
e
ca
n
ea
s
ily
cr
ea
te
m
u
ltil
ay
er
n
eu
r
al
n
etwo
r
k
m
o
d
els
o
n
co
m
p
u
ter
s
,
th
er
e
ar
e
n
u
m
er
o
u
s
ch
allen
g
es
wh
en
d
esig
n
in
g
d
ee
p
n
eu
r
al
n
e
two
r
k
s
o
n
s
ilico
n
.
T
h
ese
ch
allen
g
es
in
clu
d
e
im
p
lem
en
tin
g
th
e
m
u
ltip
licatio
n
o
f
in
p
u
t
with
weig
h
ts
at
e
ac
h
lay
er
,
d
esig
n
i
n
g
t
h
e
s
to
r
i
n
g
n
etwo
r
k
weig
h
ts
,
an
d
ac
tiv
atio
n
f
u
n
ctio
n
,
an
d
s
e
lectin
g
an
ap
p
r
o
p
r
iate
n
u
m
b
er
f
o
r
m
at
f
o
r
ca
lc
u
latio
n
s
.
T
o
ac
h
iev
e
p
e
r
f
o
r
m
an
ce
ef
f
icien
cy
,
ea
c
h
lay
e
r
in
a
d
ee
p
ANN
ar
ch
itectu
r
e,
wh
ich
will
b
e
th
e
f
o
cu
s
o
f
th
is
wo
r
k
,
n
ec
ess
itate
s
its
o
wn
in
d
ep
en
d
en
t
s
p
ac
e
f
o
r
m
em
o
r
y
an
d
its
weig
h
t.
I
n
a
d
ee
p
n
etwo
r
k
th
at
h
as
two
o
r
m
o
r
e
h
id
d
en
lay
er
s
,
th
e
m
em
o
r
y
r
eq
u
ir
e
m
en
t
f
o
r
th
e
weig
h
t
m
atr
ices
o
f
th
e
e
n
tire
n
etwo
r
k
in
cr
e
ases
at
th
e
f
astes
t
as
its
n
etwo
r
k
’
s
s
ize
an
d
m
em
o
r
y
ac
ce
s
s
es
g
r
o
w.
Fu
r
th
er
m
o
r
e,
c
o
m
p
u
tatio
n
s
with
in
ea
ch
lay
er
s
h
o
u
ld
b
e
e
x
ec
u
ted
in
p
a
r
allel
f
o
r
ap
p
licatio
n
s
r
e
q
u
ir
in
g
h
ig
h
s
p
ee
d
.
T
h
ese
ch
ar
ac
ter
is
tics
h
av
e
a
s
u
b
s
tan
tial
in
f
lu
en
ce
o
n
th
e
ef
f
icien
cy
an
d
p
er
f
o
r
m
an
ce
o
f
DNNs.
I
n
th
is
co
n
tex
t,
th
e
FP
GA
is
v
ia
b
le
f
o
r
im
p
le
m
en
tin
g
a
DN
N
ac
ce
ler
ato
r
.
T
h
e
d
esig
n
in
g
o
f
d
ee
p
ANN
an
d
th
e
u
s
e
o
f
FP
GA
is
v
er
y
p
r
e
v
alen
t
in
r
esear
ch
a
r
ea
s
[
1
0
]
,
[
1
1
]
.
c
o
n
v
o
lu
tio
n
n
eu
r
al
n
etwo
r
k
s
(
C
NNs)
h
av
e
b
ee
n
u
s
ed
f
o
r
r
ea
l
-
tim
e
em
b
ed
d
e
d
s
y
s
tem
s
,
ex
p
lo
r
in
g
th
e
im
p
ar
tial
class
if
icatio
n
o
f
cu
ttin
g
-
ed
g
e
s
y
s
tem
-
on
-
c
h
ip
s
[
1
2
]
.
M
u
lti
-
d
im
en
s
io
n
al
tr
a
d
e
-
o
f
f
s
a
r
e
in
v
esti
g
ated
to
co
n
f
ig
u
r
e
p
r
o
g
r
am
m
ab
le
ac
c
eler
ato
r
s
f
o
r
n
eu
r
al
in
ter
p
r
et
atio
n
u
s
in
g
th
e
f
in
est
s
o
f
twar
e
lib
r
ar
ies.
T
h
e
f
r
am
ewo
r
k
is
o
p
en
s
o
u
r
ce
to
p
r
o
m
o
te
f
ai
r
b
e
n
ch
m
ar
k
in
g
ass
ess
m
en
ts
.
I
t
h
as
b
ec
o
m
e
in
c
r
e
asin
g
ly
co
m
m
o
n
to
d
esig
n
m
u
ltil
ay
er
p
er
ce
p
tr
o
n
n
eu
r
al
n
etwo
r
k
s
o
n
r
ec
o
n
f
i
g
u
r
ab
le
h
ar
d
war
e
to
ca
r
r
y
o
u
t
a
wid
e
r
a
n
g
e
o
f
ap
p
licatio
n
s
,
in
clu
d
in
g
p
atter
n
r
ec
o
g
n
itio
n
an
d
class
if
icatio
n
.
T
h
e
p
u
r
p
o
s
e
o
f
th
e
r
esear
ch
b
y
Hu
y
n
h
[
1
3
]
was
to
in
v
esti
g
ate
th
e
co
m
b
in
ed
im
p
ac
t
o
f
n
eu
r
al
n
etwo
r
k
s
ize
an
d
d
ec
r
ea
s
ed
p
r
ec
is
io
n
n
u
m
b
er
f
o
r
m
ats,
wh
ich
ar
e
u
tili
ze
d
f
o
r
th
e
r
e
p
r
esen
t
atio
n
o
f
t
h
e
id
ea
l
p
a
r
am
eter
s
,
o
n
th
e
r
ec
o
g
n
itio
n
r
ate
o
f
a
h
an
d
wr
itten
d
ig
it
r
ec
o
g
n
itio
n
s
y
s
tem
th
at
is
b
ased
o
n
n
e
u
r
al
n
etwo
r
k
s
.
ANN
a
r
ch
itectu
r
e
[
1
4
]
was
ev
alu
ate
d
o
n
Vir
tex
-
5
FP
GA
f
o
r
p
atter
n
r
ec
o
g
n
itio
n
an
d
te
s
ted
f
o
r
s
ca
lab
ilit
y
an
d
h
a
r
d
w
ar
e
r
eso
u
r
ce
u
s
e.
Dee
p
lea
r
n
i
n
g
is
n
o
w
t
h
e
m
o
s
t
ad
v
an
ce
d
s
o
lu
tio
n
f
o
r
alm
o
s
t
all
s
ig
n
if
ican
t
m
ac
h
in
e
-
lear
n
in
g
task
s
in
a
wid
e
r
an
g
e
o
f
d
o
m
ain
s
[
1
5
]
,
[
1
6
]
.
Dee
p
lear
n
in
g
tech
n
i
q
u
es
o
u
tp
er
f
o
r
m
ty
p
ical
m
ac
h
in
e
l
ea
r
n
in
g
al
g
o
r
ith
m
s
t
h
at
r
eq
u
ir
e
h
u
m
a
n
f
ea
tu
r
e
ex
tr
ac
tio
n
(
h
an
d
cr
af
ted
f
ea
tu
r
es)
[
1
7
]
.
Dee
p
lear
n
in
g
m
o
d
els
ex
tr
ac
t
h
ie
r
ar
ch
ical
f
ea
tu
r
es
an
d
p
er
f
o
r
m
b
etter
as th
e
am
o
u
n
t o
f
d
ata
in
cr
e
ases
[
1
8
]
.
T
h
er
e
ar
e
v
ar
io
u
s
d
ee
p
lear
n
in
g
m
eth
o
d
s
an
d
ar
ch
itec
tu
r
es a
p
p
lied
in
th
e
d
esig
n
,
s
u
ch
as
m
u
lti
-
lay
er
p
er
ce
p
tr
o
n
(
ML
P),
d
ee
p
b
eli
ef
n
etwo
r
k
(
DB
N)
,
au
to
en
c
o
d
er
(
AE
)
,
C
NNs,
r
ec
u
r
r
en
t
n
e
u
r
al
n
etwo
r
k
(
R
NN)
with
lo
n
g
s
h
o
r
t
-
ter
m
m
em
o
r
y
(
L
STM
)
an
d
g
ated
r
e
cu
r
r
en
t
u
n
its
(
GR
U)
,
g
en
er
ativ
e
ad
v
er
s
ar
ial
n
etwo
r
k
(
GAN)
,
d
ee
p
r
ein
f
o
r
ce
m
en
t
lear
n
in
g
(
DR
L
)
[
1
9
]
.
T
h
ese
m
o
d
els
h
av
e
s
p
an
n
e
d
a
wid
e
r
an
g
e
o
f
f
ield
s
an
d
ap
p
licatio
n
s
.
Dee
p
lear
n
in
g
ar
c
h
itectu
r
es
h
av
e
en
ab
led
a
u
to
n
o
m
o
u
s
d
r
iv
in
g
(
AD)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
Desig
n
a
n
d
imp
leme
n
t
a
tio
n
o
f d
ee
p
n
e
u
r
a
l n
etw
o
r
k
h
a
r
d
w
a
r
e
ch
ip
a
n
d
its
p
erfo
r
ma
n
ce
...
(
A
r
u
n
a
P
a
n
t
)
487
s
y
s
tem
s
[
2
0
]
to
ac
q
u
ir
e
a
h
ig
h
lev
el
o
f
p
r
ec
is
io
n
in
th
ei
r
p
er
f
o
r
m
an
ce
;
th
ese
s
y
s
tem
s
ar
e
co
m
p
r
is
ed
o
f
n
u
m
er
o
u
s
p
er
ce
p
tio
n
-
lev
el
ta
s
k
s
.
T
r
ad
itio
n
al
s
u
p
e
r
v
is
ed
le
ar
n
in
g
m
eth
o
d
s
ar
e
n
o
lo
n
g
e
r
ap
p
licab
le
to
th
e
m
an
y
task
s
th
at
m
ak
e
u
p
au
to
n
o
m
o
u
s
d
r
iv
i
n
g
s
y
s
tem
s
,
ex
ce
p
t
f
o
r
p
er
ce
p
tio
n
.
I
n
co
m
p
u
ter
v
is
io
n
ap
p
licatio
n
s
,
C
NN
m
o
d
els h
a
v
e
d
em
o
n
s
tr
ated
e
n
co
u
r
ag
in
g
o
u
tco
m
es
in
ar
ea
s
in
clu
d
in
g
r
o
b
o
tics
,
au
to
n
o
m
o
u
s
v
eh
icle
v
is
io
n
s
y
s
tem
s
,
an
d
d
r
o
n
e
n
av
ig
ati
o
n
[
2
1
]
.
I
n
h
e
alth
ca
r
e,
C
NNs
h
av
e
also
o
u
tp
er
f
o
r
m
ed
o
th
e
r
m
eth
o
d
s
,
esp
ec
ially
f
o
r
im
ag
e
id
en
tific
atio
n
.
I
t
h
a
s
b
e
en
u
s
ed
to
d
e
te
c
t
t
u
m
o
r
s
o
r
o
th
e
r
t
y
p
e
s
o
f
le
s
i
o
n
s
b
y
m
o
r
e
e
x
p
e
r
ie
n
c
ed
r
ad
i
o
lo
g
i
s
t
s
[
2
2
]
.
B
r
a
i
n
i
m
ag
e
s
h
a
v
e
b
e
en
u
s
ed
f
o
r
h
u
m
a
n
s
f
o
r
m
ag
n
e
t
i
c
r
e
s
o
n
an
c
e
i
m
ag
i
n
g
(
M
R
I
)
t
h
a
t
d
e
t
e
c
t
s
d
i
s
e
a
s
e
A
l
z
h
e
i
m
er
’
s
w
i
t
h
th
e
h
e
l
p
o
f
C
N
N
[
2
3
]
.
D
e
e
p
l
e
a
r
n
i
n
g
m
e
th
o
d
o
lo
g
i
e
s
[
2
4
]
h
av
e
p
r
o
v
ed
e
f
f
e
c
t
i
v
e
i
n
a
d
d
r
e
s
s
i
n
g
ar
t
i
f
i
c
ia
l
i
n
t
e
l
l
ig
en
c
e
d
if
f
i
cu
l
t
i
e
s
,
wh
i
l
e
F
P
G
A
s
h
a
v
e
b
e
e
n
u
s
e
d
to
i
m
p
r
o
v
e
r
e
c
o
n
f
ig
u
r
a
b
le
c
o
m
p
u
t
er
h
ar
d
w
ar
e
a
n
d
s
o
f
t
w
a
r
e
f
o
r
a
r
ti
f
i
c
i
a
l
in
t
e
l
l
i
g
en
c
e
d
e
s
ig
n
.
Mu
l
t
i
l
ay
e
r
N
e
tw
o
r
k
s
h
a
v
e
b
e
en
d
e
s
i
g
n
e
d
u
s
in
g
F
P
G
A
[
2
5
]
.
Gr
a
p
h
i
c
p
r
o
c
e
s
s
i
n
g
u
n
i
t
s
(
G
P
Us
)
a
r
e
u
n
s
u
i
t
a
b
l
e
f
o
r
em
b
ed
d
ed
ap
p
l
ic
a
t
i
o
n
s
d
u
e
t
o
t
h
e
ir
h
ig
h
-
p
o
we
r
co
n
s
u
m
p
t
i
o
n
.
M
an
y
a
ca
d
em
i
c
s
h
av
e
e
x
p
lo
r
ed
t
e
ch
n
iq
u
e
s
to
c
r
e
a
t
e
d
e
ep
n
eu
r
a
l
n
e
t
w
o
r
k
s
o
n
lo
w
-
p
o
w
er
h
a
r
d
w
a
r
e
[
2
6
]
.
T
h
e
m
o
s
t
ex
p
en
s
i
v
e
o
p
er
a
t
i
o
n
s
in
d
ee
p
c
o
n
v
o
lu
t
io
n
a
l
n
eu
r
a
l
n
e
t
w
o
r
k
s
(
D
C
N
N
s
)
ar
e
g
e
n
e
r
a
t
e
d
b
y
d
e
ep
C
N
N
s
,
w
h
ic
h
a
r
e
m
ad
e
u
p
o
f
n
u
m
e
r
o
u
s
l
a
y
er
s
o
f
th
r
e
e
-
d
i
m
en
s
i
o
n
a
l
co
n
v
o
lu
t
i
o
n
s
,
ea
c
h
o
f
wh
i
c
h
h
a
s
b
e
t
w
e
en
te
n
s
an
d
h
u
n
d
r
e
d
s
o
f
f
i
l
t
e
r
s
.
T
h
e
r
e
s
t
r
i
c
t
io
n
o
f
f
l
o
w
i
n
g
d
a
ta
i
n
an
d
o
u
t
o
f
t
h
e
h
a
r
d
w
a
r
e
ch
ip
a
c
c
e
l
er
a
t
o
r
,
w
h
i
c
h
i
s
s
y
n
th
e
s
i
z
ed
o
n
X
i
l
in
x
Z
y
n
q
-
7
0
0
0
,
i
s
a
c
h
a
l
l
en
g
e
f
o
r
s
y
s
t
e
m
s
t
h
a
t
r
u
n
D
C
N
N
s
.
T
h
e
s
e
s
y
s
t
e
m
s
ar
e
r
e
q
u
i
r
ed
t
o
p
a
s
s
3
D
i
n
p
u
t
m
a
p
s
t
o
t
h
e
h
ar
d
w
ar
e
a
c
c
e
l
e
r
a
to
r
s
t
o
p
e
r
f
o
r
m
c
o
n
v
o
l
u
t
i
o
n
s
[
2
7
]
.
A
n
i
m
p
l
e
m
e
n
t
a
t
io
n
o
f
H
ig
h
-
L
ev
e
l
S
y
n
th
e
s
i
s
w
a
s
e
m
p
l
o
y
e
d
o
n
a
V
i
r
t
e
x
-
6
F
P
G
A
b
o
a
r
d
,
an
d
C
N
N
w
a
s
u
t
i
l
i
z
e
d
[
2
8
]
to
a
n
a
ly
z
e
t
h
e
d
e
s
i
g
n
f
lo
w
.
T
h
e
F
PG
A
r
e
s
o
u
r
c
e
s
c
an
b
e
d
e
c
r
ea
s
e
d
b
y
u
p
to
1
3
t
i
m
es
w
h
e
n
co
m
p
a
r
ed
to
a
cc
e
l
e
r
a
t
o
r
s
th
a
t
u
s
e
tr
ad
i
t
i
o
n
a
l
s
cr
a
t
c
h
p
a
d
m
e
m
o
r
i
e
s
,
w
h
i
l
e
y
e
t
m
a
i
n
ta
i
n
in
g
t
h
e
s
am
e
l
ev
e
l
o
f
p
e
r
f
o
r
m
a
n
c
e
.
A
f
i
x
e
d
p
o
i
n
t
D
N
N
i
s
d
e
s
i
g
n
ed
o
n
F
P
G
A
a
n
d
u
s
e
s
a
n
y
ex
t
e
r
n
a
l
m
em
o
r
y
.
I
t
s
p
o
w
e
r
c
o
n
s
u
m
p
t
io
n
an
d
t
i
m
e
o
f
e
x
e
cu
t
i
o
n
a
r
e
co
m
p
a
r
e
d
w
i
t
h
G
P
U
-
b
a
s
e
d
i
m
p
l
e
m
e
n
t
a
t
io
n
[
2
9
]
.
T
o
c
o
n
s
t
r
u
c
t
f
e
e
d
-
f
o
r
w
ar
d
d
e
e
p
n
eu
r
a
l
n
e
t
wo
r
k
s
i
n
r
e
a
l
-
t
im
e
,
GP
U
s
a
r
e
c
o
m
m
o
n
l
y
u
s
e
d
d
u
e
t
o
t
h
e
h
i
g
h
a
m
o
u
n
t
o
f
a
r
i
t
h
m
e
t
i
c
a
n
d
m
e
m
o
r
y
o
p
e
r
a
t
i
o
n
s
r
e
q
u
i
r
ed
[
3
0
]
.
T
h
e
im
p
l
e
m
e
n
t
a
t
io
n
o
f
an
o
p
t
i
m
iz
e
d
d
e
ep
l
ea
r
n
in
g
a
r
c
h
i
t
e
c
t
u
r
e
w
a
s
c
ar
r
i
ed
o
u
t
o
n
a
p
a
r
a
l
l
e
l
co
m
p
u
t
in
g
p
l
a
tf
o
r
m
.
T
h
i
s
a
r
ch
i
t
e
c
tu
r
e
f
e
a
t
u
r
e
d
f
l
e
x
ib
l
e
l
ay
e
r
t
o
p
o
lo
g
i
e
s
a
n
d
f
a
s
t
m
a
t
r
ix
o
p
e
r
a
ti
o
n
k
e
r
n
e
l
s
.
T
h
i
s
i
s
t
h
e
d
e
s
i
g
n
to
s
u
p
p
o
r
t
m
in
i
m
u
m
h
a
r
d
w
a
r
e
u
t
i
l
i
z
a
t
io
n
[
3
1
]
.
T
h
e
i
n
t
eg
r
a
t
i
o
n
o
f
s
e
v
er
a
l
t
y
p
e
s
o
f
d
e
ep
a
r
c
h
i
t
ec
t
u
r
e
s
i
n
t
o
a
u
n
i
f
i
ed
n
e
u
r
a
l
t
r
a
i
n
in
g
-
t
e
s
t
i
n
g
s
y
s
t
e
m
i
s
a
cc
o
m
p
l
i
s
h
e
d
b
y
th
e
im
p
l
em
e
n
t
a
t
i
o
n
o
f
l
ay
e
r
-
w
i
s
e
p
r
o
ce
d
u
r
e
s
t
h
a
t
h
av
e
b
e
en
m
e
t
i
c
u
lo
u
s
l
y
d
e
s
i
g
n
e
d
.
A
n
eu
r
a
l
n
e
t
wo
r
k
h
a
r
d
w
a
r
e
m
o
d
el
c
a
n
b
e
d
e
s
i
g
n
e
d
o
n
F
P
G
A
.
T
h
i
s
n
eu
r
a
l
m
o
d
e
l
c
o
n
t
a
in
s
m
o
r
e
t
y
p
e
s
o
f
la
y
er
s
,
c
o
n
n
e
c
t
io
n
s
,
an
d
d
a
t
a
[
3
2
]
.
R
e
t
r
a
i
n
in
g
v
i
a
b
ac
k
p
r
o
p
ag
a
ti
o
n
wa
s
c
r
ea
t
e
d
a
s
a
n
a
l
t
e
r
n
a
t
iv
e
t
o
d
i
r
ec
t
w
e
i
g
h
t
q
u
a
n
t
i
za
t
i
o
n
[
3
3
]
.
F
e
e
d
f
o
r
w
a
r
d
D
N
N
s
w
i
t
h
s
ev
e
r
a
l
h
i
d
d
e
n
l
ay
e
r
s
f
u
n
c
t
i
o
n
w
e
l
l
i
n
m
a
n
y
r
eq
u
e
s
t
s
,
b
u
t
t
h
e
y
r
e
q
u
i
r
e
co
m
p
le
x
h
a
r
d
w
a
r
e
.
T
h
e
s
y
s
t
e
m
h
a
r
d
wa
r
e
co
m
p
le
x
i
ty
c
a
n
b
e
m
i
n
i
m
i
ze
d
b
y
r
e
d
u
c
i
n
g
t
h
e
we
i
g
h
t
s
an
d
s
i
g
n
a
l
wo
r
d
le
n
g
t
h
.
Q
u
a
n
t
i
z
e
d
we
i
g
h
t
s
a
n
d
f
i
x
ed
-
p
o
i
n
t
s
ig
n
a
l
s
ar
e
u
s
e
d
in
b
a
c
k
p
r
o
p
a
g
a
t
io
n
f
o
r
r
e
t
r
a
in
i
n
g
an
d
o
u
t
p
u
t
co
m
p
u
ta
t
i
o
n
.
H
o
w
ev
er
,
p
r
ec
i
s
i
o
n
n
u
m
b
e
r
s
a
r
e
u
s
ed
t
o
c
h
a
n
g
e
n
e
t
w
o
r
k
s
[
3
4
]
.
T
h
e
d
e
v
e
l
o
p
ed
n
e
t
wo
r
k
s
t
y
p
i
c
a
l
ly
u
s
e
d
2
t
o
8
b
i
t
s
f
o
r
w
e
i
g
h
t
s
an
d
m
o
r
e
t
h
a
n
7
b
i
t
s
t
o
r
e
p
r
e
s
e
n
t
s
i
g
n
a
l
s
i
n
a
n
a
lo
g
o
r
h
i
g
h
-
p
r
e
ci
s
i
o
n
f
ix
e
d
p
o
i
n
t
s
[
3
5
]
.
L
i
m
i
t
ed
wo
r
k
h
a
s
b
e
en
d
o
n
e
in
th
e
d
o
m
a
i
n
o
f
h
a
r
d
w
a
r
e
c
h
ip
d
e
s
i
g
n
an
d
s
y
n
th
e
s
i
s
o
n
th
e
s
p
ec
i
f
i
c
h
a
r
d
wa
r
e
s
p
e
c
i
f
i
ca
l
l
y
t
ar
g
e
t
in
g
an
y
F
P
G
A
h
a
r
d
w
a
r
e
a
n
d
d
ee
p
n
eu
r
a
l
n
et
w
o
r
k
p
r
o
c
e
s
s
i
n
g
.
T
h
e
p
r
o
b
l
em
s
t
a
t
e
m
en
t
o
f
t
h
e
p
r
e
s
e
n
t
ed
w
o
r
k
i
s
t
o
d
e
s
i
g
n
t
h
e
s
c
a
l
ab
l
e
h
ar
d
w
ar
e
c
h
ip
o
f
a
d
e
ep
n
eu
r
a
l
n
e
t
wo
r
k
an
d
e
v
a
lu
a
t
e
t
h
e
s
y
s
t
e
m
p
er
f
o
r
m
a
n
c
e
i
n
t
e
r
m
s
o
f
d
e
l
a
y
,
f
r
eq
u
en
c
y
,
L
UT
s
,
an
d
m
e
m
o
r
y
.
3.
M
E
T
H
O
DS
T
h
e
s
y
s
tem
d
esig
n
in
clu
d
es
a
n
in
p
u
t
lay
er
,
two
h
id
d
e
n
lay
e
r
s
,
an
d
o
n
ly
o
n
e
o
u
tp
u
t
lay
er
.
T
h
e
d
esig
n
o
f
th
e
d
ee
p
n
e
u
r
al
h
ar
d
war
e
ch
ip
co
n
s
is
ts
o
f
h
id
d
en
lay
er
s
wh
ich
ar
e
m
o
r
e
th
an
o
n
e.
T
h
e
two
h
id
d
en
lay
er
s
ar
e
u
s
ed
,
an
d
m
u
ltip
lier
s
ar
e
u
s
ed
f
o
r
m
u
ltip
ly
in
g
th
e
n
e
u
r
o
n
in
p
u
ts
with
th
e
n
eu
r
o
n
’
s
we
ig
h
t,
ap
ar
t
f
r
o
m
th
e
(2
×
1
)
m
u
ltip
lex
er
c
o
m
p
o
n
en
t,
th
e
d
em
u
ltip
le
x
er
(
1
×
2
)
c
o
m
p
o
n
en
t
a
n
d
o
n
e
(
8
×
1
)
m
u
ltip
l
ex
er
co
m
p
o
n
e
n
t
ar
e
u
s
ed
.
Fig
u
r
e
1
p
r
esen
ts
th
e
b
asic
b
u
ild
in
g
b
lo
ck
s
f
o
r
th
e
m
u
ltil
ay
er
Feed
-
f
o
r
war
d
ar
tifi
cial
n
eu
r
al
n
etwo
r
k
(
MFFANN)
.
F
ig
u
r
e
2
r
ep
r
es
en
ts
th
e
d
esig
n
o
f
a
m
u
ltil
ay
er
n
eu
r
al
n
etwo
r
k
.
I
t
co
n
s
is
ts
o
f
o
n
e
in
p
u
t,
two
h
id
d
en
la
y
er
s
,
an
d
o
n
e
o
u
tp
u
t
lay
er
.
T
h
e
d
esig
n
o
f
th
is
m
u
ltil
ay
er
ANN
is
d
o
n
e
in
VHSI
C
Har
d
war
e
Descr
ip
tio
n
L
an
g
u
ag
e
(
VHDL
)
u
s
in
g
a
n
FP
GA
Sp
ar
tr
an
-
6
d
ev
ice.
T
h
e
m
ain
p
ar
ts
o
f
th
e
s
y
s
tem
d
esig
n
ar
e
as
f
o
llo
ws
.
−
Mu
ltip
lier
: T
h
e
in
p
u
t la
y
er
is
im
p
lem
en
ted
u
s
in
g
a
m
u
ltip
lier
th
at
m
u
ltip
lies
(
0
-
1
9
)
n
eu
r
o
n
in
p
u
t w
ith
2
0
n
eu
r
o
n
weig
h
ts
an
d
ea
c
h
in
p
u
t
co
n
s
is
ts
o
f
8
-
b
it
n
eu
r
o
n
i
n
p
u
t
an
d
g
iv
es
a
s
in
g
le
1
6
-
b
it
o
u
tp
u
t
af
te
r
m
u
ltip
licatio
n
.
X
0
,
X
1
……….
X
19
as
th
e
in
p
u
ts
o
f
n
eu
r
o
n
s
,
an
d
W
0
,
W
1
…….
W
19
as
th
e
weig
h
ts
b
y
m
u
ltip
ly
in
g
th
is
in
th
e
m
u
ltip
li
er
we
g
et
th
e
o
u
tp
u
t Y
0
, Y
1
…
….
Y
19
,
wh
ich
is
1
6
-
b
it.
−
C
o
n
tr
o
l
L
o
g
ic:
T
h
e
co
n
tr
o
l
lo
g
ic
is
u
s
ed
f
o
r
th
e
h
i
d
d
en
la
y
e
r
.
T
h
e
o
u
tp
u
t
o
f
t
h
e
m
u
ltip
lier
,
wh
ich
is
Y
0
,
Y
1
,
Y
2
…….
Y
19
is
g
iv
en
to
th
e
co
n
tr
o
l
lo
g
ic
th
e
in
p
u
t
o
f
th
e
co
n
tr
o
l
lo
g
ic
is
tak
en
as
M
0
,
M
1
,
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
13
,
No
.
4
,
Dec
em
b
er
20
24
:
48
5
-
4
9
4
488
M
2
…….
M
19
ea
ch
is
o
f
1
6
-
b
it
in
p
u
t
an
d
N
0
,
N
1
,
N
2
…….
N
19
is
th
e
o
u
tp
u
t
o
f
th
e
co
n
tr
o
l
lo
g
ic
wh
ich
i
s
also
1
6
-
b
it.
Af
ter
th
is
,
th
e
2
×1
m
u
ltip
lex
er
is
u
s
ed
.
−
Mu
ltip
lex
er
(
2
×
1
)
:
T
h
e
r
e
ar
e
2
0
o
u
tp
u
ts
o
f
t
h
e
co
n
tr
o
l
lo
g
ic
ea
ch
p
air
o
f
two
o
u
tp
u
ts
is
g
iv
en
to
a
(
2
×1
)
m
u
ltip
lex
er
,
s
o
th
e
2
0
o
u
tp
u
ts
o
f
th
e
co
n
tr
o
l lo
g
ic
n
o
w
b
ec
o
m
e
1
0
o
u
tp
u
ts
.
−
Dem
u
ltip
lex
er
(
1
×
2
)
:
T
h
e
1
0
o
u
tp
u
ts
o
f
th
e
(
2
×
1
)
m
u
ltip
lex
er
ar
e
ag
ain
g
iv
en
to
th
e
(
1
×
2
)
d
em
u
ltip
lex
er
w
h
ich
n
o
w
b
ec
o
m
es
2
0
o
u
tp
u
ts
,
an
d
th
e
o
u
tp
u
t
o
f
th
e
(
1
×
2
)
d
em
u
ltip
lex
er
is
ag
ain
g
iv
e
n
to
co
n
tr
o
l
lo
g
ic
wh
ich
is
th
e
s
ec
o
n
d
h
id
d
en
lay
e
r
.
T
h
e
2
0
o
u
tp
u
ts
o
f
th
e
co
n
tr
o
l
lo
g
ic
ar
e
g
iv
en
to
th
e
2
0
to
1
d
em
u
ltip
lex
er
s
o
we
g
et
t
h
e
s
in
g
le
o
u
tp
u
t
o
f
1
6
b
its
f
r
o
m
th
is
.
Fig
u
r
e
1
.
Mu
ltil
ay
er
f
eed
-
f
o
r
w
ar
d
ar
tific
ial
n
eu
r
al
n
etwo
r
k
(
MFFANN)
Fig
u
r
e
2
.
B
lo
ck
d
iag
r
am
r
ep
r
e
s
en
tatio
n
o
f
MFFANN
Ad
d
er
s
an
d
m
u
ltip
lier
s
ar
e
im
p
o
r
tan
t
c
o
m
p
o
n
en
ts
th
at
ar
e
u
tili
ze
d
in
ANNs
to
ca
r
r
y
o
u
t
t
h
e
cr
u
cial
m
ath
em
atica
l
o
p
er
atio
n
s
th
at
ar
e
r
eq
u
ir
ed
f
o
r
p
r
o
ce
s
s
in
g
an
d
lear
n
in
g
.
T
h
is
is
a
co
m
p
r
eh
en
s
iv
e
lo
o
k
at
th
e
r
o
les
th
at
th
ey
p
lay
[
3
5
]
.
I
n
a
n
ANN
ea
ch
n
eu
r
o
n
g
ets
s
ev
e
r
al
in
p
u
ts
,
ea
ch
o
f
wh
ich
h
as
a
weig
h
t
ass
o
ciate
d
with
it.
T
h
e
n
et
in
p
u
t
to
th
e
n
eu
r
o
n
is
f
o
r
m
ed
b
y
ad
d
i
n
g
u
p
all
th
ese
co
n
tr
ib
u
tio
n
s
to
th
e
n
eu
r
o
n
.
E
ac
h
in
p
u
t
th
at
is
r
ec
eiv
ed
b
y
a
n
e
u
r
o
n
is
m
u
ltip
lied
b
y
th
e
weig
h
t
th
at
co
r
r
esp
o
n
d
s
to
th
at
in
p
u
t.
T
o
alter
th
e
in
f
lu
en
ce
th
at
ea
ch
in
p
u
t
h
as
o
n
th
e
o
u
t
p
u
t
o
f
th
e
n
eu
r
o
n
is
an
e
x
tr
em
ely
im
p
o
r
tan
t
s
tep
.
His
m
u
ltip
licatio
n
tak
es
p
lace
f
o
r
e
v
er
y
in
p
u
t
th
at
th
e
n
eu
r
o
n
r
ec
eiv
es.
Mu
ltip
lier
s
ar
e
u
tili
ze
d
in
t
h
e
p
r
o
ce
s
s
o
f
ca
lcu
lati
n
g
g
r
ad
ien
ts
d
u
r
in
g
th
e
tr
ain
in
g
o
f
ANNs
th
at
ar
e
tr
ain
ed
u
s
in
g
b
ac
k
p
r
o
p
ag
a
tio
n
.
T
h
e
p
r
o
ce
s
s
o
f
m
u
ltip
ly
in
g
d
e
r
iv
ativ
es
is
p
er
f
o
r
m
ed
to
ca
lcu
late
th
e
e
r
r
o
r
g
r
a
d
ien
t
r
e
g
ar
d
i
n
g
t
h
e
weig
h
ts
an
d
in
p
u
ts
.
I
t
is
ess
en
tial
f
o
r
h
ar
d
war
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
Desig
n
a
n
d
imp
leme
n
t
a
tio
n
o
f d
ee
p
n
e
u
r
a
l n
etw
o
r
k
h
a
r
d
w
a
r
e
ch
ip
a
n
d
its
p
erfo
r
ma
n
ce
...
(
A
r
u
n
a
P
a
n
t
)
489
ac
ce
ler
ato
r
s
s
u
ch
as
FP
GAs
an
d
ap
p
licatio
n
-
s
p
ec
if
ic
in
teg
r
ated
cir
cu
its
(
ASI
C
)
th
at
ar
e
u
s
ed
f
o
r
n
eu
r
a
l
n
etwo
r
k
s
to
h
av
e
ad
d
er
s
a
n
d
m
u
ltip
lier
s
th
at
ar
e
im
p
lem
en
t
ed
p
r
o
p
er
ly
an
d
ef
f
icien
tly
.
A
NNs
ca
n
h
av
e
t
h
eir
s
p
ee
d
g
r
ea
tly
in
cr
ea
s
ed
a
n
d
th
eir
p
o
wer
co
n
s
u
m
p
tio
n
d
r
a
m
atica
lly
d
ec
r
ea
s
ed
if
th
ey
ar
e
d
esig
n
ed
with
o
p
tim
ized
ad
d
er
s
an
d
m
u
ltip
lier
s
,
s
u
ch
as
ca
r
r
y
-
lo
o
k
ah
ea
d
ad
d
er
s
an
d
b
o
o
t
h
m
u
ltip
lier
s
,
r
esp
ec
tiv
ely
.
T
h
e
co
n
ce
p
t
o
f
p
a
r
allelis
m
is
u
tili
ze
d
in
h
a
r
d
war
e
im
p
lem
en
tati
o
n
s
.
T
h
e
i
n
f
er
en
ce
an
d
tr
ain
in
g
p
r
o
ce
s
s
es
ca
n
b
e
s
p
ed
u
p
b
y
d
esig
n
i
n
g
m
u
ltip
lier
s
an
d
a
d
d
er
s
t
o
r
u
n
i
n
p
ar
allel
[
3
6
]
.
T
h
is
allo
ws
f
o
r
th
e
s
im
u
ltan
e
o
u
s
co
m
p
u
tatio
n
o
f
m
a
n
y
n
eu
r
o
n
o
u
tp
u
ts
,
wh
ich
s
p
ee
d
s
u
p
th
e
p
r
o
ce
s
s
.
T
h
e
ch
a
r
ac
ter
is
tics
o
f
a
n
eu
r
al
n
etwo
r
k
r
ef
er
to
th
e
v
ar
iab
les
a
n
d
q
u
ali
ties
o
f
th
e
d
ataset
[
3
7
]
.
T
o
e
n
h
an
ce
th
e
ac
cu
r
ac
y
o
f
y
o
u
r
m
o
d
el,
it
is
cu
s
to
m
ar
y
to
u
s
e
a
s
u
b
s
et
o
f
v
ar
ia
b
les.
T
h
er
ef
o
r
e
,
th
e
attr
i
b
u
tes
o
f
a
n
e
u
r
al
n
etwo
r
k
ar
e
in
th
e
i
n
p
u
t
l
ay
er
,
r
ath
er
th
a
n
th
e
n
o
d
es in
th
e
h
id
d
e
n
lay
er
.
I
n
ANNs,
th
e
b
asic a
ct
iv
ities
o
f
s
u
m
m
in
g
weig
h
te
d
in
p
u
ts
an
d
co
m
p
u
tin
g
g
r
ad
ien
ts
ar
e
p
er
f
o
r
m
ed
b
y
ad
d
er
s
an
d
m
u
ltip
lier
s
,
wh
ich
ar
e
ess
en
tial
co
m
p
o
n
e
n
ts
f
o
r
co
n
d
u
ctin
g
th
ese
o
p
er
atio
n
s
[
3
8
]
.
T
h
e
p
e
r
f
o
r
m
an
ce
an
d
en
er
g
y
ef
f
icien
cy
o
f
n
eu
r
al
n
etw
o
r
k
h
a
r
d
war
e
ac
ce
ler
ato
r
s
ar
e
d
ir
ec
tly
co
r
r
elate
d
to
th
e
ef
f
icien
c
y
with
wh
ich
th
ey
ar
e
im
p
lem
e
n
ted
.
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
e
d
esig
n
s
tr
u
ctu
r
e
o
f
a
m
u
ltil
ay
er
f
ee
d
-
f
o
r
war
d
n
eu
r
al
n
etwo
r
k
h
ar
d
war
e
ch
ip
co
n
s
is
t
s
o
f
th
e
m
u
ltip
lier
,
co
n
tr
o
l
lo
g
ic,
m
u
ltip
lex
er
s
(
2
×
1
)
,
d
em
u
ltip
l
ex
er
s
(
1
×
2
)
,
an
d
an
o
th
e
r
m
u
ltip
lex
er
(
2
0
×
1
)
co
m
p
o
n
en
ts
.
Fig
u
r
e
3
s
h
o
ws th
e
R
T
L
r
ep
r
esen
tatio
n
o
f
a
m
u
ltil
ay
er
n
eu
r
al
n
etwo
r
k
.
Fig
u
r
e
3
.
R
T
L
o
f
m
u
ltil
ay
er
A
NN
Fig
u
r
e
4
p
r
esen
ts
th
e
s
im
u
l
atio
n
in
p
u
ts
v
er
if
ied
i
n
th
e
s
im
u
latio
n
an
d
Fig
u
r
e
5
p
r
esen
ts
th
e
s
im
u
latio
n
wav
ef
o
r
m
o
u
tp
u
ts
o
f
a
m
u
ltil
ay
er
n
eu
r
al
n
etwo
r
k
s
im
u
lated
in
th
e
d
esig
n
.
Fig
u
r
e
6
p
r
esen
ts
th
e
m
u
ltil
ay
er
h
ar
d
war
e
c
h
ip
u
tili
za
tio
n
ch
ar
t,
an
d
Fig
u
r
e
7
p
r
e
s
en
ts
th
e
m
u
ltil
ay
er
h
ar
d
war
e
ch
ip
u
tili
za
tio
n
p
ie
ch
ar
t.
T
h
e
u
s
ag
e
o
f
R
T
L
is
t
o
p
r
esen
t
t
h
e
co
m
p
lete
d
etail
s
o
f
th
e
i
n
p
u
t
a
n
d
o
u
tp
u
t
p
in
s
in
th
e
d
esig
n
.
T
h
e
s
im
u
latio
n
in
p
u
t a
n
d
o
u
tp
u
t p
i
n
s
with
s
tim
u
li in
p
u
ts
ar
e
s
h
o
wn
in
th
e
wav
ef
o
r
m
s
in
th
e
f
o
r
m
o
f
b
in
ar
y
d
ata.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
13
,
No
.
4
,
Dec
em
b
er
20
24
:
48
5
-
4
9
4
490
Fig
u
r
e
4
.
Simu
latio
n
wav
ef
o
r
m
o
f
in
p
u
ts
f
o
r
m
u
ltil
ay
er
AN
N
Neu
r
o
n
_
X
0
_
i
n
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
8
-
b
it
in
p
u
t
o
f
t
h
e
n
e
u
r
o
n
X
0
an
d
Ne
u
r
o
n
_
W
eig
h
t_
W
0
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
i
n
g
weig
h
t
in
p
u
t.
Neu
r
o
n
_
X
1
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
8
-
b
it
in
p
u
t
o
f
th
e
n
eu
r
o
n
X
1
an
d
Neu
r
o
n
_
W
eig
h
t_
W
1
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
in
g
weig
h
t
in
p
u
t.
Neu
r
o
n
_
X
2
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
8
-
b
it
in
p
u
t
o
f
th
e
n
eu
r
o
n
X
2
a
n
d
Neu
r
o
n
_
W
eig
h
t_
W
2
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
in
g
weig
h
t
in
p
u
t.
Neu
r
o
n
_
X
3
_
in
p
u
t
(
7
:
0
)
p
r
esen
ts
th
e
8
-
b
it
in
p
u
t
o
f
th
e
n
eu
r
o
n
X
3
an
d
Neu
r
o
n
_
W
eig
h
t_
W
3
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
i
n
g
weig
h
t
in
p
u
t.
Neu
r
o
n
_
X
4
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
8
-
b
it
in
p
u
t
o
f
th
e
n
eu
r
o
n
X
4
an
d
Neu
r
o
n
_
W
eig
h
t_
W
4
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
in
g
weig
h
t
in
p
u
t.
Neu
r
o
n
_
X
5
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
8
-
b
it
in
p
u
t
o
f
th
e
n
eu
r
o
n
X
5
a
n
d
Neu
r
o
n
_
W
eig
h
t_
W
5
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
in
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
Desig
n
a
n
d
imp
leme
n
t
a
tio
n
o
f d
ee
p
n
e
u
r
a
l n
etw
o
r
k
h
a
r
d
w
a
r
e
ch
ip
a
n
d
its
p
erfo
r
ma
n
ce
...
(
A
r
u
n
a
P
a
n
t
)
491
weig
h
t
in
p
u
t.
Neu
r
o
n
_
X
6
_
in
p
u
t
(
7
:
0
)
p
r
esen
ts
th
e
8
-
b
it
in
p
u
t
o
f
th
e
n
eu
r
o
n
X
6
an
d
Neu
r
o
n
_
W
eig
h
t_
W
6
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
i
n
g
weig
h
t
in
p
u
t.
Neu
r
o
n
_
X
7
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
8
-
b
it
in
p
u
t
o
f
th
e
n
eu
r
o
n
X
7
an
d
Neu
r
o
n
_
W
eig
h
t_
W
7
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
in
g
weig
h
t
in
p
u
t.
Neu
r
o
n
_
X
8
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
8
-
b
it
in
p
u
t
o
f
th
e
n
eu
r
o
n
X
8
a
n
d
Neu
r
o
n
_
W
eig
h
t_
W
8
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
in
g
weig
h
t
in
p
u
t.
Neu
r
o
n
_
X
9
_
in
p
u
t
(
7
:
0
)
p
r
esen
ts
th
e
8
-
b
it
in
p
u
t
o
f
th
e
n
eu
r
o
n
X
9
an
d
Neu
r
o
n
_
W
eig
h
t_
W
9
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
in
g
weig
h
t
in
p
u
t.
Ne
u
r
o
n
_
X
10
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
8
-
b
it
i
n
p
u
t
o
f
th
e
n
eu
r
o
n
X
10
an
d
Neu
r
o
n
_
W
eig
h
t_
W
10
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
in
g
weig
h
t
in
p
u
t.
Neu
r
o
n
_
X
11
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
8
-
b
it
in
p
u
t
o
f
th
e
n
eu
r
o
n
X
11
an
d
Neu
r
o
n
_
W
ei
g
h
t_
W
11
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
i
n
g
weig
h
t
in
p
u
t.
Neu
r
o
n
_
X
12
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
8
-
b
it
in
p
u
t
o
f
th
e
n
eu
r
o
n
X
12
a
n
d
Neu
r
o
n
_
W
eig
h
t_
W
12
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
in
g
weig
h
t
in
p
u
t.
Neu
r
o
n
_
X
13
_
i
n
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
8
-
b
it
in
p
u
t
o
f
th
e
n
e
u
r
o
n
X
13
an
d
Neu
r
o
n
_
W
eig
h
t_
W
13
_
in
p
u
t
(
7
:
0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
in
g
weig
h
t
in
p
u
t.
Neu
r
o
n
_
X
14
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
8
-
b
it
in
p
u
t
o
f
th
e
n
eu
r
o
n
X
14
an
d
Neu
r
o
n
_
W
eig
h
t_
W
14
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
in
g
weig
h
t
in
p
u
t.
Ne
u
r
o
n
_
X
15
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
8
-
b
it
i
n
p
u
t
o
f
th
e
n
eu
r
o
n
X
15
an
d
Neu
r
o
n
_
W
eig
h
t_
W
15
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
in
g
weig
h
t
in
p
u
t.
Neu
r
o
n
_
X
16
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
8
-
b
it
in
p
u
t
o
f
th
e
n
eu
r
o
n
X
16
an
d
Neu
r
o
n
_
W
ei
g
h
t_
W
16
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
i
n
g
weig
h
t
in
p
u
t.
Neu
r
o
n
_
X
17
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
8
-
b
it
in
p
u
t
o
f
th
e
n
eu
r
o
n
X
17
a
n
d
Neu
r
o
n
_
W
eig
h
t_
W
17
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
in
g
weig
h
t
in
p
u
t.
Neu
r
o
n
_
X
18
_
i
n
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
8
-
b
it
in
p
u
t
o
f
th
e
n
e
u
r
o
n
X
18
an
d
Neu
r
o
n
_
W
eig
h
t_
W
18
_
in
p
u
t
(
7
:
0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
in
g
weig
h
t
in
p
u
t.
Neu
r
o
n
_
X
19
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
8
-
b
it
in
p
u
t
o
f
th
e
n
eu
r
o
n
X
19
an
d
Neu
r
o
n
_
W
eig
h
t_
W
19
_
in
p
u
t
(
7
:0
)
p
r
esen
ts
th
e
co
r
r
esp
o
n
d
i
n
g
weig
h
t
in
p
u
t.
Neu
r
o
n
_
m
id
_
lay
er
_
s
o
u
r
ce
_
ad
d
r
ess
(
4
:0
)
p
r
esen
ts
th
e
ad
d
r
ess
o
f
th
e
m
id
-
la
y
er
b
lo
c
k
s
to
s
u
p
p
o
r
t
th
e
d
esig
n
.
Neu
r
o
n
_
m
id
_
lay
er
_
d
esti
n
atio
n
_
ad
d
r
es
s
(
4
:0
)
p
r
esen
ts
th
e
d
esti
n
atio
n
ad
d
r
ess
o
f
th
e
m
i
d
lay
er
s
to
s
u
p
p
o
r
t
th
e
d
esig
n
.
Mu
x
2
1
_
s
elec
to
r
is
u
s
ed
to
s
elec
t
th
e
s
p
ec
if
ic
m
u
ltip
lex
er
o
f
th
e
b
lo
ck
i
n
in
p
u
t.
Dem
u
ltip
lex
e
r
1
2
_
s
elec
to
r
is
u
s
ed
to
s
elec
t
th
e
s
p
ec
if
ic
d
em
u
ltip
lex
er
o
f
th
e
b
lo
ck
in
o
u
tp
u
t.
Po
r
t_
a
d
d
r
ess
(
4
:0
)
is
th
e
ad
d
r
ess
to
s
u
p
p
o
r
t
th
e
m
em
o
r
y
b
lo
ck
s
.
T
h
e
co
m
p
lete
o
u
tp
u
t
o
f
t
h
e
en
tire
m
o
d
u
le
is
tak
en
b
y
n
eu
r
o
n
_
o
u
tp
u
t (
1
5
:0
)
.
Fig
u
r
e
5
.
Simu
latio
n
wav
ef
o
r
m
o
u
tp
u
ts
o
f
m
u
ltil
ay
er
ANN
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
13
,
No
.
4
,
Dec
em
b
er
20
24
:
48
5
-
4
9
4
492
Fig
u
r
e
6
.
Mu
ltil
ay
er
h
ar
d
war
e
ch
ip
u
tili
za
tio
n
ch
a
r
t
Fig
u
r
e
7
.
Mu
ltil
ay
er
h
ar
d
war
e
ch
ip
u
tili
za
tio
n
p
ie
c
h
ar
t
T
ab
le
1
lis
ts
th
e
h
ar
d
war
e
p
ar
am
eter
s
u
tili
ze
d
b
y
th
e
m
u
ltil
ay
er
h
ar
d
war
e
ch
ip
an
d
th
e
d
ev
ice
u
tili
za
tio
n
r
ep
o
r
t
f
o
r
th
is
n
e
u
r
o
m
o
r
p
h
ic
h
a
r
d
war
e
c
h
ip
is
ta
k
en
f
r
o
m
Xilin
x
s
o
f
twar
e.
T
h
e
n
u
m
b
er
o
f
s
lices
u
s
ed
is
9
6
0
in
th
e
m
u
ltil
ay
er
h
ar
d
war
e
ch
ip
,
wh
ich
is
o
u
t
o
f
a
to
tal
o
f
4
8
0
0
,
M
o
r
eo
v
e
r
,
th
e
n
u
m
b
er
o
f
s
lices
L
UT
s
u
s
ed
is
1
5
4
7
o
u
t o
f
a
to
tal
o
f
2
4
0
0
.
I
n
p
u
t
-
o
u
tp
u
ts
u
s
e
d
ar
e
5
3
o
u
t
o
f
a
to
tal
o
f
1
0
2
.
T
h
e
n
u
m
b
er
o
f
DSP
elem
en
ts
u
s
ed
is
8
o
u
t
o
f
a
to
tal
o
f
8
.
T
h
e
to
tal
d
ela
y
u
s
e
d
in
m
u
ltil
ay
er
h
ar
d
war
e
ch
i
p
s
is
1
1
.
6
8
8
n
s
.
T
h
e
m
em
o
r
y
u
tili
ze
d
f
o
r
th
e
m
u
lti
lay
er
h
a
r
d
war
e
ch
ip
is
4
6
6
7
9
4
4
k
ilo
b
y
tes.
T
h
e
tar
g
et
d
ev
ic
e
u
s
ed
is
s
p
ar
tan
-
6
FP
GA
with
th
e
d
ev
ice
(
XC
6
T
S4
-
2
T
QG1
4
4
)
f
o
r
s
im
u
latio
n
an
d
s
y
n
th
esis
.
T
h
e
Sp
ar
tan
-
6
f
am
ily
o
f
f
er
s
th
e
m
o
s
t
ad
v
an
ce
d
ca
p
ab
ilit
ies
f
o
r
s
y
s
tem
in
teg
r
atio
n
at
th
e
l
o
west
p
o
s
s
ib
le
o
v
er
all
co
s
t
f
o
r
ap
p
licatio
n
s
th
at
r
eq
u
ir
e
h
i
g
h
v
o
l
u
m
e
ca
p
ac
ity
.
I
t
is
d
ev
elo
p
ed
u
s
in
g
a
n
estab
lis
h
ed
lo
w
-
p
o
wer
c
o
p
p
er
p
r
o
ce
s
s
tech
n
iq
u
e
th
at
o
p
er
ates
at
4
5
n
an
o
m
eter
s
.
D
NNs
r
eq
u
ir
e
h
ig
h
m
em
o
r
y
-
to
-
p
r
o
ce
s
s
in
g
u
n
it
d
ata
tr
an
s
m
is
s
io
n
r
ates.
Me
m
o
r
y
b
an
d
wid
th
ca
n
lim
it
lar
g
e
-
s
c
ale
m
o
d
els
with
m
an
y
lay
er
s
an
d
p
ar
am
eter
s
.
W
h
en
tr
ain
ed
an
d
r
u
n
o
n
n
o
n
-
o
p
tim
al
h
ar
d
war
e,
DNNs
ca
n
u
s
e
a
lo
t
o
f
p
o
wer
.
T
h
is
af
f
e
cts
d
ata
m
o
b
ile,
ce
n
ter
s
,
an
d
em
b
ed
d
e
d
d
ev
ices
s
in
ce
en
er
g
y
ef
f
icien
cy
is
cr
u
c
ial.
W
h
en
wo
r
k
in
g
with
lar
g
er
m
o
d
els
o
r
m
o
r
e
co
m
p
lex
DNN
task
s
,
s
ca
lab
ilit
y
r
eq
u
ir
es
ef
f
icien
t
p
ar
alleliza
tio
n
ac
r
o
s
s
n
u
m
er
o
u
s
p
r
o
ce
s
s
in
g
u
n
its
.
Du
e
to
d
ee
p
n
eu
r
al
n
e
two
r
k
s
’
co
m
p
le
x
ity
an
d
r
eso
u
r
ce
r
eq
u
ir
em
en
ts
,
h
a
r
d
war
e
im
p
lem
e
n
tatio
n
h
as b
e
co
m
e
h
ar
d
er
.
T
ab
le
1
.
Har
d
war
e
p
ar
a
m
eter
s
u
tili
ze
d
b
y
m
u
ltil
ay
er
h
ar
d
wa
r
e
ch
ip
P
a
r
a
me
t
e
r
V
a
l
u
e
A
N
N
C
h
i
p
si
z
e
ANN
-
20
M
u
l
t
i
p
l
i
e
r
u
se
d
20
Ta
r
g
e
t
d
e
v
i
c
e
(
F
P
G
A
)
S
p
a
r
t
a
n
-
6
(
x
c
6
s
t
x
4
-
2
t
9
g
1
4
4
)
U
sed
LU
Ts
F
F
P
s
3
5
2
/
2
1
5
5
S
l
i
c
e
s
9
6
0
/
4
8
0
0
D
e
l
a
y
(
n
s)
6
.
3
5
9
I
O
B
s
5
3
/
1
0
2
M
e
m
o
r
y
(
K
B
)
4
6
6
5
3
0
4
960
353
352
0
200
400
600
800
1000
1200
S
l
i
c
e
re
g
i
s
te
rs
N
u
mb
e
r o
f I
O
B
s
Us
e
d
L
UT
-F
F
s
(
fli
p
-flo
p
)
Par
ame
te
r
s
u
ti
l
i
ze
d
S
i
ze
(
M
u
l
ti
l
y
er
A
N
N
n
euro
n
i
n
p
u
t
-
20)
Har
dw
ar
e
pa
r
ameters
ut
i
l
i
z
ed
20%
16,33%
51,96%
100%
F
PG
A
H
ar
d
war
e
Uti
l
i
zati
o
n
S
l
i
c
e
re
g
i
s
te
rs
L
UT
F
F
s
IO
B
s
DS
P
B
l
o
c
ks
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
Desig
n
a
n
d
imp
leme
n
t
a
tio
n
o
f d
ee
p
n
e
u
r
a
l n
etw
o
r
k
h
a
r
d
w
a
r
e
ch
ip
a
n
d
its
p
erfo
r
ma
n
ce
...
(
A
r
u
n
a
P
a
n
t
)
493
5.
CO
NCLU
SI
O
N
T
h
e
m
o
d
el
o
f
n
e
u
r
al
n
etwo
r
k
s
is
ca
lled
a
d
ee
p
n
eu
r
al
n
etwo
r
k
wh
en
it
u
s
es
two
o
r
m
o
r
e
h
id
d
e
n
lay
er
s
.
DNN
ch
ip
s
b
ased
o
n
FP
GA
s
ca
n
b
e
o
p
tim
ized
f
o
r
p
ar
ticu
lar
n
eu
r
al
n
etwo
r
k
m
o
d
els
o
r
task
s
,
wh
ic
h
co
u
ld
lea
d
to
b
etter
p
er
f
o
r
m
a
n
ce
th
an
g
en
e
r
ic
alter
n
ativ
es.
I
t
is
d
is
tin
g
u
is
h
ed
b
y
th
e
u
ti
lizatio
n
o
f
ar
tific
ial
n
eu
r
al
n
etwo
r
k
s
with
v
ar
i
o
u
s
o
th
er
la
y
er
s
o
f
p
r
o
ce
s
s
in
g
u
n
it
s
,
s
o
m
etim
es
k
n
o
w
n
as n
eu
r
o
n
s
o
r
n
o
d
es,
to
lear
n
r
ep
r
esen
tatio
n
s
o
f
in
p
u
t.
T
h
e
s
e
n
eu
r
al
n
etwo
r
k
s
m
ay
a
u
t
o
m
atica
lly
id
en
tif
y
p
atter
n
s
o
r
f
ea
t
u
r
es
in
d
ata
with
o
u
t
th
e
n
ee
d
f
o
r
e
x
p
licit
p
r
o
g
r
a
m
m
in
g
.
T
h
e
m
u
ltil
ay
er
h
ar
d
war
e
ch
i
p
f
o
r
2
0
-
n
eu
r
o
n
i
n
p
u
t
is
d
esig
n
ed
i
n
S
p
ar
tan
-
6
(
XC
6
SLX4
-
2
T
QG
1
4
4
)
FP
GA
Xilin
x
I
SE
1
4
.
7
s
u
cc
ess
f
u
lly
.
T
h
e
R
T
L
,
in
te
r
n
al
s
ch
em
atic,
an
d
s
im
u
latio
n
o
f
th
e
m
u
ltil
ay
er
h
ar
d
war
e
a
r
e
v
er
if
ied
.
T
h
e
h
ar
d
war
e
p
a
r
am
eter
u
tili
za
tio
n
r
ep
o
r
ts
s
h
o
w
th
at
L
UT
’
s
f
lip
-
f
lo
p
u
s
ed
is
3
5
2
o
u
t
o
f
2
1
5
5
wh
ic
h
is
1
6
.
3
3
%.
T
h
e
n
u
m
b
er
o
f
b
o
n
d
ed
I
O
B
s
u
s
ed
is
5
3
o
u
t
o
f
th
e
av
ailab
le
to
tal
1
0
2
I
O
B
s
,
wh
ich
is
5
1
.
9
6
%.
T
h
e
n
u
m
b
e
r
o
f
DSP
elem
en
ts
u
s
ed
is
8
t
o
tal
o
u
t
o
f
8
w
h
ich
i
s
1
0
0
.
0
0
%,
Nu
m
b
e
r
o
f
s
lice
r
eg
is
ter
s
u
s
ed
is
9
6
0
o
f
th
e
t
o
t
al
av
ailab
le
4
8
0
0
wh
ich
s
h
o
w
s
th
at
it
is
2
0
.
0
0
%
u
tili
ze
d
.
T
h
e
m
u
ltil
ay
er
ANN
d
esig
n
s
u
p
p
o
r
ted
3
1
3
MH
z
f
r
eq
u
en
c
y
,
d
ela
y
6
.
3
5
9
n
s
,
m
e
m
o
r
y
4
6
6
5
3
0
4
KB
,
an
d
o
p
tim
al
p
o
wer
esti
m
atio
n
o
f
1
1
.
4
5
n
W
.
A
n
ew,
m
o
r
e
ef
f
icien
t
d
u
al
-
r
eg
is
ter
6
-
in
p
u
t
L
UT
s
lo
g
ic
an
d
a
lar
g
e
r
a
n
g
e
o
f
b
u
ilt
-
in
s
y
s
tem
-
lev
el
b
lo
c
k
s
ar
e
b
o
th
in
clu
d
e
d
in
t
h
e
Sp
ar
ta
n
-
6
f
am
ily
o
f
in
teg
r
ated
ci
r
cu
its
,
wh
ich
o
f
f
e
r
s
th
e
b
est m
ix
o
f
c
o
s
t,
p
o
wer
,
an
d
p
e
r
f
o
r
m
an
ce
.
ACK
NO
WL
E
DG
E
M
E
NT
S
T
h
an
k
s
to
th
e
UPES
VL
SI
d
esig
n
an
d
DSP L
ab
to
p
r
o
v
id
e
th
e
p
latf
o
r
m
f
o
r
ca
r
r
y
in
g
o
u
t th
e
wo
r
k
.
RE
F
E
R
E
NC
E
S
[
1
]
M
.
W
i
e
l
g
o
sz
a
n
d
M
.
K
a
r
w
a
t
o
w
sk
i
,
“
M
a
p
p
i
n
g
n
e
u
r
a
l
n
e
t
w
o
r
k
s
t
o
F
P
G
A
-
b
a
se
d
I
o
T
d
e
v
i
c
e
s
f
o
r
u
l
t
r
a
-
l
o
w
l
a
t
e
n
c
y
p
r
o
c
e
ssi
n
g
,
”
S
e
n
so
rs
,
v
o
l
.
1
9
,
n
o
.
1
3
,
Ju
l
.
2
0
1
9
,
d
o
i
:
1
0
.
3
3
9
0
/
s1
9
1
3
2
9
8
1
.
[
2
]
A
.
K
u
mar
,
P
.
S
h
a
r
ma
,
M
.
K
.
G
u
p
t
a
,
a
n
d
R
.
K
u
mar,
“
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
b
a
sed
r
e
so
u
r
c
e
u
t
i
l
i
z
a
t
i
o
n
a
n
d
p
r
e
-
e
s
t
i
m
a
t
i
o
n
f
o
r
n
e
t
w
o
r
k
o
n
c
h
i
p
(
N
o
C
)
c
o
mm
u
n
i
c
a
t
i
o
n
,
”
Wi
re
l
e
ss
Pe
rs
o
n
a
l
C
o
m
m
u
n
i
c
a
t
i
o
n
s
,
v
o
l
.
1
0
2
,
n
o
.
3
,
p
p
.
2
2
1
1
–
2
2
3
1
,
O
c
t
.
2
0
1
8
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
1
2
7
7
-
0
1
8
-
5
3
7
6
-
3.
[
3
]
W
.
H
a
e
n
sc
h
,
T.
G
o
k
m
e
n
,
a
n
d
R
.
P
u
r
i
,
“
Th
e
n
e
x
t
g
e
n
e
r
a
t
i
o
n
o
f
d
e
e
p
l
e
a
r
n
i
n
g
h
a
r
d
w
a
r
e
:
a
n
a
l
o
g
c
o
m
p
u
t
i
n
g
,
”
i
n
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
I
EEE
,
J
a
n
.
2
0
1
9
,
v
o
l
.
1
0
7
,
n
o
.
1
,
p
p
.
1
0
8
–
1
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
JP
R
O
C
.
2
0
1
8
.
2
8
7
1
0
5
7
.
[
4
]
Y
.
Le
c
u
n
,
Y
.
B
e
n
g
i
o
,
a
n
d
G
.
H
i
n
t
o
n
,
“
D
e
e
p
l
e
a
r
n
i
n
g
,
”
N
a
t
u
re
,
v
o
l
.
5
2
1
,
n
o
.
7
5
5
3
,
p
p
.
4
3
6
–
4
4
4
,
M
a
y
2
0
1
5
,
d
o
i
:
1
0
.
1
0
3
8
/
n
a
t
u
r
e
1
4
5
3
9
.
[
5
]
L.
D
e
n
g
e
t
a
l
.
,
“
R
e
c
e
n
t
a
d
v
a
n
c
e
s
i
n
d
e
e
p
l
e
a
r
n
i
n
g
f
o
r
sp
e
e
c
h
r
e
sea
r
c
h
a
t
M
i
c
r
o
s
o
f
t
,
”
i
n
2
0
1
3
I
E
EE
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
Ac
o
u
s
t
i
c
s
,
S
p
e
e
c
h
a
n
d
S
i
g
n
a
l
Pr
o
c
e
ss
i
n
g
,
M
a
y
2
0
1
3
,
p
p
.
8
6
0
4
–
8
6
0
8
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
A
S
S
P
.
2
0
1
3
.
6
6
3
9
3
4
5
.
[
6
]
A
.
K
r
i
z
h
e
v
s
k
y
,
I
.
S
u
t
s
k
e
v
e
r
,
a
n
d
G
.
E.
H
i
n
t
o
n
,
“
I
mag
e
N
e
t
c
l
a
ssi
f
i
c
a
t
i
o
n
w
i
t
h
d
e
e
p
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s
,
”
C
o
m
m
u
n
i
c
a
t
i
o
n
s
o
f
t
h
e
A
C
M
,
v
o
l
.
6
0
,
n
o
.
6
,
p
p
.
8
4
–
9
0
,
M
a
y
2
0
1
7
,
d
o
i
:
1
0
.
1
1
4
5
/
3
0
6
5
3
8
6
.
[
7
]
D
.
S
i
l
v
e
r
e
t
a
l
.
,
“
M
a
s
t
e
r
i
n
g
t
h
e
g
a
me
o
f
g
o
w
i
t
h
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
s
a
n
d
t
r
e
e
se
a
r
c
h
,
”
N
a
t
u
r
e
,
v
o
l
.
5
2
9
,
n
o
.
7
5
8
7
,
p
p
.
4
8
4
–
4
8
9
,
2
0
1
6
,
d
o
i
:
1
0
.
1
0
3
8
/
n
a
t
u
r
e
1
6
9
6
1
.
[
8
]
D
.
W
a
n
g
,
L
.
D
e
n
g
,
P
.
Ta
n
g
,
C
.
M
a
,
a
n
d
J.
P
e
i
,
“
F
P
G
A
-
b
a
se
d
n
e
u
r
o
m
o
r
p
h
i
c
c
o
m
p
u
t
i
n
g
s
y
st
e
m
w
i
t
h
a
sc
a
l
a
b
l
e
r
o
u
t
i
n
g
n
e
t
w
o
r
k
,
”
i
n
2
0
1
5
1
5
t
h
N
o
n
-
V
o
l
a
t
i
l
e
Me
m
o
ry
T
e
c
h
n
o
l
o
g
y
S
y
m
p
o
s
i
u
m
(
N
V
MTS
)
,
O
c
t
.
2
0
1
5
,
p
p
.
1
–
4
,
d
o
i
:
1
0
.
1
1
0
9
/
N
V
M
TS.
2
0
1
5
.
7
4
5
7
4
3
2
.
[
9
]
M
.
A
.
D
i
a
s
a
n
d
D
.
A
.
P
.
F
e
r
r
e
i
r
a
,
“
D
e
e
p
l
e
a
r
n
i
n
g
i
n
r
e
c
o
n
f
i
g
u
r
a
b
l
e
h
a
r
d
w
a
r
e
:
a
su
r
v
e
y
,
”
i
n
2
0
1
9
I
EEE
I
n
t
e
r
n
a
t
i
o
n
a
l
Pa
r
a
l
l
e
l
a
n
d
D
i
st
ri
b
u
t
e
d
Pr
o
c
e
s
si
n
g
S
y
m
p
o
si
u
m
W
o
rks
h
o
p
s
(
I
PD
PS
W)
,
M
a
y
2
0
1
9
,
p
p
.
9
5
–
9
8
,
d
o
i
:
1
0
.
1
1
0
9
/
I
P
D
P
S
W
.
2
0
1
9
.
0
0
0
2
6
.
[
1
0
]
N
.
N
e
d
j
a
h
,
R
.
M
.
d
a
S
i
l
v
a
,
a
n
d
L.
d
e
M
a
c
e
d
o
M
o
u
r
e
l
l
e
,
“
C
o
mp
a
c
t
y
e
t
e
f
f
i
c
i
e
n
t
h
a
r
d
w
a
r
e
i
mp
l
e
m
e
n
t
a
t
i
o
n
o
f
a
r
t
i
f
i
c
i
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s
w
i
t
h
c
u
st
o
mi
z
e
d
t
o
p
o
l
o
g
y
,
”
Ex
p
e
r
t
S
y
s
t
e
m
s
w
i
t
h
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
3
9
,
n
o
.
1
0
,
p
p
.
9
1
9
1
–
9
2
0
6
,
A
u
g
.
2
0
1
2
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
sw
a
.
2
0
1
2
.
0
2
.
0
8
5
.
[
1
1
]
J
.
P
a
r
k
a
n
d
W
.
S
u
n
g
,
“
F
P
G
A
b
a
s
e
d
i
m
p
l
e
m
e
n
t
a
t
i
o
n
o
f
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
s
u
s
i
n
g
o
n
-
c
h
i
p
m
e
m
o
r
y
o
n
l
y
,
”
i
n
2
0
1
6
I
E
E
E
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
A
c
o
u
s
t
i
c
s
,
S
p
e
e
c
h
a
n
d
S
i
g
n
a
l
P
r
o
c
e
s
s
i
n
g
(
I
C
A
S
S
P
)
,
M
a
r
.
2
0
1
6
,
p
p
.
1
0
1
1
–
1
0
1
5
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
A
S
S
P
.
2
0
1
6
.
7
4
7
1
8
2
8
.
[
1
2
]
M
.
V
e
r
u
c
c
h
i
e
t
a
l
.
,
“
A
s
y
s
t
e
m
a
t
i
c
a
s
s
e
ssm
e
n
t
o
f
e
m
b
e
d
d
e
d
n
e
u
r
a
l
n
e
t
w
o
r
k
s
f
o
r
o
b
j
e
c
t
d
e
t
e
c
t
i
o
n
,
”
i
n
2
0
2
0
2
5
t
h
I
E
EE
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
E
m
e
r
g
i
n
g
T
e
c
h
n
o
l
o
g
i
e
s
a
n
d
F
a
c
t
o
r
y
A
u
t
o
m
a
t
i
o
n
(
E
T
F
A
)
,
S
e
p
.
2
0
2
0
,
p
p
.
9
3
7
–
9
4
4
,
d
o
i
:
1
0
.
1
1
0
9
/
E
T
F
A
4
6
5
2
1
.
2
0
2
0
.
9
2
1
2
1
3
0
.
[
1
3
]
T.
V
.
H
u
y
n
h
,
“
D
e
si
g
n
s
p
a
c
e
e
x
p
l
o
r
a
t
i
o
n
f
o
r
a
si
n
g
l
e
-
F
P
G
A
h
a
n
d
w
r
i
t
t
e
n
d
i
g
i
t
r
e
c
o
g
n
i
t
i
o
n
sy
st
e
m,
”
i
n
2
0
1
4
I
EEE
Fi
f
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
C
o
m
m
u
n
i
c
a
t
i
o
n
s
a
n
d
El
e
c
t
ro
n
i
c
s (I
C
C
E)
,
Ju
l
.
2
0
1
4
,
p
p
.
2
9
1
–
2
9
6
,
d
o
i
:
1
0
.
1
1
0
9
/
C
C
E
.
2
0
1
4
.
6
9
1
6
7
1
7
.
[
1
4
]
T.
V
.
H
u
y
n
h
,
“
E
v
a
l
u
a
t
i
o
n
o
f
a
r
t
i
f
i
c
i
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
a
r
c
h
i
t
e
c
t
u
r
e
s
f
o
r
p
a
t
t
e
r
n
r
e
c
o
g
n
i
t
i
o
n
o
n
F
P
G
A
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
C
o
m
p
u
t
i
n
g
a
n
d
D
i
g
i
t
a
l
S
y
st
e
m
s
,
v
o
l
.
6
,
n
o
.
3
,
p
p
.
1
3
3
–
1
3
8
,
2
0
1
7
,
d
o
i
:
1
0
.
1
2
7
8
5
/
I
JC
D
S
/
0
6
0
3
0
5
.
[
1
5
]
A
.
K
u
m
a
r
,
G
.
V
e
r
m
a
,
M
.
K
.
G
u
p
t
a
,
M
.
S
a
l
a
u
d
d
i
n
,
B
.
K
.
R
e
h
m
a
n
,
a
n
d
D
.
K
u
m
a
r
,
“
3
D
m
u
l
t
i
l
a
y
e
r
m
e
s
h
N
o
C
c
o
mm
u
n
i
c
a
t
i
o
n
a
n
d
F
P
G
A
s
y
n
t
h
e
s
i
s
,
”
W
i
r
e
l
e
ss
P
e
rs
o
n
a
l
C
o
m
m
u
n
i
c
a
t
i
o
n
s
,
v
o
l
.
1
0
6
,
n
o
.
4
,
p
p
.
1
8
5
5
–
1
8
7
3
,
J
u
n
.
2
0
1
9
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
1
2
7
7
-
018
-
5724
-
3
.
[
1
6
]
A
.
O
w
e
n
s
,
P
.
I
so
l
a
,
J.
M
c
D
e
r
m
o
t
t
,
A
.
T
o
r
r
a
l
b
a
,
E
.
H
.
A
d
e
l
s
o
n
,
a
n
d
W
.
T.
F
r
e
e
m
a
n
,
“
V
i
su
a
l
l
y
i
n
d
i
c
a
t
e
d
s
o
u
n
d
s
,
”
i
n
2
0
1
6
I
EEE
C
o
n
f
e
re
n
c
e
o
n
C
o
m
p
u
t
e
r
Vi
s
i
o
n
a
n
d
Pa
t
t
e
r
n
Re
c
o
g
n
i
t
i
o
n
(
C
VP
R)
,
J
u
n
.
2
0
1
6
,
p
p
.
2
4
0
5
–
2
4
1
3
,
d
o
i
:
1
0
.
1
1
0
9
/
C
V
P
R
.
2
0
1
6
.
2
6
4
.
[
1
7
]
J.
W
a
n
g
,
Y
.
M
a
,
L
.
Z
h
a
n
g
,
R
.
X
.
G
a
o
,
a
n
d
D
.
W
u
,
“
D
e
e
p
l
e
a
r
n
i
n
g
f
o
r
sm
a
r
t
man
u
f
a
c
t
u
r
i
n
g
:
m
e
t
h
o
d
s
a
n
d
a
p
p
l
i
c
a
t
i
o
n
s
,
”
J
o
u
r
n
a
l
o
f
Ma
n
u
f
a
c
t
u
r
i
n
g
S
y
s
t
e
m
s
,
v
o
l
.
4
8
,
p
p
.
1
4
4
–
1
5
6
,
Ju
l
.
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
j
ms
y
.
2
0
1
8
.
0
1
.
0
0
3
.
[
1
8
]
A
.
G
o
e
l
,
A
.
K
.
G
o
e
l
,
a
n
d
A
.
K
u
mar
,
“
Th
e
r
o
l
e
o
f
a
r
t
i
f
i
c
i
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
a
n
d
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
i
n
u
t
i
l
i
z
i
n
g
s
p
a
t
i
a
l
i
n
f
o
r
ma
t
i
o
n
,
”
S
p
a
t
i
a
l
I
n
f
o
rm
a
t
i
o
n
R
e
se
a
rc
h
,
v
o
l
.
3
1
,
n
o
.
3
,
p
p
.
2
7
5
–
2
8
5
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
0
7
/
s
4
1
3
2
4
-
0
2
2
-
0
0
4
9
4
-
x.
[
1
9
]
P
.
P
.
G
r
o
u
m
p
o
s,
“
D
e
e
p
l
e
a
r
n
i
n
g
v
s.
w
i
se
l
e
a
r
n
i
n
g
:
A
c
r
i
t
i
c
a
l
a
n
d
c
h
a
l
l
e
n
g
i
n
g
o
v
e
r
v
i
e
w
,
”
I
FA
C
-
P
a
p
e
rsO
n
L
i
n
e
,
v
o
l
.
4
9
,
n
o
.
2
9
,
p
p
.
1
8
0
–
1
8
9
,
2
0
1
6
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
i
f
a
c
o
l
.
2
0
1
6
.
1
1
.
0
9
9
.
[
2
0
]
B
.
R
.
K
i
r
a
n
e
t
a
l
.
,
“
D
e
e
p
r
e
i
n
f
o
r
c
e
me
n
t
l
e
a
r
n
i
n
g
f
o
r
a
u
t
o
n
o
m
o
u
s
d
r
i
v
i
n
g
:
A
su
r
v
e
y
,
”
I
EEE
T
r
a
n
s
a
c
t
i
o
n
s
o
n
I
n
t
e
l
l
i
g
e
n
t
T
ra
n
s
p
o
rt
a
t
i
o
n
S
y
st
e
m
s
,
v
o
l
.
2
3
,
n
o
.
6
,
p
p
.
4
9
0
9
–
4
9
2
6
,
J
u
n
.
2
0
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
TI
TS.2
0
2
1
.
3
0
5
4
6
2
5
.
[
2
1
]
A
.
G
i
u
s
t
i
e
t
a
l
.
,
“
A
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
t
o
v
i
s
u
a
l
p
e
r
c
e
p
t
i
o
n
o
f
f
o
r
e
st
t
r
a
i
l
s
f
o
r
m
o
b
i
l
e
r
o
b
o
t
s
,
”
I
E
EE
Ro
b
o
t
i
c
s
a
n
d
Au
t
o
m
a
t
i
o
n
L
e
t
t
e
rs
,
v
o
l
.
1
,
n
o
.
2
,
p
p
.
6
6
1
–
6
6
7
,
J
u
l
.
2
0
1
6
,
d
o
i
:
1
0
.
1
1
0
9
/
L
R
A
.
2
0
1
5
.
2
5
0
9
0
2
4
.
[
2
2
]
R
.
J
.
C
h
e
n
e
t
a
l
.
,
“
D
e
r
mat
o
l
o
g
i
s
t
-
l
e
v
e
l
c
l
a
ssi
f
i
c
a
t
i
o
n
o
f
s
k
i
n
c
a
n
c
e
r
w
i
t
h
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
s,
”
N
a
t
u
r
e
,
v
o
l
.
1
0
,
n
o
.
3
,
p
p
.
1
–
8
,
2
0
1
8
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
13
,
No
.
4
,
Dec
em
b
er
20
24
:
48
5
-
4
9
4
494
[
2
3
]
B
.
K
h
a
g
i
,
C
.
G
.
L
e
e
,
a
n
d
G
.
-
R
.
K
w
o
n
,
“
A
l
z
h
e
i
m
e
r
’
s d
i
s
e
a
s
e
c
l
a
ssi
f
i
c
a
t
i
o
n
f
r
o
m b
r
a
i
n
M
R
I
b
a
se
d
o
n
t
r
a
n
sf
e
r
l
e
a
r
n
i
n
g
f
r
o
m C
N
N
,
”
i
n
2
0
1
8
1
1
t
h
B
i
o
m
e
d
i
c
a
l
E
n
g
i
n
e
e
ri
n
g
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
(
BM
Ei
C
O
N
)
,
N
o
v
.
2
0
1
8
,
p
p
.
1
–
4
,
d
o
i
:
1
0
.
1
1
0
9
/
B
M
E
i
C
O
N
.
2
0
1
8
.
8
6
0
9
9
7
4
.
[
2
4
]
T.
B
e
l
a
b
e
d
,
M
.
G
.
F
.
C
o
u
t
i
n
h
o
,
M
.
A
.
C
.
F
e
r
n
a
n
d
e
s,
C
.
V
.
S
a
k
u
y
a
ma,
a
n
d
C
.
S
o
u
a
n
i
,
“
U
ser d
r
i
v
e
n
F
P
G
A
-
b
a
se
d
d
e
s
i
g
n
a
u
t
o
m
a
t
e
d
f
r
a
mew
o
r
k
o
f
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
s
f
o
r
l
o
w
-
p
o
w
e
r
l
o
w
-
c
o
st
e
d
g
e
c
o
m
p
u
t
i
n
g
,
”
I
EEE
Ac
c
e
ss
,
v
o
l
.
9
,
p
p
.
8
9
1
6
2
–
8
9
1
8
0
,
2
0
2
1
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
1
.
3
0
9
0
1
9
6
.
[
2
5
]
S
.
H
a
r
i
p
r
a
sa
t
h
a
n
d
T
.
N
.
P
r
a
b
a
k
a
r
,
“
F
P
G
A
i
mp
l
e
m
e
n
t
a
t
i
o
n
o
f
m
u
l
t
i
l
a
y
e
r
f
e
e
d
f
o
r
w
a
r
d
n
e
u
r
a
l
n
e
t
w
o
r
k
a
r
c
h
i
t
e
c
t
u
r
e
u
s
i
n
g
V
H
D
L,
”
2
0
1
2
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
C
o
m
p
u
t
i
n
g
,
C
o
m
m
u
n
i
c
a
t
i
o
n
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s,
I
C
C
C
A
2
0
1
2
,
2
0
1
2
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
C
C
A
.
2
0
1
2
.
6
1
7
9
2
2
5
.
[
2
6
]
Y
.
Zh
o
u
,
S
.
R
e
d
k
a
r
,
a
n
d
X
.
H
u
a
n
g
,
“
D
e
e
p
l
e
a
r
n
i
n
g
b
i
n
a
r
y
n
e
u
r
a
l
n
e
t
w
o
r
k
o
n
a
n
F
P
G
A
,
”
i
n
2
0
1
7
I
EEE
6
0
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
Mi
d
w
e
st
S
y
m
p
o
si
u
m
o
n
C
i
rc
u
i
t
s
a
n
d
S
y
s
t
e
m
s (
MW
S
C
A
S
)
,
A
u
g
.
2
0
1
7
,
p
p
.
2
8
1
–
2
8
4
,
d
o
i
:
1
0
.
1
1
0
9
/
M
W
S
C
A
S
.
2
0
1
7
.
8
0
5
2
9
1
5
.
[
2
7
]
A
.
D
u
n
d
a
r
,
J.
J
i
n
,
V
.
G
o
k
h
a
l
e
,
B
.
M
a
r
t
i
n
i
,
a
n
d
E
.
C
u
l
u
r
c
i
e
l
l
o
,
“
M
e
m
o
r
y
a
c
c
e
ss
o
p
t
i
m
i
z
e
d
r
o
u
t
i
n
g
s
c
h
e
me
f
o
r
d
e
e
p
n
e
t
w
o
r
k
s
o
n
a
mo
b
i
l
e
c
o
p
r
o
c
e
ss
o
r
,
”
i
n
2
0
1
4
I
EEE
H
i
g
h
P
e
rf
o
rm
a
n
c
e
E
x
t
r
e
m
e
C
o
m
p
u
t
i
n
g
C
o
n
f
e
re
n
c
e
(
H
PE
C
)
,
S
e
p
.
2
0
1
4
,
p
p
.
1
–
6
,
d
o
i
:
1
0
.
1
1
0
9
/
H
P
EC
.
2
0
1
4
.
7
0
4
0
9
6
3
.
[
2
8
]
M
.
P
e
e
m
e
n
,
A
.
A
.
A
.
S
e
t
i
o
,
B
.
M
e
s
m
a
n
,
a
n
d
H
.
C
o
r
p
o
r
a
a
l
,
“
M
e
m
o
r
y
-
c
e
n
t
r
i
c
a
c
c
e
l
e
r
a
t
o
r
d
e
s
i
g
n
f
o
r
C
o
n
v
o
l
u
t
i
o
n
a
l
N
e
u
r
a
l
N
e
t
w
o
r
k
s
,
”
i
n
2
0
1
3
I
E
E
E
3
1
s
t
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
C
o
m
p
u
t
e
r
D
e
s
i
g
n
(
I
C
C
D
)
,
O
c
t
.
2
0
1
3
,
p
p
.
1
3
–
1
9
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
C
D
.
2
0
1
3
.
6
6
5
7
0
1
9
.
[
2
9
]
Y
.
Y
.
G
h
a
d
i
e
t
a
l
.
,
“
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
so
l
u
t
i
o
n
s
f
o
r
t
h
e
sec
u
r
i
t
y
o
f
w
i
r
e
l
e
ss
sen
s
o
r
n
e
t
w
o
r
k
s:
a
r
e
v
i
e
w
,
”
I
EE
E
A
c
c
e
ss
,
v
o
l
.
1
2
,
p
p
.
1
2
6
9
9
–
1
2
7
1
9
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
4
.
3
3
5
5
3
1
2
.
[
3
0
]
Z.
C
h
e
n
,
J.
W
a
n
g
,
H
.
H
e
,
a
n
d
X
.
H
u
a
n
g
,
“
A
f
a
s
t
d
e
e
p
l
e
a
r
n
i
n
g
s
y
st
e
m
u
s
i
n
g
G
P
U
,
”
i
n
2
0
1
4
I
EE
E
I
n
t
e
r
n
a
t
i
o
n
a
l
S
y
m
p
o
si
u
m
o
n
C
i
rc
u
i
t
s
a
n
d
S
y
st
e
m
s (I
S
C
A
S
)
,
J
u
n
.
2
0
1
4
,
p
p
.
1
5
5
2
–
1
5
5
5
,
d
o
i
:
1
0
.
1
1
0
9
/
I
S
C
A
S
.
2
0
1
4
.
6
8
6
5
4
4
4
.
[
3
1
]
Y
.
Zh
a
n
g
a
n
d
S
.
Z
h
a
n
g
,
“
O
p
t
i
m
i
z
e
d
d
e
e
p
l
e
a
r
n
i
n
g
a
r
c
h
i
t
e
c
t
u
r
e
s
w
i
t
h
f
a
st
mat
r
i
x
o
p
e
r
a
t
i
o
n
k
e
r
n
e
l
s
o
n
p
a
r
a
l
l
e
l
p
l
a
t
f
o
r
m,
”
i
n
2
0
1
3
I
EEE
2
5
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
T
o
o
l
s
w
i
t
h
Art
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
,
N
o
v
.
2
0
1
3
,
p
p
.
7
1
–
7
8
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
TA
I
.
2
0
1
3
.
2
1
.
[
3
2
]
R
.
Z
h
a
o
e
t
a
l
.
,
“
H
a
r
d
w
a
r
e
c
o
m
p
i
l
a
t
i
o
n
o
f
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
s
:
a
n
o
v
e
r
v
i
e
w
,
”
i
n
2
0
1
8
I
EE
E
2
9
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
Ap
p
l
i
c
a
t
i
o
n
-
sp
e
c
i
f
i
c
S
y
st
e
m
s
,
Ar
c
h
i
t
e
c
t
u
res
a
n
d
Pr
o
c
e
sso
rs
(
AS
AP)
,
Ju
l
.
2
0
1
8
,
p
p
.
1
–
8
,
d
o
i
:
1
0
.
1
1
0
9
/
A
S
A
P
.
2
0
1
8
.
8
4
4
5
0
8
8
.
[
3
3
]
J.
K
i
m,
K
.
H
w
a
n
g
,
a
n
d
W
.
S
u
n
g
,
“
X
1
0
0
0
r
e
a
l
-
t
i
m
e
p
h
o
n
e
me
r
e
c
o
g
n
i
t
i
o
n
V
LSI
u
si
n
g
f
e
e
d
-
f
o
r
w
a
r
d
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
s,
”
i
n
2
0
1
4
I
EE
E
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
Ac
o
u
st
i
c
s
,
S
p
e
e
c
h
a
n
d
S
i
g
n
a
l
P
ro
c
e
ssi
n
g
(
I
C
A
S
S
P)
,
M
a
y
2
0
1
4
,
p
p
.
7
5
1
0
–
7
5
1
4
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
A
S
S
P
.
2
0
1
4
.
6
8
5
5
0
6
0
.
[
3
4
]
K
.
H
w
a
n
g
a
n
d
W
.
S
u
n
g
,
“
F
i
x
e
d
-
p
o
i
n
t
f
e
e
d
f
o
r
w
a
r
d
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
d
e
s
i
g
n
u
si
n
g
w
e
i
g
h
t
s
+
1
,
0
,
a
n
d
−
1
,
”
i
n
2
0
1
4
I
EE
E
Wo
r
k
sh
o
p
o
n
S
i
g
n
a
l
Pr
o
c
e
ss
i
n
g
S
y
st
e
m
s (S
i
P
S
)
,
O
c
t
.
2
0
1
4
,
p
p
.
1
–
6
,
d
o
i
:
1
0
.
1
1
0
9
/
S
i
P
S
.
2
0
1
4
.
6
9
8
6
0
8
2
.
[
3
5
]
S
.
A
n
w
a
r
,
K
.
H
w
a
n
g
,
a
n
d
W
.
S
u
n
g
,
“
F
i
x
e
d
p
o
i
n
t
o
p
t
i
mi
z
a
t
i
o
n
o
f
d
e
e
p
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s
f
o
r
o
b
j
e
c
t
r
e
c
o
g
n
i
t
i
o
n
,
”
i
n
2
0
1
5
I
EEE
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
Ac
o
u
st
i
c
s,
S
p
e
e
c
h
a
n
d
S
i
g
n
a
l
P
ro
c
e
ssi
n
g
(
I
C
A
S
S
P)
,
A
p
r
.
2
0
1
5
,
p
p
.
1
1
3
1
–
1
1
3
5
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
A
S
S
P
.
2
0
1
5
.
7
1
7
8
1
4
6
.
[
3
6
]
A
.
G
o
e
l
,
A
.
K
.
G
o
e
l
,
a
n
d
A
.
K
u
m
a
r
,
“
P
e
r
f
o
r
ma
n
c
e
a
n
a
l
y
s
i
s
o
f
m
u
l
t
i
p
l
e
i
n
p
u
t
si
n
g
l
e
l
a
y
e
r
n
e
u
r
a
l
n
e
t
w
o
r
k
h
a
r
d
w
a
r
e
c
h
i
p
,
”
Mu
l
t
i
m
e
d
i
a
T
o
o
l
s
a
n
d
Ap
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
8
2
,
n
o
.
1
8
,
p
p
.
2
8
2
1
3
–
2
8
2
3
4
,
Ju
l
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
0
7
/
s1
1
0
4
2
-
0
2
3
-
1
4
6
2
7
-
3.
[
3
7
]
A
.
D
e
v
r
a
r
i
a
n
d
A
.
K
u
mar,
“
T
u
r
b
o
e
n
c
o
d
e
r
a
n
d
d
e
c
o
d
e
r
c
h
i
p
d
e
s
i
g
n
a
n
d
F
P
G
A
d
e
v
i
c
e
a
n
a
l
y
si
s
f
o
r
c
o
mm
u
n
i
c
a
t
i
o
n
sy
s
t
e
m
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
Re
c
o
n
f
i
g
u
ra
b
l
e
a
n
d
Em
b
e
d
d
e
d
S
y
st
e
m
s
(
I
J
R
ES
)
,
v
o
l
.
1
2
,
n
o
.
2
,
p
p
.
1
7
4
–
1
8
5
,
Ju
l
.
2
0
2
3
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
r
e
s.
v
1
2
.
i
2
.
p
p
1
7
4
-
1
8
5
.
[
3
8
]
A
.
S
.
R
a
w
a
t
,
A
.
R
a
n
a
,
A
.
K
u
m
a
r
,
a
n
d
A
.
B
a
g
w
a
r
i
,
“
A
p
p
l
i
c
a
t
i
o
n
o
f
mu
l
t
i
l
a
y
e
r
a
r
t
i
f
i
c
i
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
i
n
t
h
e
d
i
a
g
n
o
s
i
s
s
y
st
e
m:
a
sy
st
e
ma
t
i
c
r
e
v
i
e
w
,
”
I
AE
S
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
Art
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
(
I
J
-
AI
)
,
v
o
l
.
7
,
n
o
.
3
,
p
p
.
1
3
8
–
1
4
2
,
A
u
g
.
2
0
1
8
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
a
i
.
v
7
.
i
3
.
p
p
1
3
8
-
1
4
2
.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
Ar
u
n
a
Pa
n
t
is
wo
rk
i
n
g
a
s
a
ju
n
i
o
r
re
se
a
rc
h
fe
ll
o
w
(JRF)
i
n
t
h
e
De
p
a
rtme
n
t
o
f
El
e
c
tri
c
a
l
a
n
d
El
e
c
tro
n
ics
En
g
i
n
e
e
rin
g
,
UPE
S
,
De
h
ra
d
u
n
,
I
n
d
i
a
.
S
h
e
h
a
s
a
n
M
.
Tec
h
.
in
El
e
c
tro
n
ics
a
n
d
C
o
m
m
u
n
ica
ti
o
n
En
g
i
n
e
e
rin
g
fro
m
Am
it
y
Un
iv
e
rsity
,
L
u
c
k
n
o
w,
In
d
ia.
He
r
a
re
a
o
f
in
tere
st
is
b
io
m
e
d
ica
l
si
g
n
a
l
p
r
o
c
e
ss
in
g
,
EE
G
sig
n
a
ls,
a
n
d
d
i
g
it
a
l
sig
n
a
l
p
r
o
c
e
ss
in
g
.
S
h
e
is
h
a
v
in
g
a
n
e
x
p
e
rien
c
e
o
f
1
2
y
e
a
rs
o
f
tea
c
h
i
n
g
.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
a
ru
n
a
p
a
n
t
4
1
@
g
m
a
il
.
c
o
m
.
Ades
h
K
u
m
a
r
h
a
s
b
e
e
n
wo
r
k
in
g
a
s
a
p
ro
fe
ss
o
r
i
n
t
h
e
De
p
a
rtme
n
t
o
f
El
e
c
tri
c
a
l
a
n
d
El
e
c
tro
n
ics
E
n
g
i
n
e
e
rin
g
,
S
c
h
o
o
l
o
f
A
d
v
a
n
c
e
d
En
g
in
e
e
rin
g
,
T
h
e
Un
iv
e
rsit
y
o
f
P
e
tr
o
leu
m
a
n
d
E
n
e
rg
y
S
tu
d
ies
(UPE
S
)
,
De
h
ra
d
u
n
,
I
n
d
ia
si
n
c
e
2
0
1
0
.
He
h
a
s
a
B.
Tec
h
.
in
El
e
c
tro
n
ics
a
n
d
Co
m
m
u
n
ica
ti
o
n
E
n
g
i
n
e
e
rin
g
fr
o
m
UPT
U,
L
u
c
k
n
o
w
I
n
d
ia
in
2
0
0
6
.
M
.
Tec
h
.
(Ho
n
s)
i
n
Emb
e
d
d
e
d
S
y
ste
m
s
Tec
h
n
o
lo
g
y
,
fro
m
S
RM
U
n
iv
e
rsit
y
,
Ch
e
n
n
a
i
in
2
0
0
8
.
P
h
.
D.
(El
e
c
tro
n
ics
En
g
i
n
e
e
rin
g
)
fro
m
UPE
S
,
De
h
ra
d
u
n
I
n
d
ia i
n
2
0
1
4
.
He
h
a
s a
lso
wo
r
k
e
d
a
s a
S
e
n
i
o
r
E
n
g
i
n
e
e
r
i
n
TATA
EL
XSI
LI
M
ITE
D
Ba
n
g
a
l
o
re
a
n
d
a
s
a
fa
c
u
lt
y
m
e
m
b
e
r
in
I
CF
AI
Un
iv
e
rsit
y
,
De
h
ra
d
u
n
.
His
a
re
a
s
o
f
in
tere
st
a
re
VLS
I
d
e
sig
n
,
e
m
b
e
d
d
e
d
sy
ste
m
s
d
e
sig
n
,
tele
c
o
m
m
u
n
ica
ti
o
n
s,
a
n
d
sig
n
a
l
p
r
o
c
e
ss
in
g
.
He
h
a
s
su
p
e
r
v
ise
d
8
P
h
.
D.
sc
h
o
lars
,
a
n
d
5
c
a
n
d
id
a
tes
a
re
d
o
i
n
g
re
se
a
rc
h
u
n
d
e
r
h
is
su
p
e
rv
isi
o
n
.
He
h
a
s
wo
rk
e
d
o
n
m
o
re
th
a
n
1
0
e
d
it
o
rial
a
ss
ig
n
m
e
n
ts
fo
r
e
d
it
e
d
b
o
o
k
s
,
c
o
n
fe
re
n
c
e
p
r
o
c
e
e
d
in
g
s
,
a
n
d
jo
u
rn
a
ls,
h
a
v
in
g
c
it
a
ti
o
n
s
o
f
m
o
re
t
h
a
n
2
0
0
0
,
H
-
i
n
d
e
x
-
2
5
i
1
0
-
in
d
e
x
-
4
5
.
He
h
a
s
1
6
y
e
a
rs
o
f
e
x
p
e
rien
c
e
in
th
e
tea
c
h
in
g
+
re
se
a
rc
h
in
d
u
str
y
,
p
u
b
li
s
h
e
d
m
o
re
th
a
n
1
2
0
+
re
se
a
rc
h
p
a
p
e
rs
in
in
tern
a
ti
o
n
a
l
p
e
e
r
-
re
v
ie
we
d
j
o
u
r
n
a
ls
(S
CI/S
c
o
p
u
s)
a
n
d
c
o
n
fe
re
n
c
e
s.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
a
d
e
sh
m
a
n
a
v
@g
m
a
il
.
c
o
m
o
r
a
d
e
sh
k
u
m
a
r@d
d
n
.
u
p
e
s.a
c
.
in
.
Evaluation Warning : The document was created with Spire.PDF for Python.