TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol.12, No.7, July 201
4, pp
. 5655 ~ 56
6
0
DOI: 10.115
9
1
/telkomni
ka.
v
12i7.572
6
5655
Re
cei
v
ed
No
vem
ber 2
1
, 2013; Re
vi
sed
Febr
uary 18,
2014; Accept
ed March 6, 2
014
Sketching Expert System for Crime Investig
ation
Purposes
Made Bagus
Yudistira*
1
, I Ketut Gede Darma Putra
2
, Anak Agu
ng Komp
y
a
ng Oka Sudan
a
3
Dep
a
rtment of Information T
e
chno
log
y
,
Ud
ay
an
a
Un
i
v
e
r
si
ty
, Ba
l
i
,
In
do
nesi
a
Bukit Jimbar
an
, Badung, Ba
li, Indon
esi
a
,
T
e
lp. 0361-
78
535
33
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: bagus
_
y
u
d
ist
i
ra@
y
a
h
o
o
.co
m
1
, duglair
e
@
y
a
h
o
o
.com
2
,
agu
ng
okas@
h
otmail.com
3
A
b
st
r
a
ct
T
he pres
ence
of polic
e sketc
her pl
ay a
n
i
m
p
o
rtant
ro
le i
n
makin
g
inv
e
stigatio
n in
pu
rpose
of
mak
i
n
g
arr
e
stme
nt to fu
gitiv
e
or s
u
sp
ect. T
he l
a
ckin
g
pr
esenc
e of
po
li
ce sketch
e
r is
makin
g
a l
a
c
k
i
n
investi
gatio
n p
r
ocess, bec
au
se lack of i
n
formatio
n
gat
h
e
red for th
e further pr
ocess
.
T
h
is limitatio
n
i
s
overcome by
developi
ng
an expert system
using
gadget as
a helpin
g device to
m
a
k
i
ng
s
k
etch, with adding
sketcher kn
ow
ledg
e. Sketchin
g metho
d
alre
a
d
y bee
n us
ed
since l
o
n
g
time
in proc
ess of i
n
vestig
atio
n an
d
effective
m
a
k
i
ng the res
u
lt. The r
e
sults
of expert syst
em
on the cas
e
hav
e
gi
ven showing the system
to
real o
b
ject w
h
i
c
h ma
de sketc
hin
g
reach
85
% of accuracy l
e
vel.
Ke
y
w
ord:
expert system
, des
ktop based, sketch, crim
e
Copy
right
©
2014 In
stitu
t
e o
f
Ad
van
ced
En
g
i
n
eerin
g and
Scien
ce. All
rig
h
t
s reser
ve
d
.
1. Introduc
tion
Crim
e in th
e
21st
centu
r
y
have in
crea
sed an
d
d
e
vel
oped.
Crimi
n
al rate
as time goe
s
by,
getting more
arises
and ye
t the crimin
al is still c
ann
ot be put in jail.
The presen
ce of witne
s
se
s
arou
nd th
e
scene i
s
ve
ry h
e
lpful in
solving
crimi
nal
case
s
by providing i
n
form
ation rega
rdi
ng
the
events ta
ke
p
l
ace, a
nd th
e
peopl
e who
were invo
lve
d
aroun
d the
scene
of
the crim
e
that wo
uld
be a co
ntribut
ing factor
sol
v
es a ca
se.
Furthe
r o
b
st
acle
s fa
ce
d
by investig
ators is
to d
e
scrib
e
the
faces
of pe
ople
involved
based on a
cha
r
a
c
teri
stic describ
ed b
y
witnesse
s to media im
age
s that ca
n later help
the
investigatio
n. Lackin
g
of hu
man re
sou
r
ce
s a
s
a
poli
c
e
sketch
er is al
so
a fla
w
of the inve
stigati
on
p
r
oc
es
s
.
Face
reco
gni
tion one
of t
he p
r
ima
r
y b
i
ometri
c tech
nologi
es
be
came mo
re
importa
nt
owin
g to
rapi
d adva
n
ces i
n
technol
ogie
s
su
ch
as
dig
i
tal cam
e
ras,
Internet
and
mobile
devi
c
es
and in
cre
a
se
d deman
ds
on se
cu
rity.
Face
re
cog
n
i
t
ion has
sev
e
ral adva
n
ta
ges ove
r
oth
e
r
biometri
c te
chnolo
g
ies, it i
s
natu
r
al, no
n
intrusiv
e
and
easy to u
s
e.
But face reco
gnition is
one
o
f
the challe
ngi
ng pro
b
lem
s
in research, till now there i
s
no unique so
lution for all face recogniti
on
appli
c
ation
s
[1-2]. The
wide range
of
variation
s
i
n
hum
an fa
ce d
ue to vi
ew p
o
int, po
se,
illumination a
nd expre
s
sio
n
deterio
rate
the re
co
gnit
i
on perfo
rma
n
ce of the F
a
ce recogniti
on
system
s. But
everyon
e
a
c
cept that t
he face
recognition
syst
em is
goo
d, if it has le
ss
comp
utationa
l complexity, good recognit
i
on perfo
rma
n
ce a
nd o
c
cu
pies le
ss me
mory.
Biometric re
cognition syst
em
is
a
syst
em that uses the uni
qu
e chara
c
te
risti
c
s of each
individual. Th
is sy
stem i
s
more
reli
able
than
the to
ken in
clu
s
ion
and
re
cogniti
on of
kno
w
le
dge.
Each individ
ual ha
s a d
i
fferent physi
ologi
cal
characteri
stic a
n
d
behavio
ral
characte
ri
stic.
Physiologi
cal
characte
ri
stic is rel
a
tively stable p
h
ysi
c
al characteri
stic like
a fingerprint, iris
, fac
e
,
and
ha
nd
g
e
o
m
etry.
While behavio
ral ch
ara
c
teri
stics such as
voi
c
e and signatu
r
e
are
influ
e
n
c
e
d
by the psych
o
logi
cal co
nd
ition that easily
chang
ed. A lot of deve
l
opers have
develop
ed th
e
recognitio
n
system ba
sed
on physi
cal o
r
physiol
ogi
ca
l characte
ri
stics [3
-6].
In the develo
p
ment p
r
oce
ss
of data in
vestigat
ion, t
he process o
f
information
retrieval
from witne
s
s is very import
ant. One of those info
rm
ation is of the face
s of wh
oe
ver involved or
happ
en to
b
e
at the
crim
e sce
ne,
whi
c
h i
n
may
b
e
a
witne
s
s
or a
suspe
c
t. An a
rre
stme
nt
pro
c
e
s
s woul
d be
helpe
d i
f
the de
script
ion ma
de by
witne
ss i
s
ea
sily re
co
gni
zable o
r
clo
s
e
l
y
lo
o
k
lik
e th
e
s
u
s
p
ec
te
d p
e
r
s
on
. T
h
e
pr
oc
es
s o
f
ma
ki
ng
sketch
es
of face
s, req
u
ire th
e fa
cto
r
s
of
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 7, July 201
4: 5655 – 56
60
5656
the sketche
r
, whi
c
h in
clu
d
e
the expe
rti
s
e of
a
sket
cher, the
avail
ability of time, and al
so
the
emotional
sta
t
e of the artist. In
addition, it is
also nece
s
sary skills of sketch
e
r to transl
a
te the
descri
p
tion gi
ven by witne
ss to
media i
m
age
s [7
-8]. The rea
s
on
a
bove is
why t
he availabl
e
of
police sketch
er is imp
o
rtan
t, and by lacki
ng of it,
will make inve
stigat
ion pro
c
e
s
s b
e
incom
p
lete.
By applying simulatio
n
ap
plicatio
n as
a medi
a to p
r
ocessin
g
face sketch , then the
process will be
completed
faster,
so that will help
the whol
e proc
ess of investigation. Expert
system i
s
on
e outgrowth
of technol
ogy
develope
d
with the aim to mimic the ab
ility of an expert
in a particular field [9].
Re
sea
r
che
s
related to
exp
e
rt sy
stem
s h
a
ve bee
n do
n
e
with
seve
ra
l re
sea
r
ch o
b
jects as
follows. Max Isch
en
ko dev
elop we
b ba
sed sketchi
ng
appli
c
ation. By that applica
t
ion use
r
can
make fa
ce
s sketch usi
ng template
s fro
m
the appli
c
a
t
ion. The appl
ication u
s
e
d
to use
r
to abl
e to
make
face sketch a
s
if b
een d
one
by polic
e sket
cher. Th
e ap
plicatio
n ha
s been
used
as
entertain
ment
purpo
se a
nd
not been u
s
e
d
for investig
ation [10].
IQ Beometrix
develop m
o
del of an
expert sy
st
em f
o
r ma
kin
g
fa
ce
sketch
ba
sed
on
model
cre
a
te
d by Max Ischen
ko ap
plication. T
he m
odel is
co
nst
r
ucted
co
nsi
s
t
ed of win
d
o
w
ed
appli
c
ation
with each wi
n
dow
co
nsi
s
t
of templates.
The mo
del
wa
s teste
d
a
nd no
w
curre
n
tly
use
d
in mostl
y
Law Enforcer in Unite
d
Stat
es, and already maki
ng
seve
ral arre
stments [11].
In sp
ring
of
2
003, the
Fresno
Cou
n
ty Sh
eriff’
s
Dep
a
rt
ment fa
ced
di
fficult investig
ation of
rape
involvin
g p
r
o
s
titutes
until on
e of
the victim
s
a
g
ree
d
to
work
with th
e d
e
t
ective. Usin
g a
expert syste
m
for maki
ng sketch,
th
e compo
s
ite sk
e
t
ch
wa
s di
stri
buted to
Police patrol u
n
it, and
within two da
ys a man wa
s app
reh
end
ed and charg
ed with seve
ral count
s of kidn
ap and rape
[11]
A
29-y
e
a
r
-
o
l
d
su
sp
ect
w
a
s a
r
r
e
st
e
d
by
B
r
oward
Cou
n
ty, Flori
da poli
c
e
sh
ortly after
America’s M
o
st Wante
d
T
V
sho
w
aire
d
a face
sket
ch made by e
x
pert system.
Prior to the
Oct.
1998
AMW b
r
oad
ca
st, several
sketche
s
of the
su
spe
c
t ha
d be
en
hand
-d
ra
wn,
but provided
few
clue
s fo
r poli
c
e. Deputy John M
c
Ma
ho
n of the Brow
ard
Cou
n
ty Sheriff’s
Office
, working
cl
o
s
ely
with youn
g vi
ctims, utili
zed
expert
syste
m
poli
c
e sket
ch gene
rato
r to
com
p
o
s
e a
pictu
r
e quality
comp
osite
of
the suspe
c
t. Soon afte
r th
e sket
ch
app
eare
d
o
n
Am
erica’s Mo
st
Wante
d
, a
wo
man
conta
c
ted
pol
ice, saying th
e po
rtrait wa
s that of
h
e
r
son.
A
su
spe
c
t
wa
s a
r
r
e
st
ed wit
h
i
n
ho
u
r
s,
and cha
r
ge
d with
sexually assaultin
g
an
11-yea
r
-o
ld
girl an
d attem
p
ting to ab
du
ct 10 oth
e
r
gi
rls
in Florid
a [11].
2. Rese
arch
Metho
d
System is a
impleme
n
tation of both a
pplication tha
t
had been
created
before usin
g
buttons,
as
a pa
ramete
r for
swit
chin
g obje
c
t int
e
rface, a
wi
ndo
wed fig
u
re to ma
ke
the
appli
c
ation u
s
er f
r
iendly,
and templ
a
te
s for ma
kin
g
the sketch. T
he data u
s
e
d
for maki
ng the
sketch
templ
a
te a
r
e
sket
ch
of p
a
rts i
n
he
ad
gain
ed
from kn
o
w
led
ge of
a
sketch
er
in *
.
png
extensio
n.
2.1. Kno
w
l
e
d
g
e Acq
u
isition
The kno
w
led
ge a
c
qui
red f
r
om the lite
r
ature a
nd
sin
g
le expe
rt, it is all in
cludi
ng hai
r,
head, eye
s
, n
o
se, m
outh, j
a
w, a
nd oth
e
r facial fi
g
u
re
su
ch as bea
rd,
musta
c
h
e
, freckle,
a
nd a
g
e
mark. All of th
ose pl
aced in
a windo
we
d form an
d se
pa
rated ba
se
d in categ
o
ry.
2.2. Kno
w
l
e
d
g
e Rep
r
ese
n
t
ation
Knowle
dge of
all sket
c
hin
g
method fro
m
the ex
pert re
pre
s
ente
d
in
a sketch fo
rm
divided
by its cate
go
ry. Each of categori
e
s
will b
e
pres
ented i
n
win
d
o
w
ed f
o
rm, an
d ea
ch win
d
o
w
will
be
showed according to button ch
osen represented on i
n
terface,
so all
of
the buttons will
representing
the categori
e
s. All of
categories will
be
placed in l
a
yer according t
o
the placem
ent,
whi
c
h mea
n
the full output will be combi
nation of a
ll sketch of each
catego
rie
s
b
e
in one sket
ch
in stack. Usi
ng Java p
r
o
g
rammi
ng to
modeling th
e appli
c
ation
,
and will di
vide all of the
templates to
each layers [12-1
3
]. The model of kn
o
w
led
ge re
pre
s
entatio
n are
sho
w
n b
e
low.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Sketchi
ng Expert System
for Crim
e Inve
st
igation Pu
rp
ose
s
(M
ade
Bagus Yu
disti
r
a)
5657
Figure 1. Win
dow a
nd Tem
p
lates b
ehin
d
Main Fram
e
Each
of the
windowed form w
ill contain
a category
represen
ting
on its figure,
and each
category will
consi
s
ting some of sket
c
h represent
ing the sket
c
h that will be drawn in the m
a
in
canvas. Each of category w
ill be shown as figure bel
o
w.
Figure 2. Obj
e
ct of each Category
3. Result a
n
d Discus
s
io
n
The p
u
rp
ose
of testing
these ap
plica
t
i
ons i
s
to
determi
ne th
e effectivene
ss and
perfo
rman
ce
of appli
c
ation
s
that h
a
ve b
een m
ade.
T
h
is te
st will
b
e
able to
pro
v
ide a
con
c
lu
sion
on ho
w effe
ctive the meth
od that allo
ws to
so
lve th
e problem
s t
hat exist an
d
how
well th
e
perfo
rman
ce i
m
pleme
n
ted.
3.1. Sy
stem
Platform
Expert sy
ste
m
develo
ped
desktop
ba
se
d pla
tform
u
s
i
ng some
software su
ch as
Adobe
Photosh
op to
build
the
kn
owle
dge
ba
se, Java
Scri
pt, Adobe
Fl
ash
Profe
s
si
onal to
buil
d
the
appli
c
ation a
nd de
sign int
e
rface.
3.2. Sy
stem
Structure
3.2.1. Kno
w
l
e
dge Acquis
i
tion
The acqui
sition of kno
w
le
dge is the de
vel
opment en
vironme
n
t used by the knowle
dge
engin
eer
to acq
u
ire
th
e kno
w
le
dge o
f
single exp
e
rt a
s
the
source. In thi
s
expe
rt sy
ste
m
developm
ent, kno
w
le
dge
acq
u
isitio
n is done th
ro
u
gh inte
rviews with
poli
c
e
sketch
er, a
n
d
sup
porte
d by literature stud
ies.
3.2.2. Kno
w
l
e
dge Base
Knowle
dge b
a
se is a dev
elopme
n
t en
vironme
n
t used by the kn
owle
dge en
gi
neer to
rep
r
e
s
ent the
kno
w
led
ge t
hat gaine
d from the a
c
qui
sition of kno
w
led
ge. The f
a
ct are consi
s
ting
the form of p
h
ysical chara
c
teri
stic of
he
ad, eyes, h
a
ir, nose j
a
w, e
y
ebro
w
an
d o
t
her detail
su
ch
as m
o
le, fre
c
kle, a
nd
scar mark in
sket
ch. Th
e
rule
s are
mad
e
b
y
combi
n
ing
the fa
cts
abo
ve.
Knowle
dge b
a
se b
u
ilt in the form of so
me wind
ow
di
splaye
d, su
ch
as wind
ow of
hair, wind
ow of
head, wi
ndo
w of eyes, wind
ow of no
se a
nd others.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 7, July 201
4: 5655 – 56
60
5658
3.2.3. User I
n
ter
f
ac
e
The interfa
c
e
is an environ
mental con
s
u
l
tancy
that
is intende
d
for use
r
s
to cho
o
se
from
option avail
a
ble in th
e exp
e
rt sy
stem
which
one
the
most
suitabl
e
by the refere
nce. T
he
option
will be shown in form of wi
ndow, whi
c
h
each wind
ow will be showi
ng figure
of each
category.
In
each window
will be showing figure that
resembl
a
nce
of sket
c
h whi
c
h w
ill be chosen by user.
3.2.4. Interfa
ce Engine
The inferen
c
e
engine
use
s
wind
owed fig
u
re to cre
a
te the interfa
c
e
whi
c
h can be
cho
s
e
n
by user, an
d
the result of the chosen
option in
oth
e
r win
d
o
w
place
d
side by
side ea
ch ot
her.
Source of tha
t
all figure ha
ve been e
s
ta
blish
ed on th
e basi
s
of kn
owle
dge ab
o
v
e.
3.2.5. Workp
l
ace
Wo
rkpl
ace
re
pre
s
ente
d
in
t
he fo
rm
of 3
parts,
which i
s
b
u
ttons,
sketch te
mplate
win
d
o
w
,
and re
sult wi
ndo
w.
3.2.6. Explanation Facilit
y
The explan
ation facility pro
v
ided in the wind
owed form, first, the system will provide al
l
the
category
in buttons, whi
c
h each button
will showi
ng a wi
ndow re
presenting a category;
second, the window
appeared
will be containing figures
whi
c
h
will
chosen by user to
show at the
result, ea
ch f
i
gure
re
prese
n
ting a
sketch sa
me to
th
e figure itself
; third, all th
e chosen fig
u
re
from all cate
gory will b
e
sho
w
n o
n
a
result win
d
o
w
, and it ca
n be chang
e
d
all over. T
h
e
example
s
for
this facility are sho
w
n at Fi
gure 3, Fig
u
re 4 and Figu
re 5
Figure 3. Use
r
Interface
Figure 4. Win
dowed Figu
re
Representin
g each Sketch
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Sketchi
ng Expert System
for Crim
e Inve
st
igation Pu
rp
ose
s
(M
ade
Bagus Yu
disti
r
a)
5659
Figure 5. Ske
t
ch Re
sult Co
mpared to Re
al Person
3.2.7. Kno
w
l
e
dge Improv
ement
Improveme
n
t of knowl
edg
e can b
e
do
ne if
there are addition
s o
r
ch
ange
s to
the new
categ
o
ry, the
chan
ge a
nd
addition va
ria
t
ion, or chan
ge of metho
d
and theo
ry to do sket
chin
g
face
s in
term
of u
s
e
for
p
o
lice
search.
Based
on
a
n
y
addition
s
or ch
ang
es, th
e sy
stem
will
do
the cre
a
tion o
f
new rule
s of
the sket
c
hin
g
and that will
be gene
rate
d.
3.3. Sy
stem
Performan
c
e
Develo
ped
a
n
expe
rt
syst
em te
sting
p
e
rfor
m
e
d
by makin
g
a sketch usi
ng
th
e
expe
rt
system
ba
se
d on
re
al face, then give
n
to re
sp
on
d
e
n
t to gue
st
which
one
of
many faces i
n
on
e
picture is the
face m
ade
by expert
syste
m
. Table 4
sh
ows that sy
stem pe
rf
orm
a
nce
as th
e re
sult
of the compa
r
ison.
Figure
Amunt
of
Audien
ce Correct
Wrong
Accuracy
(%
)
1 100
87
14
87
2 100
83
17
83
3 100
85
15
85
Average of the di
fference result
85
4. Conclusio
n
Expert sy
ste
m
for m
a
ki
n
g
face sket
ch in p
u
rpo
s
e
of poli
c
e i
n
vestigation
h
a
s
been
develop
ed on
desktop ba
sed platform to
receive in
put
in the form of characte
ri
stic given by user.
The input
s are physical ch
ara
c
teri
stic v
a
lue ba
s
ed o
n
real phy
sical cha
r
a
c
teri
stic from peop
le
who
will be sketched. The
system p
r
ovi
des o
u
tput in a form of sk
e
t
ch ba
sed o
n
output given
by
use
r
the
p
r
o
c
e
s
s can
be
don
e repe
atedly until
re
a
c
hin
g
the
ou
tput wa
nted.
System te
sting
results
sho
w
that the syste
m
develope
d has t
he
simila
rity with the real pictu
r
e at 85%.
REFERE
NC
ES
[1]
A Jain,
R Bo
ll
e
,
S Pank
anti E
d
s. BIOMET
RIC
- Pers
ona
l I
dentific
atio
n i
n
Net
w
ork
ed
So
ciet
y
. K
l
uw
er
Acade
mic Publ
ishers, Boston/
Dordrec
h
t/Lon
don
. 19
99.
[2]
D Sridhar, Dr IV Murali
Kris
hn
a. Face
Reco
g
n
itio
n us
ing
T
c
heb
ichef M
o
m
ents.
Internati
o
nal J
our
nal
o
f
Information &
Network Security (IJINS).
201
2; 1(4): 243-2
5
3
.
[3]
Agus
A
n
w
a
r,
Darma Putra, Agun
g
C
a
h
y
a
w
a
n
.
P
a
lmpr
in
t Verificati
on
Usin
g T
i
me S
e
ries
Method.
T
E
LKOMNIKA Jurna
l
of Electrical En
gin
eeri
n
g
. 2013; 1
1
(4): 749-
758.
[4]
I Ketut Gede
Darma Putra,
Erdia
w
a
n
. H
i
g
h
Perfo
rma
nce
Palmpri
n
t Ide
n
tificatio
n
S
y
stem Base
d On
T
w
o D
i
mens
io
nal Gab
o
r.
T
E
LKOMNIKA Journal of Electric
al Eng
i
ne
eri
n
g
.
2010; 8(3): 3
0
9
-31
8
.
[5]
I Ketut Darm
a
Putra, W
i
ra
Bhua
na, Er
dia
w
a
n
. P
e
mbe
n
t
u
kan
Ko
de T
e
lap
a
k T
anga
n
(Palm C
o
d
e
)
Berbas
is Meto
de Gabor 2
D
. MAKARA Jour
nal of
T
e
chnol
og
y
Seri
es. 20
11; 15(2): 1
61-
167.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 7, July 201
4: 5655 – 56
60
5660
[6] Darma
Putra.
Sistem Ver
i
fika
si Me
ngg
un
ak
an Gar
i
s-gar
is
T
e
lap
a
k T
a
nga
n.
T
e
knolo
g
i
el
ektro. 20
07
;
6(2): 46-5
1
.
[7] Douglas
Harper.
Online Etymolo
g
y Dictio
nar
y – Sketch
. Access on Dec
e
mber 20
13
[8]
Diana Davies (editor).
Harr
ap'
s Illustrated D
i
c
tionary
of Art and Artists
, Ha
rrap Bo
oks Li
mited. 19
90.
ISBN 0-245-
54
692-
8
[9]
Giarratan
o
Jos
eph, R
i
l
e
y Ga
r
y
. E
x
p
e
rt S
y
s
t
em
s Princi
ple
s
and
Progr
a
mming
3rd E
d
ition. U
n
ite
d
States of America: PW
S P
ubli
s
hin
g
Comp
an
y. 19
98.
[10]
Munir R. Dig
ita
l
Image Proces
sing W
i
th Alg
o
r
i
th
m Appro
a
ch.
Informatika Pu
blish
i
n
g
, Band
ung. 20
04.
[11] IQ
Biometrics.
Faces. www.facesid.com
.
Acc
e
ssed o
n
Mach
8 2012
[12]
High Pri
o
rit
y
F
r
ee Soft
w
a
re Pr
ojects. F
r
ee So
ft
w
a
re F
o
und
at
ion. Access on
Janu
ar
y
26 2
0
1
2
[13]
W
i
nokur, Dan
n
y
. F
l
ash to F
o
cus on PC Bro
w
sin
g
an
d Mobil
e
Apps; A
dob
e to More Aggress
i
vel
y
Contri
bute to H
T
ML5. Access
on Jan
uar
y, 26
2012.
Evaluation Warning : The document was created with Spire.PDF for Python.