IEEE Referance Paper PDF

C65:E2C5C2F56E64E:@?2?5&C6G6?E:@?
J
246'64@8?:E:@?

DE ?5
D9H:?:#/:?;FC56 +:=2D!2>3=6
""#" "" !!" !! ""

FC2?82325?5:2 FC2?82325?5:2
2D9H:?:K:?;FC56 8>2:=4@> G3<2>3=6J29@@4@>

, !&$!& !! % !") " ! & "%& " 7:ED2?%)& H:==36D6?EE@E962AAC@AC:2E6C68:DE6C65 >@3:=6
"!* '% "$ " !! &$!%&"!% "!!&&* ! A9@?6)964FDE@>6C>FDEAFEE9:DC2?5@>=J86?6C2E654@56:?
" #$" % ! & #$"%% " &$"! #'$%% E96E6IE3@I,96?4FDE@>6CDAFE2=68:E:>2E6%)&E96JH:==
&$"$ !&$"' !) ##$" &" #$(!& && '$!
"!! &$!%&"!% ! "$$ &" #$"&& !"$ &"! &$"'
4@?E:?F6 H:E9 E96 EC2?D24E:@? )96 D68>6?E 42? 4@?E:?F6 (@
&)"%&# !% " '&!&&"! #$ $* %&# " E96 4@>3:?2E:@? @7 2 7246C64@8?:E:@? 2=8@C:E9> 2?5 2? %)&
'&!&&"!%($&"! &%! D:8?:7:42?E=J 564C62D6D E96 C:D<D @7 E967E A=FD 7C66D E96 FD6C
&%"'$"!+ &%$"& ! 7C@>E966IEC23FC56?@7C642==:?84@>A=:42E65A2DDH@C5D
& '&"$+&"! "$ "& &$' ! $''!& &$!%&"!% %
%' &&&"& !!)$&$ %!!! ( % &'%&%( #%"
!$&$%&%%'%$&!&'%$$!!%%
)96 AFCA@D6 @7 E96 AC@;64E :D E@ 4C62E6 2 7C2>6H@C< E92E
& #'$#"% " & ##&"! % &" $' $& $
$' *!")"& *%& '%&" $%&& "%& FD6D7246C64@8?:E:@?E@2FE96?E:42E62=68:E:>2E6A6CD@?:CDE
%%!* !& &$"!!! #$"$ E96 4FDE@>6C H:== 6?E6C E96 56E2:=D @7 E96 4C65:E 42C5 2?5 E96
,'&!&&"! $& $ %!!! :?7@C>2E:@? H:==36 4964<65 :?E9632?<52E232D6
$"!+ " !$* &&$!% #! 7E6CE96 G6C:7:42E:@?AC@46DD%)&:DAC@5F4652?5D6?EE@
&$!%&"!($&"!)

E96 4FDE@>6C 7 E96 %)& 92D 366? 4964<65 E96 FD6C H:== 36
$)'%*)%$ 2D<657@C72462FE96?E:42E:@?)96H6342>7246A:4EFC6H:==36
4@==64E652?57@CH2C565E@E9632?<D6CG6CD7@C2FE96?E:42E:@?
D 2 C6DF=E @7 C2A:5 25G2?46D :? D4:6?46 2?5 E649?@=@8J
:? 6?4CJAE65 7@C> )96 :>286 H@F=5 36 564CJAE65 :? E96
A@E6?E:2= E649?@=@8:6D 2C6 36:?8 D6E FA H:E9 D@=:5 AC@E64E:@?
52E232D6 2?5 FD65 7@C 2FE96?E:42E:@? AFCA@D6D )96 '(
FEE96C6QD2=H2JD2;FDE:7:42E:@?E@4C24<E9:DD64FC:EJDJDE6>
2=8@C:E9> :D FD65 7@C :>286 6?4CJAE:@? 2?5 564CJAE:@? )96
@? E96 @E96C D:56 ,9:=6 2FE@>2E:@? 6?92?46>6?E 92D 925 2
&JE9@? =2?8F286 :D FD65 7@C D4C:AE:?8 2?5 >2?:AF=2E:?8 E96
A@D:E:G6 @G6C2== 67764E ?F>6C@FD 7:?2?4:2= :?DE:EFE:@?D DF49
%A6?+=:3C2CJ7:=6E92E:D6>365565:?E@&JE9@?7E6CE92E
2D 32?<D 2?5 2AA=:42E:@?D DF49 2D )# 2C6 DE:== DF3;64E E@
E96 "& 2=8@C:E9> :D FD65 E@ 2FE96?E:42E6 E96 7246 *?=6DD
E967E2?57C2F5)96?6H)#4@?7:8FC2E:@?FD6D242C52?5
E96:56?E:EJDF:EDE96A:4EFC6 DE@C65:?E9652E232D6E96 FD6CQD
2&$E@>2I:>:K6D64FC:EJ:?E9642D6@74@F?E6C76:E42C5D
4C65:E42C542AH:==36C6G:6H652?5:7:ED2E:D7:6DE964C:E6C:2
@C 5F6 E@ 2C3:EC2C:=J :DDF65 &$D 5FA=:42E6 42C5D 2?5 G2C:@FD
E96 FD6C H:== 36 23=6 E@ >2<6 2 492C86 @C E96 492C86 H:== 36
@E96C 492==6?86D )@ D@=G6 2 9J3C:5 :?E6C7246 4@?D:DE:?8 @7
42?46==65
EC25:E:@?2= 762EFC6D 4@FA=65 H:E9 ?6H 762EFC6D DF49 2D 724:2=
'('(2=8@C:E9>:DFD657@C6?4CJAE:@?2?5564CJA
C64@8?:E:@?2?52@?6E:>6A2DDH@C5:D FD652E232D6DE@C6D
E:@? AFCA@D6 E >2<6D FD6 @7 EH@ <6JD &F3=:4 <6J 2?5
:?7@C>2E:@?23@FEE96 FD6CQD244@F?E52E2A9@E@8C2A9D@79:D
&C:G2E6<6J)96AF3=:4<6J:DFD65E@6?4CJAE >6DD286D
96C 7246 2?5 2 >@3:=6 ?F>36C E92E H:== D:8?:7:42?E=J 3@@DE
#6DD286D 6?4CJAE65 FD:?8 E96 AF3=:4 <6J 42? @?=J 36
AC@E64E:@? )96 AC:>2CJ E9:?8 E96 4FDE@>6C 92D E@ 5@ :D E2A
564CJAE65H:E9E96AC:G2E6 <6J
E96)#42C5)96=:G6:>286:DE2<6?2FE@>2E:42==JE9C@F89
%A6?+%A6?+%A6?@>AFE6C+:D:@?:D2=:3C2CJ
2 H6342> >@F?E65 @? E96 )# H9:49 :D 4@>A2C65 E@ E96
>2:?=J 2:>65 2E C62=E:>6 4@>AFE6C G:D:@? E AC@G:56D
:>286DD2G65:?E9652E232D67:E>2E496D2?%)&H:==
8C62E DFAA@CE 7@C 7246 56E64E:@? 2?5 7246C64@8?:E:@?
36 D6?E E@ E96 4@CC6DA@?5:?8 C68:DE6C65 >@3:=6 ?F>36C )9:D
E649?:BF6DFD:?8&JE9@?
C2?5@>=J86?6C2E654@5692DE@36AFE3JE964FDE@>6C:?E96
"&"&:D2EJA6@72=8@C:E9>FD657@C4=2DD:7:42E:@?:?
E6IE3@I7E964FDE@>6CDAFEG2=:5%)&E96EC2?,96?E9:D
4@>AFE6C G:D:@? E 92D 366? 7@F?5 E@ 36 2 A@H6C7F=
762EFC66IEC24E:@?2?54=2DD:7:42E:@?AFCA@D6D

N L
86
Authorized licensed use limited to: Western Sydney University. Downloaded on June 15,2021 at 20:26:01 UTC from IEEE Xplore. Restrictions apply.
")')*' (*'+. >2?J 492?46D E@ DE62= E96 AC:G2E6 :?7@C>2E:@? E9C@F89 E96
2EE24<6CD @C 924<6CD (@ H6 6?5@CD6 2 3C2?5 ?6H >6E9@5 E@
#@9D:? !2C@G2=:J2 0 1 :D AC@A@D65 E9:D DJDE6> :D E@ >2<6 <66A2H2J7C@>7C2F5F=6?E2==E9C@F89@?=:?6EC2?D24E:@?D2?5
EC25:E:@?2=)#>@56=>@C6D276,62=D@AFE7@CH2C52?6H E@ 4@KJ E96 724ED 3J >62?D @7 2 DE6A G6C:7:42E:@? AC@465FC6
4@?46AE E92E H:== :>AC@G6 E96 @G6C2== EC2?D24E:@? 6IA6C:6?46 )96 :?7@C>2E:@? :D AC@46DD65 2?5 E96 24<?@H=6586>6?E :D
C6=:23:=:EJ 2?5 4@>7@CE 2E E96 )# 62EFC6D DF49 2D 7246 5:DA2E4965 E@ E96 32?< 7@C 6249 E96 =68:E:>2E6 2?5 :?G2=:5
C64@8?:E:@? 2?5 %?6):>6 &2DDH@C5 %)& 2C6 FD65 7@C EC2?D24E:@?D 2 ?6H 2AAC@249 @7 4C65:E 42C5 D42??:?8 92D
:>AC@G:?8 244@F?E D64FC:EJ 2?5 FD6C AC:G24J246 A@AF=2C:EJ 36?67:4:2= 2EEC:3FE6D :? A9C2D6D @7 4@DE 7:?2?4:2= D2G:?8D 2?5
86?6C2E:@? 2==@HD E96 82586E E@ :56?E:7J 6249 2?5 6G6CJ E:>6A6C7@C>2?46
A6CD@?F?:BF6=J96?46>2<:?872462D2<6J)9:D4@>A=6E6=J
6=:>:?2E6D E96 AC@323:=:E:6D @7 7C2F5 3642FD6 @7 E967E 2?5 + -()$(.()# &&'%
5FA=:4:EJ @7 E96 )# 42C5D FCE96C>@C6 E96 C2?5@>=J
C2F5 :D @?6 @7 E96 3:8 =682= AC@3=6>D :? E96 4C65:E D4@C6
86?6C2E65%)&7C66DE96A6CD@?7C@>C6>6>36C:?8&$D2D:E
42C5 :?5FDECJ )96 >2:? 8@2=D 2C6 7:CDE @7 2== E@ C64@8?:K6
:ED6=724ED2D2&$
E96F?FDF2=7@C>D@74C65:ED4@C642C57C2F52?5D64@?5=JE@

E6DE E96 5:776C6?E >6E9@5D FD65 E@ 56E64E 7C2F5 01 )96 DF3
'FA:?56C (2:?: 0 1 :?EC@5F4:?8 E9:D DJDE6> 277@C5D 4@?EC2DE
AFCA@D6:DE@AC@G:566G2=F2E62?52?2=JK6C646?E=JAF3=:D965
4@?46C?:?8 5:G6CD6 3:@>6EC:4 DJDE6>D :? C62=:EJ 3J FD:?8
7:?5:?8D :? E96 56E64E:@? @7 4C65:E D4@C6 42C5 7C2F5 )9:D
567:?:?8E96:C25G2?E286D2?55:D25G2?E286D D9@CE4C62E:@?
:D 86?6C2==J @776C65 C682C5:?8 ?@C>2==J FD65 3:@>6EC:4D 2CE:4=6 56D4C:36D DA64:7:4 E6C>:?@=@8J 7@C 4C65:E D4@C6 42C5
7C2F5 2?5 9:89=:89ED :>A@CE2?E 724ED 2?5 7:8FC6D :? E9:D 7:6=5
:?4=FD:G6@7246C:D:?86CAC:?E:?86C+6:?":AD+@:46
2D65 @? E96 ?2EFC6 @7 E967E 6?4@F?E6C65 3J 32?<D @C 4C65:E
)964@?EC2DEDE2?52C5D=:DE:?825565:DC6DEC2:?65E@244FC24J
42C5 4@>A2?:6D 2 G2C:6EJ @7 DE6AD 42? 36 :>A=6>6?E65 2?5
D:K6@7E6>A=2E64@DED276EJDE2862?5=@?8E:>632=2?46
6?7@C465)96C64@>>6?52E:@?D>256@?E9:DA2A6C>2J92G6

!9J2E: 92F592CJ 0 1 AC@A@D65 :? AC6D6?E DE2E6 @7 2772:CD G2=F23=6 492C24E6C:DE:4D :? E6C>D @7 AC:46 C65F4E:@?D 2?5 E:>6
H9:=DE E96 E:>6 A6C:@5 7C2F5 4@>6D :?E@ 2 5:D4FDD:@? 4C65:E 677:4:6?4J)96:>A@CE2?46@7E96FD6@7E96>6E9@5D6I2>:?65
42C5 7C2F5 4=:4<D E@ >:?5 E9FD 72C ,:E9 E96 6IEC2@C5:?2CJ C:89E 96C6 =:6D :? E96 >:?:>:K2E:@? @7 4C65:E C6A@CE 42C5 E967E
8C@HE9 :? 4C65:E 42C5 EC2?D24E:@?D 4C65:E 42C5 7C2F5 92D @H6G6C E96C6 2C6 DE:== =682= :DDF6D H96? 8@@5 4C65:E C6A@CE
:?4C62D:?8 6I46DD:G6=J :? C646?E J62CD C2F5 56E64E:@? 42C54@?DF>6CD2C6>:D4=2DD:7:652D5:D9@?6DE?255:E:@?
4@?D:DED @7 EC24<:?8 @7 E96 DA6?5:?8 3692G:@C @7 4FDE@>6CD E@ E96 6E9:42= ?2EFC6 @7 5:D9@?6DE 24E:@?D E96C6 92D 366? 2
4=:6?ED :7 J@F H2?E E@ H:==A@H6C 56E64E:@? @C 2G@:52?46 @7 8C@H:?8 :?E6C6DE :? 32?<:?8 6E9:4D 7@C D@>6 E:>6 4C:E:42=
F?56D:C23=6 4@?5F4E D 4C65:E D4@C6 42C5 H:== 364@>6 E96 G6?EFC6E@96=A3FD:?6DD6D2?5&@=:E:42=:?DE:EFE:@?D2=@?8H:E9
>@DE H:??:?8 >@56 @7 492C86 7@C 6249 @?=:?6 :? 255:E:@? E@ 32?<D >FDE E2<6 >62DFC6D E@ 56E6C 2?5 C6D@=G6 E967E BF:4<=J
C68F=2C AFC492D6 7C2F5 C6=2E6 H:E9 :E 2C6 2=D@ 2446=6C2E:?8 2?5 4@CC64E=J

C2F556E64E:@?:D:?G@=G65H:E9?@=@?86CD:>A=6DE42AEFC:?8
+ #)%%"%. *(
E96 7C2F5F=6?E 6G6?ED 3FE 255:E:@?2==J D9@@E:?8 @7 DF49
24E:G:E:6D2DBF:4<2D G:23=6 !!

?:DD2":?E2?8'2>2592?:0192DAC@A@D6525565246:DE96 G6CJ :>286 :D 7@C>2E:@? @7 ' D9256D G6CJ 42AEFC65
A:4EFC692DD@>6?@:D6F?H2?E65324<8C@F?5)96C67@C6E96C6
>@DE>6>@C23=62A2CE@7E967C2>6:?24EF2==:76E92E>2<6D
:E 2? 4C:E:42= G2C:23=6 @? E9:D DEF5:6D H6 FD6 7246 2AAC696?5 :DH2?E@7>2??6CE96@?6D42AEFC65:>286367@C62DD:8?E@@FC
C64@8?:E:@? >@5F=6 &C6AC@46DD:?8 F?:E >256 :D FA @7 ?@:D6
E649?:BF6 E92E :?4@CA@C2E65 :? 'J* C@3@E:4 )96 C@3@E :D
5:28?@D65 H:E9 E96 2:5 @7 2 56E64E65 G@:46 4@>>2?5 =@@<:?8 C6>@G2=8C6J:>2864@?G6CD:@?3:?2CJA:4EFC64@?G6CD:@?@7
7@C D@>6@?6 2?5 H96? D@>6@?6QD 7246 92D 366? 56E6C>:?65 6?E6CA:I27E6CE92E492C24E6C:DE:46IEC24E:@?42CC:65@FE@?E9@D6
D2>A=6D ? 7FEFC6 6IEC24E:@? 7:G6 DE6AD :>A=6>6?E65 H96C6:?
7246 C6AFE2E:@? :D 6?E:C6 )9:D E6IE H:== 2AA=J E96 9F>2? 7246
C64@8?:E:@? 56G:46 E96 FD286 @7 E96 :86?7246 2AAC@249 7:?5:?8E966446?EC:4:EJ$6IE6=@?82E:@?D@7A:I2C66G2=F2E65
G:242=4F=2E:?8A:I6=D68>6?E2E:@?2DH6==2DC@E2E:@?@76?E6C
:86?7246 :D @?6 @7 E96 724:2= C64@8?:E:@? DEC2E68:6D 32D65
E@E2==J2EE967F?52>6?E2=:DDF66G2=F2E:@?&D6E@7CF=6D A9@E@8C2A9D
& :?G@=G652 >2E96>2E:42=AC@465FC6E@56C:G62 92C52?5
! %
72DE @7 762EFC6D 7@C 7246 C6AFE2E:@? 246 C64@8?:E:@? =6G6=
368:?D H:E9 7246 56E64E:@? E649?:BF6 E96 FD286 @7 42D4256 #249:?686EE:?8E@<?@H:D24@>A=6I2C62)96:>A=6>6?
4=2DD:7:6C >6E9@5 7246 AC6AC@46DD 24BF:C6 2?5 65F42E6 E96 E2E:@? :? >249:?6 82:?:?8 <?@H=6586 @7 2?5 :?EC@5F4E:@? @7
724656E64E652?5DF3D6BF6?E=JE967246A@AF=2C:EJ >@56=D:D2=@EE@F892?55:77:4F=EE92?:E H2D H2J E@82586E
>2DE6C:?8 E649?@=@8J 2?5 7C2>6H@C<D ?4=FD:G6 @7 @@8=6QD
2?2?: ( ' 01 DE2E6D 4C65:E 42C5 277@C5D AC@>:?6?E FD6 @7 )6?D@C 5C:7E E92E >2<6D @FC 492==6?86 D:>A=6 :E :D AC@465FC6
492C86 >6E9@5 D@ :EQD 72C 244@>A2?:65 :? >2?J D46?2C:@D @7@3E2:?:?8C64@C5DEC2:?:?8>@56=DD6CG:?8AC65:4E:@?D2?5
DH6<?@H:?E964@FCD6@7@?=:?6EC2?D24E:@?DE96C62C6 C67:?:?856DE:?J67764ED
87
$#" # "% !

? AC@A@D65 H@C< H6 2C6 FD:?8 $$ H9:49 E2<6D 7246
:>286D2D2:?AFE7E6C86EE:?8:>286D7C@>@A6?4GAJE9@?:E
H:== AC@46DD65 FD:?8 :>286 AC@46DD:?8 E649?:BF6D 7@C 762EFC6
6G2=F2E:@? ,6 6IEC24E 5:776C6?E 762EFC6D 7C@> E9@D6 :>286D

FD:?82C42D4256JFD:?82D6C:6D@7>2E96>2E:42=7F?4E:@?D

H6 2C6 8@:?8 E@ :56?E:7J E96 F?:BF6 7246D G6CJ =2J6C :?
$$ 92D 42A23:=:EJ E@ 7:?5 @FE H6:89ED @7 :>286D 3J FD:?8
>2EC:I 6G2=F2E:@?D H9:49 4@?G6CED :?AFE E@ @FEAFE H:E9 G2=F
23=6 7F?4E:@?D "2J6CD @7 $$ FD65 E@ :56?E:7J >2E4965 7246
7C@>6IEC24E65:>286D2?58:G6AC65:4E:@?3JAC6D6CG:?89:89
244FC24J2?5=6DDE:>6

(E6A ?AFE7246 :>286 :8 &@@=:?8=2J6C
(E6A 2466IEC24E:@?
(E6A >286AC@46DD:?83JFD:?8@A6?4G
(E6A62EFC6IEC24E:@?7C@>:>286D

(E6A#@56=86?6C2E:@?

4=2EE6?:?8
(E6A7246C64@8?:E:@?

@FC>2:?=2J6CH@C<:?82AAC@249@7$$6IA=2:?65 6?6C2==J 96C6 H6 AFE E96 A@@=65 762EFC6 :?E@ 2 D:?8=6
36=@H 4@=F>?2D2D2>A=6:?AFE 7@C7FCE96C=2J6CEC2?D7@C> E96
>2EC:I52E2E@ >2EC:I 52E2
2@?G@=FE:@?2="2J6C

,62C68@:?8E@6IEC24E5:776C6?E762EFC6D@77246:>286D=:<6
A:I6= H6:89E >2EC:I 42=4F=2E:@?D 3J FD:?8 762EFC6 <6C?6=D
&6C7@C> >2E96>2E:42= 4@?G@=FE:@?D @? :>286 H96C6 6G6CJ
7F?4E:@?FD6D2F?:BF67:=E6C)9:D@FE4@>6H:==36:?5:776C6?E
762EFC6 >2AD E E96 6?5 H6 H:== 4@==64E 2== @7 E96D6 762EFC6
>2AD 2?5 5C27E E96> 2D E96 56DE:?2E:@? @FEAFE >2EC:I @7 E96
4@?G@=FE:@?=2J6C
:8 =2EE6?:?8#2EC:I

5F==J@??64E:@?

7F==J4@??64E65=2J6C92D7F==4@??64E:@?D@7?6FEC@?DE@2==
E96 ?@56D :? E96 AC6G:@FD =2J6C )96 7FD:@? @7 >@C6 ?6FC@?D
E@6G2=F2E6D244FC2E6=J

:8 @?G@=FE:@?2="2J6C

3&@@=:?8
)96 6IAC6DD:@? @7 A@@=:?8 :D E@ 4@?DE2?E=J 564C62D6 E96

5:>6?D:@?2=:EJE@=:>:EDE96?F>36C@7724E@CD2?542=4F=2E:@?
:?E96?6EH@C<)9:D=:>:EDE96E:>6@7EC2:?:?82?5>2:?E2:?D
@G6C 7:EE:?8 AC@3=6> )96 >2I &@@=:?8 6IEC24ED @FE E96
=2C86DE A:I6= G2=F6 @FE @7 2 762EFC6 ,9:=6 A@@=:?8 2G6C286
:D 42=4F=2E65 7@C E96 2G6C286 A:I6= G2=F6 E92E 92D E@ 36
6G2=F2E65&@@=:?8=2J6C@A6C2E6D@?6249762EFC6>2A
:?56A6?56?E=J :8F==J@??64E65"2J6C
88
:>286D 7@C 762EFC6 6IEC24E:@? 2?5 :>286 AC@46DD:?8 24:2=
724E@C 6G2=F2E:@? :D E96 AC@465FC6 @7 86EE:?8 7246 A2CED =:<6
6J6D ?@D6 >@FE9 6E4 7C@> C62= E:>6 7246 :>286D 24:2=
#249:?6"62C?:?8H9:49:D2A2CE@7EC2:?DE96DJDE6>E@ 724E@C6G2=F2E:@?:DG6CJ>F496DD6?E:2=7@CE96:?:E:2=:K2E:@?@7
=62C? 2FE@>2E:42==J 2?5 6IA6C:6?46D E@ :>AC@G:D6 H:E9@FE AC@46DD:?8E649?:BF6D=:<6724656E64E:@?2?57246C64@8?:E:@?
36:?86IA=:4:E=JAC@8C2>>65"@42=:?2CJ(2>A=6:DE@8C2>
:D 2 4@>>@? 2?5 62DJE@FD6 E6IEFC6 @A6C2E@C E92E >2C<D E96 C2F5 &C6G6?E:@? 6E64E:@? 7E6C 7246 56E64E:@? :7
A:I6=D @7 2? :>286 3J E9C6D9@=5:?8 E96 G:4:?:EJ @7 6249 A:I6= 4FDE@>6C :D =682= 4FDE@>6C E96? 2446DD 8C2?E E@ 4FDE@>6C
2?54@?D:56C:?8E966?5C6DF=EE@3623:?2CJD6E%A6?(@FC46 @E96CH:D6 42C5 3=@4<65 3J 32?< ,9:49 2C6 >2:? E2D< E@
@>AFE6C +:D:@? :D 2 =:3C2CJ @7 AC@8C2>>:?8 7F?4E:@?D E92E 5:776C6?E:2E686?F:?64FDE@>6C2?5:==682=4FDE@>6CDJFD:?8
2C6>2:?=J56D:8?657@CC62=E:>6 >249:?6 G:D:@? C62= E:>6 7246 C64@8?:E:@? H6 2C6 ECJ:?8 E@ AC@G:56 A9JD:42=
=6G6=D64FC:EJE@4C65:E42C52FE96?E:42E:@?7C2>6H@C<
+ &'%&%((.()# &&'%
+ '(*")($ (*((%$
? 2 AC@A@D65 DJDE6> H6 92G6 AC@A@D65 C62= E:>6 724:2=
C64@8?:E:@?32D654C65:E42C5G6C:7:42E:@?2?57C2F5AC6G6?E:@?

)"
DJDE6> %#&'(%$)"%'"'%$)%$

$ " #$%"!$ &$
"

? @FC AC@A@D65 4C65:E 42C5 7C2F5 7C2F5 56E64E:@? DJDE6>

>2:?=JD9@HDC6DF=E@7 5:776C6?EA2C2>6E6C7@C4@>2A2C:D:@?
=:<6 "&$$ 2?5 7:?2= 244FC24J? 7@==@H:?8 E23=6 H6
D66?E92EG2=F6@7"&:D$$:D 2?57:?2=244FC24J
:D

:8=@4<:28C2>@7&C@A@D65(JDE6>

? 2 AC@A@D65 DJDE6> H6 2C6 8@:?8 E@ @G6C4@>6 6I:DE:?8

:8@>A2C:D:@?8C2A97@C724:2=C64@8?:E:@?
5C2H324<D 2?5 AC@G:56 C62= E:>6 7C2F5 56E64E:@? >6492?:D>
32D65@?>249:?6=62C?:?82?5@A6?+AJE9@?
,62C68@:?8E@:?G6?E7@==@H:?8DF3>@5F=6D
D A6C @FC :>A=6>6?E2E:@? ?665 H6 2C6 H@C<65 @? C62=
2E2)C2:?:?8?E9:D724625>:?@74C65:E42C5DJDE6> E:>67246:>286D42EFC657C@>DJDE6>42>6C2H9:=64FDE@>6C
42? 42AEFC6 >F=E:A=6 7246 :>286D @7 4FDE@>6CD 2?5 C68:DE6C AC@G6D :ED :56?E:EJ &9JD:42= =6G6= 2FE96?E:42E:@? 4@?D:DED @7
C6E:?2 D42? E9F>3 D42? 2?5 7246 D42? )96 =2DE 2?5 >2:?
E96> :?E@ E96 DJDE6> 7E6C 86EE:?8 7246 52E2D6E :E EC2:? E96
7246>@56=7@C62492?56G6CJ4FDE@>6C?EC2:?:?8A92D6H6 =6G6= @7 2FE96?E:42E:@? :D A9JD:42= =6G6= 2FE96?E:42E:@? 4964<
92G6 FD65 $$ 2?5 "& 7@C 6IEC24E:?8 7246 762EFC6D 7@C >@?82==23@G6D64FC:EJ724E@CDH62C6FD65724:2=D42?FD:?8
4C62E:?87246>@56=7E6C4C62E:?87246>@56=H62C6E2<:?8 C62= E:>6 7246 C64@8?:E:@? FD:?8 @A6?4GAJE9@? 2?5 >249:?6
C62= E:>6 7246 :>286 H9:=6 =@8:? :?E@ 4C65:E 42C5 =62C?:?8 ? H9:49 H6 92G6 E2<6? 724:2= :>286 D2>A=6 @7
2FE96?E:42E:@?DJDE6> 44FDE@>6C2?5AC@G6:ED:56?E:EJ3JFD:?8@FC>249:?6=62C?:?8
EC2:?65 >@56= )96 "& :D FD65 E@ C64@8?:K6 C62= E:>6 7246
246762EFC66IEC24E:@?*D:?8@A6?4@>AFE6CG:D:@? :>286 FD:?8 92C 42D4256 4=2DD:7:6C 2?5 7C@?E2= 7246 I>= 7:=6D
=:3C2CJ ,6 2C6 8@:?8 E@ 42AEFC6 C62= E:>6 7246 :>286D @7 *D:?8@A6?4GAJE9@?:>286D42?EC2?7@C>7C@>@?64=2DD:7:6C
4FDE@>6CD7E6C86EE:?87246DH62C67@CH2C5:?8 E96D6 E@@E96CFE@?=JFD:?8"& H642?E86EE:?86IA64E65
89
244FC24J:E86E?62C=J23@FEA6C46?E244FC24JH9:=6H@C<:?8
H:E9 C62= E:>6 7246 :>286D )96 D64@?5 2AAC@249 FD65 E@
C64@8?:K6C62=E:>67246:>286D:D$$?H9:49H62=C625J
EC2:??65 4FDE@>6CD 7246 :>286D E@ 4C62E6 EC2:? >@56= 7@C @FC
?6FC2= ?6EH@C<J FD:?8$$2AAC@249 H686E ?62C=J23@FE
A6C46?E 244FC24J 7E6C E92E H6 4@>3:?6 3@E9 4=2DD:7:6CD
"& 2?5 $$ EC2:?65 >@56= E@ 244FC2E6=J C64@8?:K6 7246D
2?5 8C2?E 4FDE@>6CD 7 F?2E96?E:42E65 A6CD@? EC:6D E@ =@8:?
FD:?8 4C65:E 42C5 56E2:=D E92E E:>6 E96J 72:=65 E@ AC@G6 724:2=
C624@8?:E:@? )96? H6 86E :?E:>2E:@? 23@FE HC@?8 2EE6>AED
E9C@F89C68:DEC656>2:=2?542C586EE6>A@C2C=J3=@4<65
+ %$"*)%$$*)*',%'!
,6 2C6 4C62E:?8 2 4C65:E 42C5 7C2F5 >@?:E@C:?8 DJDE6>
7@4FD65@?7246E@7246:56?E:7:42E:@?:?C62=E:>6H:E92G:6H
E@ 6?DFC:?8 E96 D276EJ @7 4@?DF>6CD 2?5 2G@:5:?8 23FD6 @7
4C65:E42C5D2D2H9@=6)96AC@A@D652AA=:42E:@?:D4@?EC2DE65
H:E9E96724:2=C64@8?:E:@?2DD:DE2?5E96DEF5JF?:E@C7FEFC6
AC24E:46E962F5:@2?5A9JD:42==6G6=762EFC6D@7E96C6E:?22?5
E9F>3H:==92G6>F=E:=6G6=AC@E64E:@?
!$%,"#$)
)9:DH@C<:DDFAA@CE65:?24C65:E42C57C2F556E64E:@?2?5
AC6G6?E:@? 3J 7246 C64@8?:E:@? @7 2?J DE2E6 :? :?5:2 FE9@CD
2C6E92?<7F=E@24F=EJ&(%FC2?823257@CAC@G:5:?8E96
724:=:EJE@42CCJ@FEE96C6D62C49H@C<
''$(
0 1 #@9D:? !2C@G2=:J2 O?92?465 D64FC:EJ 7@C )# >249:?6 H:E9 %)&
2?524:2=C64@8?:E:@?762EFC6DP?E6C?2E:@?2=@?76C6?46@?5G2?465
@>AFE:?8)649?@=@8:6D2?5AA=:42E:@?D) >2J
0 1 'FA:?56C(2:?:O4@>A2C:D@?@7G2C:@FD3:@>6EC:4>6E9@5DP E9
?E6C?2E:@?2= (J>A@D:F> @? AA=:65 @>AFE2E:@?2= ?E6==:86?46 2?5
?7@C>2E:4D(
0 1 !9J2E: 92F592CJ O C6G:6H @7 C2F5 6E64E:@? )649?:BF6D C65:E
2C5P ?E6C?2E:@?2= @FC?2= @7 @>AFE6C AA=:42E:@?D M
+@=F>6M$@ #2J
01 ?:DD2 ":?E2?8 '2>2592?: OF>2? 246 '64@8?:E:@? AA=:42E:@?
*D:?8 & 2?5 :86?7246 AAC@249P (64@?5 ?E6C?2E:@?2= @?
76C6?46@??7@C>2E:4D2?5@>AFE:?8$%+
01 2?2?:('OD64FC654C65:E42C5EC2?D24E:@?DFD:?8H6342>P ?E6C?2
E:@?2='6D62C49 @FC?2=@7?8:?66C:?82?5)649?@=@8J' )
90
Proceedings of the Fifth International Conference on Intelligent Computing and Control Systems (ICICCS 2021)
IEEE Xplore Part Number: CFP21K74-ART; ISBN: 978-0-7381-1327-2
ENHANCED SECURITY FEATURE OF

ATM’S
THROUGH FACIAL RECOGNITION
Ms. SOUNDARI D V
Department of ECE
2021 5th International Conference on Intelligent Computing and Control Systems (ICICCS) | 978-1-6654-1272-8/21/$31.00 ©2021 IEEE | DOI: 10.1109/ICICCS51141.2021.9432327
Sri Krishna College of Engineering

and Technology
Coimbatore, India
soundaridv@skcet.ac.in
ARAVINDH R EDWIN RAJ K ABISHEK S

Department of ECE Department of ECE Department of ECE
Sri Krishna College of Engineering Sri Krishna College of Engineering Sri Krishna College of Engineering
and Technology and Technology and Technology
Coimbatore, India Coimbatore, India Coimbatore, India
aravindh10100@gmail.com mathaenggvdsr@gmail.com abishek.srinivasan@gmail.com
Abstract— Automated Teller Machines also Keywords—ATM, Face-id, Eigenface algorithm,

known as ATM’s are widely used nowadays by Machine Learning.
each and everyone. The ATM machine (Automated
Teller Machine) is an electronic device that is used I. INTRODUCTION
by the banks to perform banking tasks like Nowadays, the uses of ATMs are inevitable in our
withdrawal of money, transferring of money, and day-to-day lives. Before the introduction of ATM,
many to get many information about a user's bank people used to travel with cash in their hands. This
account without the need to visit a bank. This caused the risk of getting robbed or the cash might get
System revolutionised the way of transactions. lost due to the negligence of the user. To overcome this
There were no long lines of queue in front of the issues ATM played a big role. It is like cash in a card.
bank for a simple withdrawal of money. The As the ATM technology got improved the idea of mis-
number of ATM’s a bank has can be a factor in using it also got improved. The main disadvantage of
considering the strength of a bank. As there is this system is that if the PIN is known anybody can use
increase in the number of ATM’s, there is also it to take money.
increase in the fraudulent activities in the ATM.
To address this issue, Face-id is used as a key. The
The main motivation of this project is to increase
main use of the Biometric is that it is unique for a
the security feature of the use of ATM. The current
person. The proposed innovation is not an alternative
method uses static key (PIN) for security. The
for the conventional ATM security system. The
proposed method uses Face-id as a key
proposed innovation is considered as an additional
incorporated with current method. The advantages
support for the existing method. In India, nearly 2000
can be found as that the face-id is unique for
ATM crimes are occurring every year. The proposed
everybody; it cannot be used by anybody other
method will surely reduce this rate to a greater extent.
than the user. For the implementation of the face-id
scan, the machine learning and image processing
algorithms (Eigenface algorithm) are used.
978-0-7381-1327-2/21/$31.00 ©2021 IEEE 1252
Authorized licensed use limited to: BOURNEMOUTH UNIVERSITY. Downloaded on July 04,2021 at 12:33:32 UTC from IEEE Xplore. Restrictions apply.
II. RELATED WORK relation from the user's cell phone. Since no special
Mohsin Karovaliya in paper [2] proposes subscription to an SMS service provider is needed,
Eigenface based method for the face recognition. This GSM modems can be a fast and easy way to get started
system analyzes the algorithms used in the previous with SMS. A GSM modem can also be a regular GSM
systems. PCA based algorithm is more reliable, very phone with the required cable and software. A GSM
fast and the storage space takes is very less. The main modem may also be a regular GSM phone with the
drawback this method is that it can be manipulated by required cable and software driver to connect to a serial
using photos of users. This method can be improved by port or USB port.
using 3D face masks but the cost of making 3D masks
is very high. III. PROPOSED METHOD
Most people use an ATM for a lot of purposes and
The paper [3] suggests a vibration sensor which in a lot of scenarios. The most common three scenarios
senses vibrations produced from ATM machine are considered here and the solutions are found.
whenever robbery occurs. This system processes real-
x The user as himself using the ATM machine.
time data collected by the vibration sensor using an
ARM controller-based embedded system. The buzzer x If the user is threatened by somebody to get
money without the user’s willing.
will emit a beep sound once the vibration is detected.
The ATM door is closed with a DC motor. There are x If the user’s friend uses the card of the user.
also some additional security measures in place. This First of all, when a user enters the ATM, two choices
will deter theft and make it possible to arrest the are given. Whether the user comes with his card
suspect. Two software programs are used to execute the (Number 1) or if the user’s friend comes with the card
(Number 2).
software, the first of which is Keil Vision 3.0. The
Flash magic simulator is the second. Keil Vision A. User uses the ATM by himself
Debugger correctly simulates on-chip peripherals. This When a user by himself uses the ATM machine,
device aids in rapid reaction and minimization of first the card must be inserted. Now a camera captures
failure by detecting the ATM machine at real-time the image of the user. Now, the captured image is
when it has been stolen can be identified by GSM compared with the image stored in the database. If both
technology. the image is matched (Eigenface algorithm does the
In paper [4], the finger print recognition is done with comparison part) then a successfully authenticated
curvelet transform by finding the Euclidean distance message is displayed. Now the user needs to enter the
between the two corresponding finger codes. The test PIN. If the PIN entered is correct then further
finger code is compared with the entire finger codes in transactions can be done.
the database. An OTP will be sent to the matching
registered mobile number if it matches. The built-in B. User is threatened by somebody
MATLAB feature ‘imread' is used for preprocessing. Now the user is threatened by somebody to get money.
The histogram equalization approach improves an For this scenario we have a robust solution. First the
image's global contrast by shifting the intensity face of the user will match. Now for entering the right
distribution on a histogram marginally. This helps low- PIN, the user must enter a wrong PIN. This causes a
contrast areas to achieve more contrast without alert at the police portal which is at the backend. The
affecting the overall contrast. This is achieved by police portal will be at idle stage if the transactions
histogram equalization, which essentially spreads out were smooth. If the wrong PIN is entered then the alert
the most common strength values. Curvelet transform is received.
and FFT can also be used for function extraction.
C. User’s friend user the card
When the user’s friend uses the card, obviously the
The ATM security is enhanced by adding GSM face will not be matched. For this case, we proposed a
module to generate OTP in paper [5]. Where there is a clear method. The user should set a PIN and
network issue with GSM technologies, this system uses withdrawal limit in the mobile app. This must be set for
Bluetooth to link to an ATM, which produces an OTP each and every transaction. Now the user’s friend must
978-0-7381-1327-2/21/$31.00 ©2021 IEEE 1253
enter this secret PIN. If the PIN matches, then the

amount set by the user will be automatically
withdrawn.
Result of csae1. The face is authenticated, The PIN

Flowchart is matched. Now the amount is entered. The process is
done successfully.
Result of case 2:if the face is matched and the PIN

is entered wrong then an alert is sent at the police
terminal at the back end.
Result of case 3: If the users friend uses the card,

Then the PIN (set by the user) must be entered
correctly After correct entry of the PIN amount is
withdrawn.
DRAWBACKS OF THE EXISTING SYSTEM

x Anybody who knows the PIN can use the card
for the transactions.
x Since the PIN is static it is easy for criminals to
extract it from the users, through hacking or
even through brute force applied to the person.
x Scanning the magnetic strip in the ATM cards
will give the complete details of the card. This
causes the duplication of the cards.
x Threatening people in front of ATM is done in
some places and the security feature of PIN
cannot be useful in these situations.
978-0-7381-1327-2/21/$31.00 ©2021 IEEE 1254
ADVANTAGES OF PROPOSED SYSTEM (2) Train the Recognizer: Feed the recognizer the face
data as well as the names of each face so that it can
x The proposed method uses a biometric, this remember.
increases the security feature of the existing (3) Recognition: Feed the face recognizer with new faces
system drastically. of certain individuals to see how it knows them.
x Since we use biometric it is unique for each and OpenCV has two built-in face recognizers. The
every individual. names of those face recognizers are: Eigenfaces and
x The condition of threatening the user is well Fisher faces.
addressed here, and this can result in the safety
of the user and the user’s money.
x The withdrawal limit is set if the user’s friend
3.3.4 Eigenfaces Face Recognizer Algorithm
uses the card. This helps the user to be in control
of his account. In this algorithm, a facial image is a point from a
high-dimensional image space and a lower-dimensional
x Implementation of this method is easy, cheap, representation is found, where classification becomes
and efficient. easy. The lower-dimensional subspace is found with
Principal Component Analysis (PCA), which identifies
the axes with maximum variance. While this kind of
Biometrics Performance Accuracy Cost transformation is optimal from a reconstruction
Iris Medium Medium High standpoint, it doesn’t take any class labels into account.
Imagine a situation where the variance is generated
Fingerprint High Medium Low from external sources, let it be light. The axes with
Retina High High High maximum variance do not necessarily contain any
discriminative information at all, hence a classification
becomes impossible.
The functioning of the algorithm is that it considers
ALGORITHMS USED
that not all parts of a face are important equally to
Since we use face data as the key here. We use recognize a face. Instead, the main feature like the
image processing algorithms here. The implementation shape of nose, ears, fore-head are used and considered
is done using OpenCV in python which has built in how they vary from one another The main idea is to
image recognising algorithms. There algorithms are find the region with maximum difference. Let us
trained using machine learning. We use Eigenface consider that there will a drastic amount of variation
Algorithm. will be there when we compare the region of eyes and
nose. When multiple faces are to be compared, the
FACE RECOGNITION comparison between the faces is done looking at the
Face recognition is an easy task for humans. Seeing maximum variation among the faces, and they help to
the geometric features of the face is considered as the distinguish the faces. This is the process behind
most easy and efficient way in facial recognition. On Eigenfaces recognizer. This recognizer works by
early days the first automatic face recognition uses the looking at all the pictures that are used to train and
marker points like position of ears, position of nose etc. extracts the key components that are considered to be
These points were used in framing the feature vector relevant and ignores the other components. Those key
like distance between them or angle between them. The components are called as Principal components and
facial recognition process was done by finding the these functions as main source for the recognizer.
distance between the feature vector of the captured This face recognizer also maps the records of the
image with a reference image. Some of the latest work face to the person whose record is that. So, when a new
on geometric face recognition was, a 22-dimensional record needs to be added it follows the following
feature vector and experiments on large datasets have process. The principal component is extracted first. The
shown that geometrical features alone may not carry comparison process is done next with the images used
enough information for face recognition. to train. Find the image which matches the best. Now
Mainly Computer coding facial recognition can be map the record with the person name. In eigenface
broken down into three simple stages that are identical recognizer the light acts as one of the important needs.
to the steps used by the human brain to recognize faces. Only with the help of light the, shadowed and
These are the steps: illuminated parts of the face is found. These details are
used to find an eigenface which are used to represent a
(1) Data Collection: Gather face data (in this case, face face. The shadows and shapes of two eyes, a nose, a
images) of the individuals you want to recognize. mouth is the most used parameter in facial recognition.
978-0-7381-1327-2/21/$31.00 ©2021 IEEE 1255
[4] [4]C. Bhosale, P. Dere, C. Jadhav, "ATM security using face and
fingerprint recognition", Interna- tional Journal of Research in
Engineering, Technology and Science, Volume VII, Special Issue,
Feb 2017.
[5] [5]Manoj V , M. Sankar R , Sasipriya S , U. Devi E, Devika T ,
"Multi Authentication ATM Theft Prevention Using iBeacon",
International Research Journal of Engineering and Technology
(IRJET).
[6] L. Wang,H. Ji, Y. Shi, " Face recognition using maximum local
fisher discriminant analysis",18th IEEE International Conference
Image showing how the Eigenface algorithm works.
on Image Processing, 2011
[7] K.Shailaja and Dr.B.Anuradha, "Effective Face Recognition using
IV. CONCLUSION Deep Learning based Linear Discriminant Classification ", IEEE
On summing up, the method what we proposed is International conference on Computational Interlligence and
Computing Research,2016.
far better in increasing the security feature. The main
[8] H. R. Babaei, O. Molalapata and A.H.Y Akbar Pandor, "Face
goal of our paper is to incorporate the facial recognition Recognition Application for Au- tomatic Teller Machines
feature along with the existing conventional method for (ATM)",International Conference on Information and Knowledge
Management(ICIKM),2012.
the betterment of the user. The Eigenface algorithm
[9] Chen, Joy Iong Zong. "Smart Security System for Suspicious
used here is used for the comparing the face of the user Activity Detection in VolatileAreas." Journal of Information
with the face in the database. Machine learning is used Technology 2, no. 01 (2020): 64-72.
to train (used in OpenCV inbuilt face recogniser) the [10] Suma, V. "Computer vision for human-machine interaction-
review." Journal of trends inComputer Science and Smart
face recogniser. The Adaboost face recognition technology (TCSST) 1, no. 02 (2019): 131-139.
algorithm has the success rate of 75% but the eigenface [11] Face Recognition using Eigenface algorithm on Laptop camera by
Algorithm produces the success rate of 80%. Rika Rosnelly, Mutiara S. Simanjuntak, Ade Clinton sitepu,
Mulkan Azhari, Sandy Kosasi in 2020 8th International Conference
The main limitation of this system is that it requires on Cyber and IT service Management.
periodic maintenance of the cameras. Twins can be an [12] Comparison of the performance of Eigen face and Fisher Face
Algorithm in the face recognition process by Delpiah
exception in this system. In rare cases, photos can be
Wahyuningsih, Chanda Kirana, Rahmat Sulaiman, Hamidah,
used to bypass the security. The future scopes of this Triwanto in 2019 7th International Conference on Cyber and IT
method are that the use of high-quality durable service Management(CITSM).
cameras. 3-d cameras can be used for the condition of [13] Face Detection based ATM security system using embedded Linux
platform Jignesh J. Patoliya, Miral M Desai in 2017 2nd
twins and photo bypassing. International Conference for Convergence in Technology(I2CT).
[14] Real Time Facial Expression and emotion recognition using
REFERENCES Eigenfaces, LBPH and Fisher Algorithms Shrayan Mukhopadhyay,
Shilpu Sharma 2020 10th International Conference on cloud
[1] J.J.Patoliya, M.M. Desai, "Face Detection based ATM Security
computing, Data Science and Engineering.
System using Embedded Linux Platform ", 2nd International
Conference for Convergence in Technology (I2CT), 2017. [15] Face recognition alert mechanism and idealness for monitoring user
motion by Web Services C. Jayaprakash, V Maheshwari 2016 10 th
[2] M.Karovaliyaa, S.Karediab, S.Ozac, Dr.D.R.Kalbande, "Enhanced
International Conference on Intelligent Systems and
security for ATM machine with OTP and Facial recognition
Control(ISCO).
features",International Conference on Advanced Computing Tech-
nologies and Applications (ICACTA), 2015.
[3] [3]Sivakumar T. 1 , G. Askok 2 , k. S. Venuprathap, "Design and .
Implementation of Security Based ATM theft Monitoring system",
International Journal of Engineering Inventions, Volume 3, Issue
1, 2013.
978-0-7381-1327-2/21/$31.00 ©2021 IEEE 1256
WK,QWHUQDWLRQDO&RQIHUHQFHRQ$GYDQFHG&RPSXWLQJ &RPPXQLFDWLRQ6\VWHPV ,&$&&6
Autonomous Face Detection System from Real-time

Video Streaming for Ensuring the Intelligence
Security System
Mohammad Ashraful Hoque* Thouhidul Islam Tanvir Ahmed Al Amin
Dept. of Computer Sci. and Engg., Dept. of Computer Sci. and Engg., Dept. of Computer Sci. and Engg., Dept. of Computer Sci. and Engg.,
Southeast University. Southeast University. Ahsanullah University of Science and Southeast University.
Dhaka, Bangladesh. Dhaka, Bangladesh. Technology (AUST). Dhaka, Bangladesh.
ashraful@seu.edu.bd bappythouhidrx15@gmail.com Dhaka, Bangladesh. alaminwww07@gmail.com
tanvir.cse@aust.edu
Abstract—Face detection is a popular research topic to the to detect a human face or faces, which is coming from the
researchers at present in the Biometric system. The aim of this AdaBoost algorithm.
research is to develop a system which can detect human face or
faces from a live video streaming and warn by giving the siren
for ensuring the significant role in the surveillance security
II. RELATED WORK
aspects. The core of this control system includes ATMega328p Ahmad et al. [3] presents a surveillance system of video-
Microcontroller based Arduino Uno with Pan-Tilt mechanism based face detection and recognition and the analysis of its
and Open Source Computer Vision System (OpenCV) is the evaluated result. Depending on few conditions such as facial
platform for the image processing system. For detecting human emotions, pose and light effects it’s tested from an affluent
face or faces existing methods are Haar–Cascade, Camshift, database of human face. As AdaBoost classifier using Haar
Hausdorff distance, AdaBoost, viola jones, etc. Amongst all of and LBP features and as Support Vector Machine (SVM)
this we used Haar classifier Cascade algorithm for face classifier using Histogram of Oriented Gradients (HOG)
detection in this research. features; it summarized the datasets between AdaBoost and
LBP.
Keywords—Face detection, OpenCV, Processing, Arduino,
ATmega328P, Pan and Tilt mechanism, Servo motor, Haar- Rady [4] presents a face recognition system which helps
Cascade, Face detection algorithm. to identify and verify a person. For verification system it
used Principal Component Analysis (PCA) with different
I. INTRODUCTION distance classifier. Olivetti Research Lab (ORL) face
database’s test result used to produce the result of this
In 21st century, security is a pivotal issue. If we can track
system. These results are produced from the recognition
a human face from a live video streaming, we will be aware
success rate and robustness of that particular face recognition
in a no-man’s land where human movement is restricted such
algorithm. This system used different classifiers such as The
as border area, central jail, bypass road, airport and so on.
Euclidian Distance Classifier, The City Block Distance
This research consists of a combination of hardware and Classifier, The Squared Chebyshew Distance Classifier and
software. As per computer vision system, we do use The Squared Euclidian Distance Classifier. The different
OpenCV version 1.0. OpenCV provides built-in functions for classifiers were executed a comparative recognition rate with
face tracking as well as face detection and recognition. We each and other.
are using JAVA programming language to use the OpenCV
libraries. Servo motors are functioning here as a control
system to balance the camera according to X and Y
coordinates. They are mechanically joined with Pan Tilt III. PROPOSED SYSYTEM
servo bracket. Piezo buzzer and Relay module are used as a The main objectives of this system are to detect the
warning instrument. Servo motors, Piezo buzzer and Relay human face/faces via any type of camera (such as webcam,
module are controlled by Arduino microcontroller. PS4 Camera etc.) and can give the warning by a peizo buzzer
In order to read the video, webcam is used, which is as well as a light bulb will turn on automatically so that it
connected to windows OS-based computer system. Using will be noticeable from a great distance. This camera can
serial communication through Arduino; servo will be moved, change its position according to the human movement. For
where webcam is to be fixed on a base. changing camera position we used here Pan Tilt mechanism.
In this mechanism one pair of MG996R Servo Motor was
For face detection there are few popular methods. used. The flow chart of this system is given in figure 1.
Knowledge based approach, Statistics-based approach and
integration methods are mostly used. The knowledge based For image processing we used Processing software. For
method is used for the images which have complex implementing haar classifier cascade algorithm, we used
background, though it requires more integration to achieve OpenCV’s built in library functions. We used we camera
the flexibility [1]. After judging all possible areas of an which read the input. USB serial communication was
image with the help of classifier, Facial part will detect on completed using Arduino Uno. Pan-tilt mechanism was
statistics based method. In this method classifier is formed working with the help of servo motor. Whole experimental
by using plenty of “Face” and “non-Face” training sample setup is given in figure 2.
[2]. In this research we will use the Haar-Cascade algorithm
978-1-7281-5197-7/20/$31.00 ©2020 IEEE 261

Under the open source BSD license its absolutely free for
using. [6]. Its library and interface was written in C. Since
version 2.0 it includes a new C++ interface with the
traditional C interface [7].
B. Processing
Processing is a java based text programming language.
It’s specially designed for generating and modifying images.
It also helps to acquire a balance between the clearness and
advanced features. It simplifies the training methods for
computer graphics activities such as vector/raster drawing,
color models and image processing etc. Its enriched library
system extends the ability of generating sound, sending and
receiving data in diverse formats. From source it can import
or export 2D and 3D file formats. [8]
C. Arduino
Arduino is a micrcontroller based physical
programable circuit board, which is capable to transfer the
data from arduino board to the output device. It’s formed
with I-I-C bus system for intra board communication.
Arduino board is programmed over RS232 with the help of
serial interface connections. For writing and debuggnig
prgrammes it has own IDE named as Arduino IDE. This
IDE supports the stanard library functions of C and C++ and
C/C++ programming language to write a chunk of code. It
has dfferent types of storage (memory) system such as flash
memory, EEPROM, SRAM etc. Based on the processor
speed, size, memory spaces, I/O ports and using purposes
Fig. 1. Flowchart of proposed system. there is a variation in Arduino board. Arduino UNO,
Arduino Leonardo, Arduino Mega, Arduino Duemilanove,
Arduino Nano, LilyPad Arduino, and Micro Arduino are
most commonly used [9]. Among of them we used Arduino
Uno; which is ATmega328P microcontroller based
development board. It has 14 digital I/O ports, 6 analogue
Input ports, 16 MHz quartz crystal, ICSP header for
ATmega328.
V. FACE DETECTION
In this section, we will discuss about the algorithm,
which is used to detect the human faces.
A. AdaBoost Algorithm
AdaBoost stands for Adaptive Boosting. It was firstly
formulated by Yoav Foreund and Robert Schapire in 1995.
Fig. 2. Bock Diagram of entire system AdaBoost is a machine learning meta-algorithm with the
sensitivity for noisy data and outliners. In many times this
algorithm is further used with other algorithms to augment
their optimal performance [10]. This algorithm is given
IV. DESCRIPTION OF TOOLS below,
This section will cover detail description of the hardware
and software methodologies to implement the face detection
using OpenCV are given.
A. OpenCV
OpenCV stands for Open Source Computer Vision. It’s a
type of llibrary functions introduced for real time computer
vision by Intel and supported by Willow Garage [5].
OpenCV’s library is a cross platform which is focusing on
real-time image processing. OpenCV is also portable for
some specific platforms such as Digital Signal Processors.
978-1-7281-5197-7/20/$31.00 ©2020 IEEE 262
First picture seems to focus on the properties of eyes and

a. Input: Give sample set
second picture focus on the properties of nose [11].
s = ( xi − y i )........( x n = y n )
x i∈ X , y i ∈ Y = { - 1, + 1} Number of iterations T
b. Initialize: Wi, j =
1
, i = 1, 2, …… N
N
c. For t = 1, 2 , …………. T
i) Train weak classifier using distribution W

ii) Calculate the weight (w) training error for each hypothesis
N
hn ∈n = ¦Wi, j | hi − yi | Fig. 4. Haar-like features (Extended).
i =1
iii) Set VI. EXPERIMENTAL SETUP
1 § 1− ∈t · For completing internal circuitry design, we used
a t = log¨¨ ¸¸ solderless breadboard instead of a custom PCB. Power
2 © ∈t ¹ source of the Arduino comes from the 9-volt DC adapter.
iv) Update the weights Schematic diagram of the internal circuit is given in figure 5.
W t =1 , i = 1 +
W i ,t
Zt
× { e − at
e at
W t , i exp (− a t y i h t ( x i ) )
=
Zt
Zt is normalization constant
d. Output: the final hypothesis, also the stronger classifier.

§ T ·
H ( x) = sign ¨ ¦ at ht ( x ) ¸
© t =1 ¹
B. Haar Feature-based Cascade Classifiers

Among of all other object detection method Haar
feature-based cascade classifier is the most effective Fig.5. Schematic Diagram of circuits
method. For training the classifier of this algorithm need a
lot of image of faces and without faces. Image of faces are
called ‘Positive Image’ and image of without faces are VII. RESULT ANALYSIS AND DISCUSSION
called ‘Negative Image’. The value of each feature is
calculated by doing the subtraction between the sum of We kept this system under observation for 128
Pixels of white rectangle and black rectangle [11]. Haar faces within 300 minutes. It was operated on three different
feature is shown in figure 3. conditions; normal, light variations and size and shapes
variations. Normal refers to the natural lighting condition;
here is no extra lighting effects and shape variations due to
the facial expression. Light variations refer to the absence of
natural light and extra lighting effects. Size and shape
variations refer to the different sizes of faces and emotional
effects of faces. After the careful observation, we made a
comparison of Success and Error rates using the formula 1
and 2 as given below,
y × 100 ; y = {all positive integers} (1)

x =
128
Where, x is the success rate of this system and y is the

Fig. 3. Haar like features. number of perfectly detected images of human face or faces.
Figure 4 shows the feature calculation from the
sum of the pixels between white and black rectangle. First
two pictures of the first row show the two good features.
978-1-7281-5197-7/20/$31.00 ©2020 IEEE 263
β × 100 ; ȕ = {all positive integers} (2)

α =
128
Where, Į is the error rate of this system and ȕ is the number
of images which couldn’t detect human face or faces.
This calculation was evaluated for each detected face from
itself among all images within the time and this evaluated
result is shown in table 1.
TABLE I. SUCCESS AND ERROR RATES OF THE SYSTEM.

Condition Success Rate (%) Error Rate (%)
Normal 92.188 7.813
Light Variations 75.656 27.344
Size and Shape Variations 83.594 16.406
The success and failure ratio within that dedicated time

period (300 min.) was represented in a graphical Fig.7. Face detection in normal condition.
representation as given in figure 6.
Fig.6. Graphical representation of table 1.
Figure 6 represents the line graph of Face Detection Rate.

The success rate for normal lighting condition is higher than
those of other conditions. And error rate is lower than those Fig.8. Multiple faces detection in normal condition.
of the other conditions. For changing the lighting condition,
success rate is decreased gradually to 75.656 and error rate is
increased gradually to 27.344. For images, shape and size
variations success rate is plummeted rapidly to 83.594 and
error rate is rocketed rapidly to 16.406.
From live video streaming of this system we took some
screenshots as following in figure 7, 8, 9, 10, 11 and 12.
Fig.9. Multiple faces detection in light variations.
978-1-7281-5197-7/20/$31.00 ©2020 IEEE 264

VIII. LIMITATINS AND CONCLUSION

In this system, video quality absolutely depends on
the camera. For taking a better quality and noise free image
from a real-time video streaming excellent camera should be
used. Depending on the camera size, angular position and its
heaviness; servo motor’s model could be varied. From the
result of the analysis, it is proved that system’s performance
will be affected by lighting condition. Therefore, this system
should cover up the requirement of daylight instead of
darkness.
In this research, well known face detection
algorithm (Haar classifier Cascade) was implemented to
establish this system with the help of OpenCV and
microcontroller on various conditions. In future, this project
will be extended for other face detection algorithms; with
facial recognition from a dedicated database to compare
their performances on the same conditions.
Fig.10. Face detection in light variations.

REFERENCES
[1] A. Faizi, “Robust Face Detection using Template Matching
Algorithm”, University of Toronto, Canada, 2008.
[2] Z. Zhang, “Implementation and Research of Embedded Face
Detection using Adaboost”, Shanghai JiaoTong University, China,
2008.
[3] F. Ahmad, A. Najam and Z. Ahmed, “Image-based Face Detection
and Recognition: State of the Art”, IJCSI International Journal of
Computer Science Issues, Vol. 9, Issue. 6, No. 1, 2013, pp. 169-172.
[4] H. Rady, “Face Recognition using Principle Component Analysis
with Different Distance Classifiers”, IJCSNS International Journal of
Computer Science and Network Security, Vol. 11, No. 10, 2011, pp.
134–144.
[5] CH. Y. Lu, CH.SH. Zhang & F. Wen, “Regional Feature based Fast
Human Face Detection”, J. Tsinghua Univ. (Sci. and Tech.), China,
Vol. 39, 1999, pp. 101–105.
[6] Open Source Computer Vision Library Reference Manual –intel
[7] Gary Bradski& Adrian KaehlerOெReilly, “Learning OpenCV”,
O’REILLY Media, 2008.
[8] Ben Fry & Casey Reas, “Processing: A Programming Handbook for
Visual Designers and Artists”, MIT, 2007.
[9] R. H. Sudhan, M. G. Kumar, A. U. Prakash, S. A. R. Devi and P.
Sathiya, Arduino ATmega-328 Microcontroller”, International
Journal of Innovation Research in Electrical, Electronics,
Fig.11. Multiple faces detection in Size and Shape variations for Instrumentations and Control Engineering, Vol. 3, Issue 4, 2015, pp.
27-29.
normal lighting.
[10] K.J. Wang, SH.L. Duan& W.X. Feng, “A Survey of Face Recognition
using Single Training Sample”, Pattern Recognition and Artificial
Intelligence, China, Vol. 21, 2008, pp. 635–642.
[11] P. Viola and M. Jones, “Rapid Object Detection using a Boosted
Cascade of Simple Features”, Accepted Conference on Computer
Vision and Pattern Recognition, 2001, pp. 1-9.
Fig.12. Multipriations for low YDULDWLRQVIRUQRUPDOOLJKWLQJ

265
2017 2nd International Conferences on Information Technology, Information Systems and Electrical Engineering (ICITISEE)
Design of Face Detection and Recognition System

for Smart Home Security Application
Dwi Ana Ratna Wati Dika Abadianto

Department of Electrical Engineering Department of Electrical Engineering
Universitas Islam Indonesia Universitas Islam Indonesia
Yogyakarta, Indonesia Yogyakarta, Indonesia
dwi_ana@uii.ac.id dikaabadianto@gmail.com
Abstract— This research designs face detection and Home security system can be developed by using face
recognition systems for smart home security application. The recognition method. Face is used as a key to access home. By
design is implemented using MyRIO 1900 and programmed using real face, the process of opening the door will be more
using LabVIEW. The connection between myRIO and computer effective and efficient because it just needs to direct a face on
is wifi network. The image of a person is acquired via webcam the camera, so the camera can identify whether the person is
connected to MyRIO using USB cable. The face detection system allowed for coming in or not. By using the face, the level of
is built based on the template matching, while the face security becomes higher because the face can not be duplicated
recognition is based on the principle component analysis. The as well as changed hands.
testing is done to examine the performance of the face detection
in various change of distance, light intensity, light position angles,
person’s accessories and shirt colour. The face detection modul II. LITERATURE REVIEW
has good performance in some conditions as distance between the Previous research was conducted by Rajiv on home
person and the camera is less than 240 cm, person doesn’t use security systems with email-based cameras [1]. The research
accessories that cover part of face, person doesn’t use shirt with produced system that could identify a person based on skin and
colour similar to skin colour, and background colour is difference
face. The system still requires the authorization of the home
from skin colour. While the face recognition system has 80% of
accuracy when it is tested using realtime image. The combination
owner using email so that the system is considered less
with password is needed in order to increase the security level as effective and efficient. Another study was also conducted
it is applied in real smart home security systems. regarding office security systems using LabVIEW-based face
recognition [2]. The system uses a histogram as a feature for
Keywords—face detection; face recognition; template face recognition. The use of histrogram as a feature is
matching; principle component analysis; smart home; security considered to have poor accuracy.
Another study also conducted regarding finger print-based
I. INTRODUCTION door security system [3]. Fingerprint-based security system is
Home security becomes one of the important things that considered less effective because the user is required to put a
must be considered by the community as well as in the smart finger to the sensor. It is also considered to cause the spread of
home systems. Home security system currently used, is a the virus. Systems with finger print can also be duplicated so
conventional home security system, that is a security system that it is rated as having low security. Another study was
with a mechanical system that requires users to always use the conducted regarding face recognition using principal
key to open or close the door. This makes the home security component analysis method (PCA) [4]. Produced a system that
level is low due to several factors, namely: the ease of can recognize faces with variations of pose and orientation.
duplication of keys, the probability of a lost key or changing The research is still done simulation with face database. The
hands, and others. selection of face recognition method should consider accuracy,
time limitations, process speed and availiability. By judging
In addition to the low security side, the use of conventional these elements, face recognition using PCA is selected because
security systems is also considered less effective and efficient. it is really a simplest and easiest approach to implement,
This is because the user is required to open the door by first extremely fast computation time [5].
inserting the key into the key's mother then turning the key in a
certain direction so that the door can be opened. The process of This research designs and implements face detection and
opening the door is long enough, this makes the conventional face recognition system based on PCA algorithm. It is realized
security system becomes ineffective and inefficient. Therefore, in MyRIO 1900 microcontroller which is programmed using
it is needed a home security system which is more effective, LabVIEW. The design is proposed for application in the smart
efficient and has a high level of security. home security system.
978-1-5386-0658-2/17/$31.00 ©2017 IEEE 342

III. METODOLOGY Fig. 3 shows the state diagram of the software in MyRIO.
The components of face recognition system are image First, the system begins with process of learning using dataset
acquisition, face detection and face recognition as shown by which consist of images of authorized persons in the home
Fig. 1. This system is implemented using hardware as security systems. The learning phase is only execute once,
presented in Fig. 2. Camera is a webcam used as image except there is a new dataset added or template modification. In
acquisition device. MyRIO 1900 is main controller which normal operation, the system will execute the face detection
contains the software of image acquitision, face detetction, as and face recognition continuously.
well as face recognition. Personal computer (PC) is applied as
user interface, image display and monitoring. Both MyRIO and
PC is programmed using LabVIEW.
Fig. 1. Steps in face recognition system for smart home security application.
Fig. 2. Hardware implementation.
Start
Open
Initialization Open Dataset
Template
Modify
Dataset Learning
Modify
Template Change dataset is true
Face
Detection
Exit is false
Change template is true
Face Recognition
Exit is true
Exit
Finish
Fig. 3. State diagram of face recognition software. Fig. 4. Flowchart of face detection.
343
Fig. 4 presents the flowchart of face detection modul. The Start

face detection is developed based on the template matching
algorithm. Each image will be compared to template. If the
matching score is greater than the minimum score, then the Read Datatest
face is detected. The learning process using principle
component analysis (PCA) is shown by Fig. 5. It can be seen
Change to Grayscale
that in the learning process, firstly the dataset colour images
will be changed to grayscale images, then it is resized to
standard size used in the face detection system. Next, the Resize to standard
calculation of the eigenvalues and eigenvectors of the dataset
images. Finally, the dataset images are projected to the
principal component to calculate the euclidean distances of the Change to Vector
dataset images.
Calculate Projection
Calculate Euclidian
Distance to all Dataset
Find the Minimum

Distance
No
Minimum Distance < Face is not
Threshold ? Recognized
Yes
Face is
Recognized
Fig. 6. Flowchart of face recognition.
IV. RESULT AND ANALYSIS

In order to evaluate the performance of the proposed
method of face detection and face recognition, many testing are
done and the results are summarized as follow.
The face detection algorithm based on template matching is
tested on variaous distance of person to camera as shown by
Fig.7 and Table I. It can be seen that the face position in an
image still can be detected when the distance is less than or
equal to 240 centimeters.
Fig. 5. Flowchart of dataset learning.
Fig. 6 shows flowchart of face recognition modul. The

image acquire from the webcam (real images) is processed first
by the face detection modul. After that, the face vector is
projected onto the principal component, then euclidean
distances are calculated between the projected datatest to all
dataset. The minimum distance then is compared to the
threshold. If it is less than the threshold, the face is recognized
as a person in the dataset which means he is authorized to
access the house. Otherwise the face is not recognized and the
person is not an authorized one to access the house. Fig. 7. Results of face detection with various distance of person to webcam.
344
TABLE I. FACE DETECTION PERFORMANCE ON VARIOUS DISTANCE OF

PERSON TO CAMERA
Distance of
Person to Detection Result
Camera (cm)
80 Face is Detected
160 Face is Detected
240 Face is Detected
360 Face is not Detected
Fig. 8 and Fig. 9 presents the samples of face detection

results regarding variation of light source distance and position.
Based on Table II and Table III, the face detection has good
performance in the variation of light source distance position as
well as angle position. The range of the light source distance is Fig. 10. Results of face detection with various accessories.
20 to 80 centimeters, while the angle position is in the interval
of [-90o, 90 o].
TABLE IV. FACE DETECTION PERFORMANCE ON VARIOUS ACCESSORIES
Accessories Detection Result
Glasses Face is Detected
Muslim Cap Face is Detected
Cap Face is hard to be detected
Hijab Face is hard to be detected
Fig. 8. Results of face detection with various light intensity which similar to The proposed face detection method is also tested
various distance of light source to person. considering the variation of accessories used by the person.
Fig. 10 shows the results of face detection regarding the usage
of accessories by the person. The use of glasses and moslem
cap don’t influence the ability of the algorithm to detect the
face position. In other case, the use of cap and hijab cause the
face detection becomes fail as summarized by Table IV.
The colour of shirt used by a person can cause the face
detection fail to detect the face position. It is shown by sample
results presented in Fig. 11. When the person uses shirt with
colour almost similar to the skin colour, it makes the algorithm
more difficult to detect a face. Table V resumes the
performance of face detection for variety of the shirt colour
used by a person.
Fig. 9. Results of face detection with various light source angle positions.
TABLE II. FACE DETECTION PERFORMANCE ON VARIOUS LIGHT

INTENSITY
Distance Light
Source to Person Detection Result
(cm)
80 Face is Detected
60 Face is Detected
40 Face is Detected
20 Face is Detected Fig. 11. Results of face detection with various shirt colour.
TABLE V. FACE DETECTION PERFORMANCE ON VARIOUS SHIRT

TABLE III. FACE DETECTION PERFORMANCE ON VARIOUS LIGHT COLOUR
SOURCE ANGLE POSITION
Shirt Colour Detection Result
Angle Position of Black Face is Detected
Light Source to Detection Result Red Face is Detected
Camera Blue Face is Detected
90o Face is Detected White Face is Detected
45o Face is Detected Green Face is Detected
0o Face is Detected Yellow Face is not Detected
-45o Face is Detected Orange Face is not Detected
-90o Face is Detected
345
After the face position is detected successfully, the next c. The person doesn’t use shirt with colour similar
process is face recognition. In face recognition phase, the to skin colour
euclidean distance of the acquired image face (real image) to d. The background colour is difference from skin
each dataset images will be calculated. The image in the dataset colour
with the minimum euclidean distance is the candidate of 2. The performance of face recognition model is
recognized image, but this minimum distance must be less than determined by the value of minimum euclidean
the threshold of euclidean distance of recognized image. In this distance threshold. Using 0,14x1014 of minimum
study the threshold is 0,14x1014. If the minimum euclidean euclidean distance threshold, the accuracy is 80%.
distance is less than the threshold, the face is recognized as one 3. When the proposed face detection and face
of image in the dataset and he has access to the house, recognition systems are implemented in real smart
otherwise the face is unkown which mean he does not have home security systems, it is needed to combine with
access in the house. Table VI summarized the testing of face the password as additional authentification in order to
recognition system using 40 real images of four subjects taken increase the level of security.
using the webcam, which consists of 30 images of 3 subjects
where are in dataset (authorized person), while the rest 10 TABLE VI. FACE RECOGNITION TESTING RESULTS
images are taken from subject D which is not included in the Subject A
dataset (unauthorized one). Fig. 12 shows the sample of face Number Euclidean
recognition display in testing. According to this testing, the Recognition Results Predicate
Distance
accuracy of the face recognition is 80%, which detailed as 1 0.09 A Correct
follow. From 40 real images, the 32 images are recognized 2 0.04 A Correct
correctly, while 8 images are recognized uncorrectly. 3 0.06 A Correct
4 0.08 A Correct
5 0.08 A Correct
6 0.33 A Correct
7 0.03 A Correct
8 0.02 A Correct
9 0.145 Unknown Uncorrect
10 0.03 A Correct
Subject B
Number Euclidean
Recognition Results Predicate
Distance
1 0.005 B Correct
2 0.01 B Correct
3 0.001 B Correct
4 0.009 B Correct
5 0.0019 B Correct
6 0.002 B Correct
7 0.013 B Correct
Fig. 12. Example results of face recognition with various subject. 8 0.001 B Correct
10 1 Unknown Uncorrect
In order to apply the face detection and face recognition
Subject C
systems in real smart home security systems, the person with Number Euclidean
correct status of face recognition will be authorized to enter a Distance
gate or open a door to get access inside the house. While the 1 0.42 Unknown Uncorrect
unrecognized person is not allowed to access the house. The 2 0.09 C Correct
problem may be arised when the recognition is false, an 3 0.46 Unknown Uncorrect
unauthorized person may enter to the house or the authorized 4 0.02 C Correct
one may not enter to the house. To solve this problem, it needs 6 0.47 Unknown Uncorrect
additional security method such as combination with password 7 0.02 C Correct
of an authorized person. 8 0.05 C Correct
9 0.06 C Correct
V. CONCLUSION
Subject D
This research has designed a model for face detection and Number Euclidean
face recognition for smart home security systems application. Distance
Based on the testing results, it can be concluded as follow. 1 9.1 Unknown Correct
2 9.1 Unknown Correct
1. The face detection model performed well by the 3 9.6 Unknown Correct
conditions: 4 10.1 Unknown Correct
a. The distance between the person and the camera 5 9.2 Unknown Correct
is less than 240 cm. 7 9.9 Unknown Correct
b. The person doesn’t use accessories that cover part 8 9.5 Unknown Correct
of face 9 8.4 Unknown Correct
346
REFERENCES [4] L. C. Paul and A. Al Sumam, “Face Recognition Using Principal

Component Analysis Method,” Int. J. Adv. Res. Comput. Eng. Technol.,
[1] P. Rajiv, R. Raj, and M. Chandra, “Email based remote access and vol. 1, no. 9, pp. 135–139, 2012.
surveillance system for smart home infrastructure ଝ,” Perspect. Sci., pp.
3–5, 2016. [5] R. Kaur and E. Himanshi, “Face recognition using Principal Component
Analysis,” in Advance Computing Conference (IACC), 2015 IEEE
[2] S. R. K, T. C. Sarma, and K. S. Prasad, “Face Recognition Office International, 2015.
Security System Using LabVIEW 8.6,” Int. J. Electron. Commun.
Instrum. Eng. Res. Dev., vol. 3, no. 2, p. 8, 2013.
[3] A. A. Shankar, P. R. K. Sastry, A. L. V. Ram, and A. Vamsidhar,
“Finger Print Based Door Locking System,” nternational J. Eng.
Comput. Sci., vol. 4, no. 3, pp. 10810–10814, 2015.
347
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/354249199
Face Detection Open CV Based ATM Security System
Article in International Journal for Modern Trends in Science and Technology · August 2021
DOI: 10.46501/IJMTST0708016
CITATION READS
1 3,828
1 author:
Editor Ijmtst
International Journal for Modern Trends in Science and Technology
384 PUBLICATIONS 208 CITATIONS
SEE PROFILE
All content following this page was uploaded by Editor Ijmtst on 31 August 2021.
The user has requested enhancement of the downloaded file.

Face Detection Open CV Based ATM Security System
Praveena.P1|Savithri.V1|Saratha.R1|Monisha.M1|Ashwini.R2
1 UG Students, Department of C.S.E, Jansons Institute of Technology, Coimbatore, India

2 Assistant Professor, Department of C.S.E, Jansons Institute of Technology, Coimbatore, India
Abstract: Automated Teller Machines are widely used nowadays by people. But It’s hard to carry their ATM card everywhere,
people may forget to have their ATM card or forget their PIN number. The ATM card may get damaged and users can have a
situation where they can’t get access to their money. In our proposal,use of biometrics for authentication instead of PIN and ATM
card is encouraged.Here, The Face ID is preferred to high priority, as the combination of thesebiometrics proved to be the best
among the identification and verificationtechniques. The implementation of ATM machines comes with the issue of beingaccessed
by illegitimate users with valid authentication code.The users are verified by comparing theimage taken in front of the ATM
machine, to the images which are present in the. If the user is legitimate the new image is used to train the model forfurther
accuracy. This system uses openCV to process the image being obtainedandHaar Cascade Classifier to detect the faces in the
image. The face recognitionis done using Local Binary Pattern.
KEYWORDS: Haar cascade, Automatic Teller Machine, Webxel, Machine learning, EigenFaces, Fisher Faces.
DOI of the Article: https://doi.org/10.46501/IJMTST0708016
Available online at: http://www.ijmtst.com/vol7issue08.html
As per UGC guidelines an electronic bar code is provided to seure your paper
Praveena.P; Savithri.V; Saratha.R; Monisha.M and Ashwini.R. Face Detection Open CV Based
To Cite this Article: ATM Security System. International Journal for Modern Trends in Science and Technology 2021, 7,
0708009, pp. 84-89. https://doi.org/10.46501/IJMTST0708016
Article Info. Received: 05 July 2021; Accepted: 28 July 2021; Published: 01 August 2021
Copyright © 2021 International Journal for Modern Trends in Science and Technology, ISSN : 2455-3778 http://www.ijmtst.com
International Journal for Modern Trends in Science and Technology 2021, 7, 0708009 85
INTRODUCTION DESIGN PHASE

An Automatic Teller Machine (ATM) is a The system contains System board as the main
computerized machine that is used to withdraw cash processor. The system is used to contain the dedicated
from a customer’s respective bank account. As financial operating system which is compatible with the System
users prefer ATM for cash withdrawals, cash deposits board. The Human face is captured by the Web Camera
and many other transaction, the banks are focusing a lot which can be directly interfaced with the System board.
over the security of ATMs.Hence ATM should be pro- The monitor displays the mes- sages for user interface.
tected properly from the criminal activities or from any
unwanted things.
Due to rapid development in science and technology,
upcoming innovations are being built-up with strong
security. But on the other hand, threats are also being
posed to destroy this security level. Though
enhancement in automation has made a positive impact
overall, various financial institutions like banks and
applications like ATM are still subjected to thefts and
frauds. The existing ATM model uses a card and a PIN
which gives rise to increase in attacks in the form of
stolen cards, or due to statically assigned PINs,
duplicity of cards and various other threats.Then Figure:System flow
another major problem is hacking of PIN. There are

other fraudulent attacks like eavesdropping, spoofing , RECOGNITION FACE DETECTION:
brute force attacks, blackmail- ing the user. In the worst The algorithm needs a lot of positive images (images of
case there can also be ATM machine Robbery. faces) and negative images (images without faces) to
To overcome these problems,the project ’ATM train the classifier. Then it is required to extract features
Security system based on Face recognition, PIN and from it. Features are nothing but numerical information
OTP’ consists of conventional features ie is Personal extracted from the images that can be used to
Iden- tification Number (PIN) along with additional distinguish one image from another; for example,
features like face recognition and one-time password histogram (distribution of intensity values) is one of the
(OTP) is used. Database holds information about a features that can be used to define several
user’s account details, images of his/her face and a characteristics of an image even without looking at the
mobile number which will improve security to a large image, such as dark or bright image, the intensity range
extent. of the image, contrast, and so on. Using Haar features is
First, the user will come to the ATM machine and a a efficient method for face detection. These features are
live image is captured through the Web Camera just like the convolution kernel. The convolution can be
interfaced with System defining as the ATM system, summarized by locating a Webxel from the image, then
which is compared with the images stored in the crop out a sub-image with the selected Webxel as the
database. If the face is recognised, then the user is center from the source image with the same size as the
notified to type the PIN. If the PIN matches, an OTP will convolution kernel. Cal- culate an element-wise product
be sent to the corresponding registered mobile number. between the values of the kernel and sub- image. Add
If the user correctly enters the OTP, the transaction can the result of the product. Put the resultant value into the
proceed. Therefore, the combination of face recognition new image at the same place where you Webcked up
algorithm,PIN and an OTP drastically reduces the the Webxel location.
chances of fraud. Inorder to obtain better accuracydeep Each feature is a single value obtained by subtracting
learning based linear discrimintclassifica- tion method the sum of thealgorithm needs a lot of positive images
is utilized. And excecuted the same in OS. (images of faces) and negative images (images without
faces) to train the classifier. Then it is required to extract
features from it. Features are nothing but numerical systems was marker points (position of eyes, ears, nose
information extracted from the images that can be used etc.) were used to build a feature vector (distance
to distinguish one image from another; for example, between the points, angle between them etc). The
histogram (distribution of intensity values) is one of the recognition was performed by calculating the euclidean
features that can be used to define several distance between feature vectors of a probe and
characteristics of an image even without looking at the reference image. Some of the latest work on geometric
image, such as dark or bright image, the intensity range face recognition was, a 22-dimensional feature vector
of the image, contrast, and so on. Using Haar features is and experiments on large datasets have shown that
a efficient method for face detection. These features are geometrical features alone may not carry enough
just like the convolution kernel. The convolution can be information for face recognition.
summarized by locating a Webxel from the image, then
crop out a sub-image with the selected Webxel as the METHOD ONE BASED ON MACHINE LEARNING
center from the source image with the same size as the
Machine learning (ML) is the scientific study of
convolution kernel. Cal- culate an element-wise product
algorithms and statistical models that computer systems
between the values of the kernel and sub- image. Add
use to effectively perform a specific task without using
the result of the product. Put the resultant value into the
explicit instructions, relying on patterns and inference
new image at the same place where you Webcked up
instead. It is seen as a sub- set of artificial intelligence.
the Webxel location.
Machine learning algorithms build a mathematical
Each feature is a single value obtained by
model based on sample data, known as "training data",
subtracting the sum of the Webxels under the white
in order to make pre- dictions or decisions without
rectangle from the sum of the Webxels under the
being explicitly programmed to perform the task.
black rectangle. Now, all possible sizes and locations
Machine learning algorithms are used in a wide variety
of each kernel are used to calculate plenty of
of applications, such as email filtering, and computer
features. Each feature calculation, requires to find
vision, where it is infeasible to develop an algorithm of
the sum of the Webxels under the white and black
specific instructions for performing the task Machine
rectangles. The concept of integral image is very
learning uses types of automated algorithms which
useful to solve this. Integral images are those images
learn to predict future decisions and model and model
in which the Webxel value at any (x,y) location is the
functions using data fed to it.
sum of the all Webxel values present before the
The Eigenfaces and Fisherfaces methods took a
current Webxel.
holistic approach to face recognition. Recently various
methods for a local feature extraction emerged. To
avoid the high-dimensionality of the input data only
local regions of an image are de- scribed, the extracted
features are hopefully more robust against partial
occlusion, illumination and small sample size.
Algorithms used for a local feature extraction are Gabor
Wavelets , Discrete Cosinus Transform and Local
Binary Patterns.
Mainly there are three easy steps to computer coding
facial recognition, which are similar to the steps that
Figure:Haar cascade
human brain use for recognizing faces. These steps are:
FACE RECOGNITION: 1.Data Gathering: Gather face data (face images in
this case) of the people you want to identify.
Face recognition is an easy task for humans. Face 2.Train the Recognizer: Feed that face data and
recognition based on the geometric features of face is respective names of each face to the recognizer so
probably the most intuitive approach to face that it can learn.
recognition. One of the first automated face recognition
3.Recognition: Feed new faces of that people and components from the new Webcture. Compare those
see if the face recognizer, just trained before, features with the list of elements stored during training.
recognizes them. Find the ones with the best match. Return the ‘person’
OpenCV has two built-in face recognizers. The names of label associated with that best match component. In
those face recognizers are: EigenFaces and FisherFaces. simple words, it’s a game of matching.However, one
thing to note in above image is that EigenFaces
EIGENFACES FACE RECOGNIZER ALGORITHM algorithm also considers illumination as an important
feature. In consequence, lights and shadows are
In this algorithm, a facial image is a point from a
Webcked up by EigenFaces, which classifies them as
high-dimensional image space and a lower-dimensional
representing a ‘face’.Face recognition Webcks up on
representation is found, where classification becomes
human things, dominated by shapes and shadows: two
easy. The lower-dimensional subspace is found with
eyes, a nose, a mouth.
Principal Component Anal- ysis(PCA), which identifies
the axes with maximum variance. While this kind of
FISHERFACES FACE RECOGNIZER ALGORITHM
transformation is optimal from a reconstruction
standpoint, it doesn’t take any class labels into account. This algorithm is an improved version of the
Imagine a situation where the variance is generated Eigenfaces. Eigenfaces looks at all the training faces of
from external sources, let it be light. The axes with all the people at once and finds principal components
maximum variance do not necessarily contain any from all of them combined. By doing that, it doesn’t
discriminative information at all, hence a classification focus on the features that dis- criminate one individual
becomes impossible. So a class-specific projection with a from another. Instead, it concentrates on the ones that
Linear Discriminant Analysis was applied to face represent all the faces of all the people in the training
recognition. The basic idea is to minimize the variance data, as a whole. Since EigenFaces also finds
within a class, while maximizing the variance between illumination as a useful component, it will find this
the classes at the same time. variation very relevant for face recognition and may
discard the features of the other people’s faces,
This algorithm considers the fact that not all parts of a
considering them less useful. In the end, the variance
face are equally important or useful for face
that Eigen- Faces has extracted represents just one
recognition. Indeed, when you look at someone, you
individual’s facial features. This can be done by
recognize that person by his distinct features, like the
tunningEigenFaces so that it extracts useful features
eyes, nose, cheeks or fore- head; and how they vary
from the faces of each person separately instead of
respect to each other. Focus is on the areas of maximum
extracting them from all the faces combined. In this
change. For example, from the eyes to the nose there is a
way, even if one person has high illumination changes,
significant change, and same applies from the nose to
it will not affect the other people’s features extraction
the mouth. When multiple faces are given, comparison
process.
is done by looking at these areas, because by catching
the maximum variation among faces, they help to The Principal Component Analysis (PCA), which is
differentiate one face from the other. This is how the core of the Eigenfaces method, finds a linear
EigenFaces recognizer works. It looks at all the training combination of features that maximizes the total
images of all the people as a whole and tries to extract variance in data. While this is clearly a powerful way to
the components which are relevant and useful and represent data, it doesn’t consider any classes and so a
discards the rest. These important features are called lot of discriminative information may be lost when
principal components. throwing components away.
So, EigenFaces recognizer trains itself by extracting The Linear Discriminant Analysis performs a
principal components, but it also keeps a record of class-specific dimensionality reduction. In order to
which ones belong to which person. Thus, whenever a find the combination of features that separates best
new image is introduced to the algorithm, it repeats the between classes the Linear Discriminant Analysis
same process as follows: Extract the principal maximizes the ratio of between-classes to within-classes
scatter, instead of maximizing the overall scatter. The processing.The key here is to get a deep Convolutional
idea is simple: same classes should cluster tightly Neural Network(CNN) to produce a bunch of numbers
together, while different classes are as far away as that describe a face (known as face encodings). When
possible from each other in the lower-dimensional two different images of the same person are passed to
representation. the network, the network should return similar outputs
(i.e. closer numbers) for both images, whereas when im-
ages of two different people are passed, the network
should return very different outputs for the two
images. This means that the neural network needs to be
trained to automatically identify different features of
faces and calculate numbers based on that.
CONCLUSION
Facial recognition has proven to be one of the most
secure methods of all biometric systems to a point for
high level security and to avoid ATM robberies and
provide security for ATM.It replaces the traditional
Figure:Face recognition
ATM system. It has advantages such as saves
Precisely, FisherFaces face recognizer algorithm extracts
manufacturing cost of cards and overcomes drawbacks
principal components that differentiate one person from
of the traditional system like carrying the ATM card,
the others. In that sense, an individual’s components
losing of card, fraud calls related to ATM card, etc.With
do not dominate (become more useful) over the
new improved techniques in the field of artificial
others.Below is an image of principal components using
Intelligence that help eliminate more disturbances and
FisherFaces algorithm.
distortions, the rate of effectiveness of the systemcan be
One thing to note here is that FisherFaces only prevents
improved.
features of one person from becoming dominant, but it
still considers illumination changes as a useful feature.
FUTURE SCOPE
But light variation is not a useful feature to extract as it
The real time security applications like inATM security
is not part of the actual face, another face recognizer
systems, military applications, high security companies.
Algorithm must be used.
This can also be used in bank locker access. Lighting
Machine Learning yielded about 75% accuracy for face
provided to the system is a key factor to be taken care
recognition. Moreover the model was very sensitive
of. Usage of high-speed computers can improve the
to lighting conditions.Hence a new method was
efficiency.
designed for face recognition and that is deep learning.
This model gives an accuracy upto 95% . REFERENCES
Deep learning (also known as deep structured 1. J.J.Patoliya, M.M. Desai, "Face Detection based ATM
learning or hierarchical learning) is part of a broader Security System using Embedded Linux Platform ", 2nd
family of machine learning methods based on artificial International Conference for Convergence in Technology
neural networks.In the case of machine learning, the (I2CT), 2017.
algorithm needs to be told how to make an accurate 2. M.Karovaliyaa, S.Karediab, S.Ozac, Dr.D.R.Kalbande,
prediction by providing it with more information, "Enhanced security for ATM machine with OTP and
Facial recognition features",International Conference on
whereas, in the case of deep learning, the algorithm is
Advanced Computing Tech- nologies and Applications
able to learn that through its own data processing. It is
(ICACTA), 2015.
similar to how a human being would identify
3. Sivakumar T. 1 , G. Askok 2 , k. S. Venuprathap, "Design
something, think about it, and then draw any kind of and Implementation of Security Based ATM theft
conclusion.Deep learning interprets data features and Monitoring system", International Journal of Engineering
its relationships using neural networks which pass the Inventions, Volume 3, Issue 1, 2013.
relevant information through several stages of data
4. C. Bhosale, P. Dere, C. Jadhav, "ATM security using face

and fingerprint recognition", Interna- tional Journal of
Research in Engineering, Technology and Science, Volume
VII, Special Issue, Feb 2017.
5. Manoj V , M. Sankar R , Sasipriya S , U. Devi E, Devika T ,
"Multi Authentication ATM Theft Prevention Using
iBeacon", International Research Journal of Engineering and
Technology (IRJET).
6. L. Wang,H. Ji, Y. Shi, " Face recognition using maximum
local fisher discriminant analysis",18th IEEE International
Conference on Image Processing, 2011.
7. K.Shailaja and Dr.B.Anuradha, "Effective Face
Recognition using Deep Learning based Linear
Discriminant Classification ", IEEE International Conference
on Computational Intelligence and Computing Research, 2016.
8. H. R. Babaei, O. Molalapata and A.H.Y Akbar Pandor,
"Face Recognition Application for Au- tomatic Teller
Machines (ATM)", International Conference on Information
and Knowledge Manage- ment (ICIKM), 2012.
9. https://docs.opencv.org/2.4/modules/contrib/doc/facerec/
facerec-tutorial.htmlface- recognition
10. https://www.superdatascience.com/opencv-face-recogniti
on/
11. https://www.rankred.com/face-recognition-algorithms-te
chniques/
View publication stats

Authentic Face Detection and Encryption
for Security Assurance
1 2
Lahari Vaidya Dr.Veena Devi Shasthrimath V
Department of Electronics and Communication Department of Electronics and Communication
NMAMIT, Nitte NMAMIT, Nitte
Karnataka,India Karnataka,India
laharikvaidya@gmail.com veenadevi@nitte.edu.in
Abstract—The rapid evolution of authentic face detection has of huge interest for the users. It has always been an issue of
been witnessed from the past along with the security assurance in personal authentication that needs to be fixed for the purpose of
the field of information technology. Now being a necessary note access control of the info-security in the wider context via
of issue the concept of face detection along with encryption, two physical security [2]. Researchers found that the face detection is
non-connected field, as a discuss are brought together in this an issue that needs to be taken into consideration. In terms of
paper for the purpose of privacy assurance of data. Considering
the security issues of information, which is the topic of concern in appearance, human face has high degree of variability, making it a
lack of confidence and negative impact profound by the user, the dynamic object of study. Application of face detection is found in
method is deployed to solve the problem by using face detection crowd surveillance, video conferencing, biometrics etc. The
along with feature extraction using PCA algorithm and concept of human face detection makes it difficult for computer
encryption of the facial image detected by generating random key vision. Detected face is stored with high level of secrecy and
that is bound to the image. Securing the information and certainty. Assuring that the data is safe [3], is the most important
maintaining the privacy of data is a mainstream topic nowadays. aspect under discussion. The image data consists of properties
To ensure that the privacy of information is not violated by associated with, such as high level of redundancy, bulk
vulnerabilities such as spoofing and tampering there is a need of
high security assurance to the data stored in the system which is
capabilities and also high correlation between the pixels.
the case in biometrics. There is a need of general awareness and Critical infrastructure of nation accounts when the security
widespread knowledge about security and privacy assurance of assurance matters the most for the purpose of maintaining the
data which is an important threshold. The paper presents the confidentiality of the documents of both public as well as
promising method of authentication of user, data security and private sectors and also is the case of high level authentication
operational efficiencies. The novel solution of generating a key of information in the department of defence. The absolute
for the facial image detected is a method of encryption that
maintains the high level secrecy and certainty of the data. Face protection in the communication media is in the transfer of
detection and encryption is tested and the method is found to be image which is a challenging issue. The method involved to
reliable and persistent. maintain the secrecy of images is encryption [4]. Encryption
being a tool for info protection is used along with facial image
Keywords—Face detection, Feature extraction, encryption, PCA detection. The interactive media in the communication is
algorithm, Spoofing, Tampering, Biometrics. widely used nowadays [5]. Sprouting of info-technology in
and around the globe is swiftly carried on. In such a scenario,
expecting the secrecy of the data has reached a level of scope
I. INTRODUCTION and an issue of concern. The protection of images when
Human face detection is the most promising field of image amassed by the users ends up as unreliable repository. This is
processing that has a vast area of research oriented real life an issue of transferring the data over the media which is
applications. In the real world the concept is widely used for unsecure. Now the concerning issue is to provide absolute
the content annotation, access control, profiling and potential protection against the cyber attacks that is the huge threat of
discrimination in the web world. There is always constructive privacy assurance of the image. The intension is to maintain
scope of new inventions in the field of technology which is as the authenticity, confidentiality and secrecy of the image using
vast as galaxy on its own. This leads to the better future. There the method of encryption. A carrier image is used that transfers
has been a supportive development in the field of technology
the original image into cryptic image. It is the revolutionary
by the humans since the beginning of mankind. The motive
change in the technology when practicality of implementing
was in rapid development and also in the advancement of
the ideas in real sense is noted possible. One such example is
technology to ensure the minimization of risk that is prone
along with the new inventions which would make life easier, implementation of face detection concept in the system. It’s a
better and much faster [1]. The main intention of face treat that is evolving with the every attempt made in
detection is to find out the human face in the given input. The implementation of the concept, face detection in the real
Psychological process of locating the human face in the visual world. It is important to assurance the genuine rate of the data
frame is also possible. It is also categorized as a special case considered for the process which is done by authentication of
of object class detection. The Eigen face approach is the method used.
considered as a promising technique of face detection. In the
field of marketing the facial image detection is playing a role
978-1-5386-7706-3/18/$31.00 2018
c IEEE 131
II. LITERATURE SURVEY III. METHODOLOGY
1. Carmen Bisogni, “An encryption approach using A rapid evolution in the field of information technology is
information fusion techniques involving prime numbers being witnessed for increased security assurance [6]. The
and face biometrics” describe the novel solution of novel solution to this issue is brought by considering two non-
bringing together two non-connected areas of study connected areas of face detection and encryption. The
face biometrics and public key cryptography for the Authentic face detection and encryption for security assurance
purpose of guarantee high level secrecy of the data in involves promising benefits of stronger authentication and
the system. improved security when the image is cryptic [7].
The system architecture shown in Fig.1 below explains the
flow of methods being inculcated to achieve the required
2. Rupali Suralka, “A real sense based multilevel security
outcome.
in cloud framework using face recognition and image
processing” presents a multiple authentication for the
security of the system using AES algorithm for
encryption of data and face recognition.
3. Valarmati R, “A biometric encryption using face

recognition system for watch list” explains the
generation of digital key for encryption to account for
increased convenience and security. The paper also
describes the unique method of biometric encryption.
4. Chang Liu, “A face image recognition algorithm based

DFT encryption domain” this paper gives the solution
to the encrypted face recognition, where the encryption
uses DFT domain of approach. Feature vector is used
as the keyword for the purpose of face recognition of
encrypted image.
5. S.V.Tathe, “Human face detection and recognition in

videos” describes the reduced processing time for face
detection that improves the overall system efficiency
and also minimize the human intervention.
6. Rahul Chandrashekaran, “Security assurance using the

face recognition & detection system based on the
neural networks” a paper that proposes the idea of
assuring the security to the facial information by
inculcating the idea of using the self organizing map
(SOM).
7. Ann Cavoukian, “Biometric encryption: Technology

for strong authentication, Security and Privacy” a paper Fig. 1. System architecture
that considers the privacy and security as the focal
point of issue of the data in the communication field. It A. Image Preprocessing
explains the biometric encryption to provide the
Image preprocessing is enhancing the image clarity by
information to the stake holders concerning the security
suppressing the unwanted distortions in the input image
of the system
considered. This is concerned with:
x Brightness of the image pixels.
Conclusion from literature survey:
x Restoration of image with the prior knowledge of
The information considered by referring to the papers
image considered as input.
mentioned above introduces to the new method of
authentic face detection and encryption for the purpose The noise amount present in the data is minimized and the
of security assurance and privacy of the data. quality of the image is improved.
132 4th International Conference on Applied and Theoretical Computing and Communication Technology iCATccT 2018
B. Face detection E. Encryption
The algorithm for this process follows Viola Jones method of It has always been an issue of privacy concern when it is
analysis. The specific uses of this algorithm is that, related to securing the data. The authentic face detected needs
to be secured before there is any threat concerning the
x High rate of detection i.e., robustness. maintenance of the secrecy of the image. This is achieved by
x Reduced false positive rate. encryption that is the unique concept studied under
x The real time application of face detection is possible cryptography. The method implemented consists of the
using this algorithm. original image to which the carrier image is bound to obtain
x It also goals to distinguish between the faces the cryptic image Fig.3 below shows the encryption process
and those of non-faces.
The four prominent stages that are categorized under face

detection are: Haar Feature Selection, Adaboost training,
Cascading Classifiers, Creating an Integral image.
Fig. 3. Encryption process
The method designed accounts for the encryption of the

original image with certain variations that is acceptable. It is
expected that the original image and the carrier image
dimensions need to be same in order to encrypt the image. The
necessary concern that is associated with the method deployed
is public trust and confidence. The benefits of proposed
Fig. 2. Haar feature that shows the similar properties of nose region encryption and authentic face detection can be noted with the
and the darker portions of eyes. minimized threat and risks that the system would be prone.
A human face consists of similarities that match with those of

Haar features. Fig.2 shows some similar looks of face that IV. ADVANTAGES OF PROPOSED METHOD
matches with the Haar characters. The eye region is darker
When the concern is about privacy and security assurance in
compared to the upper cheeks and the nose region is brighter
the communication media the authentic face detection and
than that of eyes.
encryption proves be the best method of implementation
C. Feature Extraction because:
It is a case where the analysis of the data initially considered
measures and develops the features that are informative and x Security against vulnerabilities such as
also non- redundant. This facilitates human interpretation in a spoofing, tampering, masquerade attack etc.
better way. The method explained in this paper is concerned x There is no retention of the template or image.
with dimensional reduction of the image. Irrespective of the x Improved authentication, security assurance.
input image considered the algorithm is restricted to certain x Maintaining Privacy and secrecy.
range of dimensions that is to be considered. Feature x It can be implemented in large scale application and
extraction restricts the amount of data to be taken in order to public domain with required authorization.
reduce the size of the image.
V. APPLICATIONS
D. Image Thresholding
The most eminent verse of image processing is that to x The method can be used in biometric encryption.
isolate the feature of image with various desired portions. x Security concern in high level institutions.
Segmentation of the image is a technique that involves x Features of proposed method can be used in
thresholding. The conversion of image from grayscale into application oriented software.
binary is done by this method. Image pixel is converted to x Social media for privacy assurance.
a binary form which is the threshold effect.
4th International Conference on Applied and Theoretical Computing and Communication Technology iCATccT 2018 133
VI. RESULTS AND DISCUSSION
The authentic face detection and encryption for the security To ensure that the detected data or image is safe and secure the
assurance involves the series of process that is conducted in image is encrypted. For his original image is considered along
order to confirm the privacy pertaining to secrecy maintaining with the carrier image as shown in Fig.6 below. The
of the data. The input image considered is preprocessed dimensions of both original and carrier image is confirmed to
be same.
initially were the clarity of the image is improved. The
brightness of the image is enhanced were each pixels quality is
improved and then the facial structure is detected using viola
Jones algorithm. Fig.4 shows the face detected along with the
number of face detected is also displayed using haar features.
Fig. 6. Original image and carrier image
The encryption process involves binding of carrier image with

that of original image to obtain the cryptic image. It is evident
from the fig.7 that the variation is observed in the cryptic
image when compared to that of original image.
Fig. 4. Face detected in the input image
Now the detected facial is subjected to the series of operation

that is feature extraction and thresholding of image. The main
Purpose of feature extraction under image processing is to
ensure that the dimensionality of the image is reduced. The
input resource is reduced and then thresholded. Fig 5 shows
the thresholded image were the detected face image is
converted from the grayscale to binary form.
Fig .7. Encrypted image.
VII. CONCLUSION AND FUTURE WORK

The authentic face detection and encryption for security
assurance is the method implemented that generalizes the
privacy concerns of the confidential data that requires secrecy
conviction. The proposed method can be improved in terms of
security assurance i.e. working in the area of face recognition
and decryption for high level data authentication and security.
The information security can be improved in the
communication field by strengthening encryption process.
Fig. 5. Thresholded image
134 4th International Conference on Applied and Theoretical Computing and Communication Technology iCATccT 2018
REFERENCES
[1] Gerardo Iovane, Carmen Bisogi, Luigi De maio, Michel Nappi,
“An Encryption Approach Using Information Fusion
Techniques Involving Prime Numbers and Face Biometrics”, in
IEEE Transactions on Sustainable computing, Jan . 2018, pp. 1-
1.
[2] K. S. Warke, Rupali Suralkar, Snehal Pawar, Priya Sonawane,

ShraddhaWani.“A Real Sense Based Multilevel Security In
Cloud Framework Using Face Recognition And Image
Processing”, in Proc. 2nd international conference for
convergence in technology (I2CT), Dec 2017, pp. 531-533.
[3] R. Valarmathi, S. Sridevi Sathiya Priya, P. Karthigai Kumar, N.

M. SivaMangai.“A Biometric Encryption Using Face
Recognition System For Watch List”, in Proc. International
Conference on Information Communication and Embedded
Systems, Chennai 2014, pp. 1-5.
[4] Chang Liu, Jingbing Li, Yucong Duan. “A Face Recognition

Algorithm Based On DFT Encryption Domain”, in Proc. First
International Conference On Electronics Instrumentation &
Information Systems (EIIS), Harbin 2017, pp. 1-6.
[5] S. V. Tathe, A. S. Narote, S. P. Narote, “Face Detection And

Recognition In Videos”, in Proc. IEEE Annual India
Conference (INDICON), Bangalore, 2016, pp. 1-6.
[6] R. Chandrasekaran, "Security Assurance Using Face

Recognition & Detection System Based On Neural
Networks," International Conference on Neural Networks and
Brain, Beijing, 2005, pp. nil9-1106.
[7] Ann Cavoukian, Alex Stoianov, Fred Carter. “Biometric

Encryption: Technology for Strong Authentication, Security
and Privacy” in IFIP International Federation for Information
Processing, Vol 261, Boston, 2008, pp. 57–77.
4th International Conference on Applied and Theoretical Computing and Communication Technology iCATccT 2018 135

IEEE Referance Paper PDF

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

IEEE Referance Paper PDF

Uploaded by

Copyright:

Available Formats

C65:E2C5C2F56E64E:@?2?5&C6G6?E:@?

    N  L 

)96 6IAC6DD:@? @7 A@@=:?8 :D E@ 4@?DE2?E=J 564C62D6 E96

? @FC AC@A@D65 4C65:E 42C5 7C2F5 7C2F5 56E64E:@? DJDE6>

? 2 AC@A@D65 DJDE6> H6 2C6 8@:?8 E@ @G6C4@>6 6I:DE:?8

ENHANCED SECURITY FEATURE OF

Sri Krishna College of Engineering

ARAVINDH R EDWIN RAJ K ABISHEK S

Abstract— Automated Teller Machines also Keywords—ATM, Face-id, Eigenface algorithm,

978-0-7381-1327-2/21/$31.00 ©2021 IEEE 1252

978-0-7381-1327-2/21/$31.00 ©2021 IEEE 1253

enter this secret PIN. If the PIN matches, then the

Result of csae1. The face is authenticated, The PIN

Result of case 2:if the face is matched and the PIN

Result of case 3: If the users friend uses the card,

DRAWBACKS OF THE EXISTING SYSTEM

978-0-7381-1327-2/21/$31.00 ©2021 IEEE 1254

978-0-7381-1327-2/21/$31.00 ©2021 IEEE 1255

978-0-7381-1327-2/21/$31.00 ©2021 IEEE 1256

Autonomous Face Detection System from Real-time

978-1-7281-5197-7/20/$31.00 ©2020 IEEE 261

First picture seems to focus on the properties of eyes and

i) Train weak classifier using distribution W

d. Output: the final hypothesis, also the stronger classifier.

B. Haar Feature-based Cascade Classifiers

y × 100 ; y = {all positive integers} (1)

Where, x is the success rate of this system and y is the

β × 100 ; ȕ = {all positive integers} (2)

TABLE I. SUCCESS AND ERROR RATES OF THE SYSTEM.

The success and failure ratio within that dedicated time

Fig.6. Graphical representation of table 1.

Figure 6 represents the line graph of Face Detection Rate.

Fig.9. Multiple faces detection in light variations.

978-1-7281-5197-7/20/$31.00 ©2020 IEEE 264

VIII. LIMITATINS AND CONCLUSION

Fig.10. Face detection in light variations.

Fig.12. Multipriations for low YDULDWLRQVIRUQRUPDOOLJKWLQJ

Design of Face Detection and Recognition System

Dwi Ana Ratna Wati Dika Abadianto

978-1-5386-0658-2/17/$31.00 ©2017 IEEE 342

Fig. 2. Hardware implementation.

Fig. 4 presents the flowchart of face detection modul. The Start

Find the Minimum

Fig. 6. Flowchart of face recognition.

IV. RESULT AND ANALYSIS

Fig. 5. Flowchart of dataset learning.

Fig. 6 shows flowchart of face recognition modul. The

TABLE I. FACE DETECTION PERFORMANCE ON VARIOUS DISTANCE OF

Fig. 8 and Fig. 9 presents the samples of face detection

TABLE II. FACE DETECTION PERFORMANCE ON VARIOUS LIGHT

TABLE V. FACE DETECTION PERFORMANCE ON VARIOUS SHIRT

REFERENCES [4] L. C. Paul and A. Al Sumam, “Face Recognition Using Principal

Face Detection Open CV Based ATM Security System

The user has requested enhancement of the downloaded file.

1 UG Students, Department of C.S.E, Jansons Institute of Technology, Coimbatore, India

DOI of the Article: https://doi.org/10.46501/IJMTST0708016

Available online at: http://www.ijmtst.com/vol7issue08.html

INTRODUCTION DESIGN PHASE

another major problem is hacking of PIN. There are

extent. of the image, contrast, and so on. Using Haar features is

algorithm,PIN and an OTP drastically reduces the the Webxel location.

4. C. Bhosale, P. Dere, C. Jadhav, "ATM security using face

C65:E2C5C2F56E64E:@?2?5&C6G6?E:@?

N L

? @FC AC@A@D65 4C65:E 42C5 7C2F5 7C2F5 56E64E:@? DJDE6>

? 2 AC@A@D65 DJDE6> H6 2C6 8@:?8 E@ @G6C4@>6 6I:DE:?8

Fig.12. Multipriations for low YDULDWLRQVIRUQRUPDOOLJKWLQJ