scispace - formally typeset
Open AccessJournal ArticleDOI

Fast indexing and visualization of metric data sets using slim-trees

Reads0
Chats0
TLDR
The slim-tree is a metric access method that tackles the problem of overlaps between nodes in metric spaces and that allows one to minimize the overlap, and how to improve the performance of a metric tree through the proposed "slim-down" algorithm is shown.
Abstract
Many recent database applications need to deal with similarity queries. For such applications, it is important to measure the similarity between two objects using the distance between them. Focusing on this problem, this paper proposes the slim-tree, a new dynamic tree for organizing metric data sets in pages of fixed size. The slim-tree uses the triangle inequality to prune the distance calculations that are needed to answer similarity queries over objects in metric spaces. The proposed insertion algorithm uses new policies to select the nodes where incoming objects are stored. When a node overflows, the slim-tree uses a minimal spanning tree to help with the splitting. The new insertion algorithm leads to a tree with high storage utilization and improved query performance. The slim-tree is a metric access method that tackles the problem of overlaps between nodes in metric spaces and that allows one to minimize the overlap. The proposed "fat-factor" is a way to quantify whether a given tree can be improved and also to compare two trees. We show how to use the fat-factor to achieve accurate estimates of the search performance and also how to improve the performance of a metric tree through the proposed "slim-down" algorithm. This paper also presents a new tool in the slim-tree's arsenal of resources, aimed at visualizing it. Visualization is a powerful tool for interactive data mining and for the visual tracking of the behavior of a tree under updates. Finally, we present a formula to estimate the number of disk accesses in range queries. Results from experiments with real and synthetic data sets show that the new slim-tree algorithms lead to performance improvements. These results show that the slim-tree outperforms the M-tree by up to 200% for range queries. For insertion and splitting, the minimal-spanning-tree-based algorithm achieves up to 40 times faster insertions. We observed improvements of up to 40% in range queries after applying the slim-down algorithm.

read more

Content maybe subject to copyright    Report

!"#$ %&'()*&+ "&' ,*#-".*/"$*0& 01
2($3*4 5"$" 6($# 7#*&+ 6.*89:3((#
;"($"&0 :3"*&" <3=> !"#$"%> &'''>?+8":3"*&">!"#$"%> &''' ()#*+,"% -)./",0>
;@3*#$0# !".0-$#0#> !"#$"%> &'''>"&'A(3&@"3'6((+(3>!"#$"% > &''' ()#*+,"% -)./",0
!"#$%&'$B2"&C 3(4(&$ '"$"D"#( "EE.*4"$*0&# 8-#$ '(". F*$@ #*8*."3*$C G-(3*(#= !03 #-4@ "EE.*4"$*0&#> *$ *# *8E03$"&$ $0 8("#-3( $@(
#*8*."3*$C D($F((& $F0 0DH(4$# -#*&+ $@( '*#$"&4( D($F((& $@(8= !04-#*&+ 0& $@*# E30D.(8> $@*# E"E(3 E30E0#(# $@( 6.*89$3((> " &(F
'C&"8*4 $3(( 103 03+"&*/*&+ 8($3*4 '"$" #($# *& E"+(# 01 1*)(' #*/(= :@( 6.*89$3(( -#(# $@( $3*"&+.( *&(G-".*$C $0 E3-&( '*#$"&4(
4".4-."$*0&# &(('(' $0 "&#F(3 #*8*."3*$C G-(3*(# 0I(3 0DH(4$# *& 8($3*4 #E"4(#= :@( E30E0#(' *&#(3$*0& ".+03*$@8 -#(# &(F E0.*4*(# $0
#(.(4$ $@( &0'(# F@(3( *&408*&+ 0DH(4$# "3( #$03('= J@(& " &0'( 0I(31.0F#> $@( 6.*89$3(( -#(# " 2*&*8". 6E"&&*&+ :3(( $0 @(.E F*$@
$@( #E.*$= :@( &(F *&#(3$*0& ".+03*$@8 .("'# $0 " $3(( F*$@ @*+@ #$03"+( -$*.*/"$*0& "&' *8E30I(' G-(3C E(31038"&4(= :@( 6.*89$3(( *# $@(
1*3#$ 8($3*4 "44(## 8($@0' $0 $"4K.( $@( E30D.(8 01 0I(3."E D($F((& &0'(# *& 8($3*4 #E"4(# "&' $0 E30E0#( " $(4@&*G-( $0 8*&*8*/( *$=
:@( E30E0#(' L1"$91"4$03M *# " F"C $0 G-"&$*1C F@($@(3 " +*I(& $3(( 4"& D( *8E30I(' "&' ".#0 $0 408E"3( $F0 $3((#= J( #@0F @0F $0
-#( $@( 1"$91"4$03 $0 "4@*(I( "44-3"$( (#$*8"$(# 01 $@( #("34@ E(31038"&4( "&' ".#0 @0F $0 *8E30I( $@( E(31038"&4( 01 " 8($3*4 $3((
$@30-+@ $@( E30E0#(' L6.*89'0F&M ".+03*$@8= :@*# E"E(3 ".#0 E3(#(&$# " &(F $00. *& $@( "3#(&". 01 3(#0-34(# 01 6.*89$3(( "*8(' "$
I*#-".*/*&+ *$= ,*#-".*/"$*0& *# " E0F(31-. $00. 103 *&$(3"4$*I( '"$" 8*&*&+ "&' 103 $@( I*#-". $3"4K*&+ 01 $@( D(@"I*03 01 " $3(( -&'(3
-E'"$(#= !*&"..C> F( E3(#(&$ " 1038-." $0 (#$*8"$( $@( &-8D(3 01 '*#K "44(##(# *& 3"&+( G-(3*(#= N(#-.$# 1308 ()E(3*8(&$# F*$@ 3(".
"&' #C&$@($*4 '"$" #($# #@0F $@"$ $@( &(F ".+03*$@8# 01 $@( 6.*89$3(( .("' $0 E(31038"&4( *8E30I(8(&$#= :@(#( 3(#-.$# #@0F $@"$ $@(
6.*89$3(( 0-$E(31038# $@( 29$3(( -E $0 OPP E(34(&$ 103 3"&+( G-(3*(#= !03 *&#(3$*0& "&' #E.*$> $@( 2*&*8".96E"&&*&+9:3((9D"#('
".+03*$@8 "4@*(I(# -E $0 QP $*8(# 1"#$(3 *&#(3$*0&#= J( 0D#(3I(' *8E30I(8(&$# -E $0 QP E(34(&$ *& 3"&+( G-(3*(# "1$(3 "EE.C*&+ $@(
6.*89'0F& ".+03*$@8=
()*+, -+%.#B2($3*4 '"$"D"#(#> 8($3*4 "44(## 8($@0'#> *&'() #$3-4$-3(#> 8-.$*8('*" '"$"D"#(#> #(.(4$*I*$C (#$*8"$*0&> #*8*."3*$C
#("34@=
!
/(0-12345-(20
!
"# $% &'( )(* +(,(-+.' /0+(.&0$), %$+ /-&-1-,( 2-)-3(4
2()& ,5,&(2, 6789:,; 0, &'( '-)/<0)3 $% (=$&0. /-&-
&5>(,? ,@.' -, 02-3(, 6,&0<< $+ /5)-20.;? ,$@)/, 0) &'(0+
2-)5 %$+2-&,? '5>(+&(=&,? >+$&(0) ,(A@().(,? %0)3(+>+0)&,?
(&.B C'(,( D0)/, $% /-&- -+( &5>0.-<<5 2@<&02(/0- /-&- -)/ 0&
0, )$& ,02><( &$ ,&-&( &'(0+ /02(),0$)-<0&5B E'0<( &'(+( '-,
1(() - <-+3( )@21(+ $% >+$>$,-<, %$+ 2@<&0/02(),0$)-<
-..(,, 2(&'$/, FGHI? -<2$,& )$)( $% &'(2 -+( ->><0.-1<( &$
2@<&02(/0- /-&-1-,(, ,0).( &'(5 -,,@2( &'-& /-&- 1(<$)3 &$
- 2@<&0/02(),0$)-< J(.&$+ ,>-.(B K$*(J(+? &'( /-&- $%
2@<&02(/0- /-&-1-,(, $%&() -+( )$& 0) J(.&$+ ,>-.(, 1@& 0)
2(&+0. ,>-.(,B C'-& 0,? &'( $)<5 0)%$+2-&0$) -J-0<-1<( -+( &'(
$1L(.&, -)/ - /0,,020<-+0&5 %@).&0$) ,&-&0)3 &'( /0,&-).(
1(&*(() &'( $1L(.&,B
C'0, >->(+ -//+(,,(, &'( >+$1<(2 $% /(,03)0)3 (%%0.0()&
2(&+0. -..(,, 2(&'$/, 69M9;B M 9M9 $+3-)0N(, - <-+3(
2(&+0. /-&- ,(& -<<$*0)3 0),(+&0$),? /(<(&0$),? -)/ ,(-+.'(,B
M 2(&+0. /-&- ,(& 0, - ,(& $% $1L(.&, -)/ - /0,&-).( %@).&0$) !6;
1(&*(() &*$ $1L(.&, "# $B C'( /0,&-).( %@).&0$) !6 ; ,-&0,%0(,
&'( &'+(( +@<(, $% - 2(&+0. ,>-.( 6
,522(&+5O !!"# $"#!!$# ""P
)$))(3-&0J0&5O! %!!"# $" % $#"%# $ &'! !!"# ""#!P-)/
&+0-)3<( 0)(A@-<0&5O !!"# $"&!!"# ("'!!(# $";B Q$),(A@()&<5?
- 9M9 0, )$& >(+20&&(/ &$ (2><$5 >+020&0J( $>(+-&0$),
,@.' -, -//0&0$)? ,@1&+-.&0$)? $+ -)5 &5>( $% 3($2(&+0.
$>(+-&0$)B 9M9, @,@-<<5 ,@>>$+& ,020<-+0&5 A@(+0(,? &'-& 0,?
%&'() *+)%,)- -)/ ')&%)-. '),(/01% *+)%,)-B "(-+(,& )(03'1$+
A@(+0(, -,D %$+ &'( %0+,& 2 )(-+(,& $1L(.&, %+$2 - 30J() $1L(.&?
%$+ (=-2><(? R:(<(.& &'( %0J( )(-+(,& ,&-+, %+$2 &'( ,@)BS
T-)3( A@(+0(, -,D %$+ &'( $1L(.&, *0&'0) - 30J() /0,&-).(
%+$2 - 30J() $1L(.&? %$+ 0),&-).(? RU0)/ &'( ,&-+, &'-& -+(
*0&'0) GV <03'&45(-+, %+$2 &'( ,@)BS W) >-+&0.@<-+? +-)3(
A@(+0(, *0&' - +-/0@, $% N(+$ -+( .-<<(/ 31,'. *+)%,)-4
C'( (%%0.0().5 $% - 9M9 0, /(&(+20)(/ 15 ,(J(+-<
%-.&$+,B U0+,&? ,0).( &'( /-&- ,(& 0, 3()(+-<<5 &$$ <-+3( &$ %0& 0)
2-0) 2(2$+5? $)( 2-L$+ %-.&$+ %$+ (%%0.0().5 0, &'( )@21(+
$% /0,D -..(,,(, + (A@0+(/ %$+ >+$.(,,0)3 A@(+0(, -)/
0),(+&0$),B E( -,,@2( '(+( &'-& - 9M9 $+3-)0N(, /-&- 0)
>-3(, $% %0=(/ ,0N( $) - /0,D -)/ &'-& /0,D -..(,, +(%(+, &$
+(-/ 6*+0&(; $)( >-3( %+$2 - /0,D 0)&$ 2-0) 2(2$+5B
:(.$)/? &'( .$2>@&-&0$)-< .$,& $% &'( /0,&-).( %@).&0$) .-)
1( J(+5 '03' ,@.' &'-& &'( )@21(+ $% /0,&-).( .-<.@<-&0$),
'-, - 2-L$+ 02>-.& $) (%%0.0().5B E( (=>(.&? '$*(J(+? &'-&
&'(+( 0, - ,&+$)3 +(<-&0$),'0> 1(&*(() &'( )@21(+ $% /0,D
-..(,,(, -)/ &'( )@21(+ $% /0,&-).( .-<.@<-&0$),B C'0+/?
OQQ %RRR :N?S6?;:%TS6 TS USTJVR5WR ?S5 5?:? RSW%SRRN%SW> ,TV= XQ> ST= O> 2?N;YZ?[N%V OPPO
! 54 6%&,'& 7%4 &'! 84 6%&,'& &%) 9,./ ./) :)3&%.;)'. 1< 51;3+.)% =>,)'>)#
?',@)%-,.$ 1< =&
A
1 B&+C1 &. =&
A
1 5&%C1-# 54 B1-.&C DDE# FGHDIJKLI =&
A
1
5&%C1-# =B# M%&N,C4 OJ;&,CP Q>&).&'1# &(;&RS,>;>4->4+-340%4
! 54 T&C1+.-1- ,- 9,./ ./) :)3&%.;)'. 1< 51;3+.)% =>,)'>)# 5&%')(,) U)CC1'
?',@)%-,.$# HIII T1%0)- 8@)4# B,..-0+%(# B8 FHVFGJGEKF4
OJ;&,CP >/%,-.1-S>-4>;+4)!+4
! M4 =))()% ,- 9,./ ./) :)3&%.;)'. 1< U&./);&.,>- &'! 51;3+.)% =>,)'>)#
?',@)%-,.$ U&%0+%(# W&'-JU))%9),'J=.%4# :JGHIGV U&%0+%(# X)%;&'$4
OJ;&,CP -))(&%SU&./);&.,Y4?',JU&%0+%(4!)4
U&'+->%,3. %)>),@)! E :)>4 FKKKZ %)@,-)! VF 7+') VIIIZ &>>)3.)! E =)3.4 VIIIZ
31-.)! .1 :,(,.&C [,0%&%$ VH 7+C$ VIIF4
T1% ,'<1%;&.,1' 1' 10.&,','( %)3%,'.- 1< ./,- &%.,>C)# 3C)&-) -)'! )J;&,C .1P
.Y!)S>1;3+.)%41%(# &'! %)<)%)'>) \OOO5= [1( 2+;0)% FFFIHV4
XPQX9Q\Q]ZPOZ^X]=PP ! OPPO %RRR
Authorized licensed use limited to: Carnegie Mellon Libraries. Downloaded on February 21,2010 at 14:58:48 EST from IEEE Xplore. Restrictions apply.

,&$+-3( @&0<0N-&0$) 0, -)$&'(+ 02>$+&-)& %-.&$+? -<&'$@3' 0&
'-, +-+(<5 1(() >+(J0$@,<5 .$),0/(+(/ 0) &'0, .$)&(=&B C'(
+(-,$) *( -+( .$).(+)(/ -1$@& ,&$+-3( @&0<0N-&0$) 0, )$&
1(.-@,( $% &'( ,&$+-3( .$,&? 1@& >+02-+0<5 1(.-@,( $% &'(
)@21(+ $% /0,D -..(,,(, +(A@0+(/ &$ -),*(+ R<-+3(S +-)3(
A@(+0(,B U$+ &'$,( A@(+0(,? &'( )@21(+ $% -..(,,(, 0, <$*
$)<5 *'() &'( ,&$+-3( @&0<0N-&0$) 0, ,@%%0.0()&<5 '03'B W)
$&'(+ *$+/,? - 9M9 2-5 1( ,<$*(+ &'-) - ,02><(
,(A@()&0-< ,.-) %$+ ,@.' .-,(,B
C'( 1-,0. ,&+@.&@+( $% 2(&+0. &+((, -02, &$ >-+&0&0$) &'(
/-&- ,>-.( 0) +(30$), @,0)3 %)3%)-)'.&.,@)- $+ >)'.)%- &$ *'0.'
&'( $&'(+ $1L(.&, 0) (-.' >-+&0&0$) *0<< 1( -,,$.0-&(/B #-.'
>-+&0&0$) '-, - .$J(+0)3 +-/0@, -)/ $)<5 $1L(.&, *0&'0) &'0,
+-/0@, -+( -,,$.0-&(/ &$ &'( +(>+(,()&-&0J(B !@+ >+$>$,(/
:<024&+(( 0, - )(* /5)-20. 9M9B C'( :<024&+((? <0D( $&'(+
2(&+0. &+((,? ,@.' -, 94&+(( FGGI? ,&$+(, &'( /-&- 0) 0&, <(-J(,
-)/ .+(-&(, -) ->>+$>+0-&( .<@,&(+ '0(+-+.'5 $) &$>B C'(
0))$J-&0$), 0) &'( :<024&+(( -+( &'( %$<<$*0)3O U0+,&? 1-,(/
$) &'( 90)02-< :>-))0)3 C+(( 69:C;? - )(* )$/( ,><0&&0)3
-<3$+0&'2 0, >+(,()&(/B C'0, -<3$+0&'2 >(+%$+2, %-,&(+ &'-)
$&'(+ ,><0& -<3$+0&'2, *0&'$@& ,-.+0%0.0)3 ,(-+.' >(+%$+4
2-).(B :(.$)/? - )(* -<3$+0&'2 0, >+(,()&(/ &$ 3@0/(
0),(+&0$) $% )(* $1L(.&, &$ -) ->>+$>+0-&( ,@1&+((B W)
>-+&0.@<-+? $@+ )(* -<3$+0&'2 <(-/, &$ .$),0/(+-1<5 '03'(+
,&$+-3( @&0<0N-&0$)B C'0+/? -)/ >+$1-1<5 &'( 2$,& 02>$+&-)&?
&'( R:<024/$*)S -<3$+0&'2 0, >+(,()&(/ &$ 2-D( &'( 2(&+0.
&+(( &03' &(+ -)/ %-,&(+ 0) - >$,&>+$.(,,0)3 , &(>B C '0,
-<3$+0&'2 *-, /(+0J(/ %+$2 $@+ %0)/0)3, &'-& '03' $J(+<->
0) - 2(&+0. &+(( 0, <-+3(<5 +(,>$),01<( %$+ 0&, 0)(%%0.0().5B
X) %$+&@)-&(<5? &'( *(<<4D)$*) &(.')0A@(, &$ 2(-,@+(
$J(+<-> $% - >-0+ $% 0)&(+,(.&0)3 )$/(, 6(B3B? .0+.<(, 0) -
&*$4/02(),0$)-< ,>-.(; .-))$& 1( @,(/ %$+ 2(&+0. /-&-B
W),&(-/? *( >+$>$,( &'( R-1,$<@&( %-&4%-.&$+S -)/ &'(
R+(<-&0J( %-&4%-.&$+S &$ 2(-,@+( &'( /(3+(( $% $J(+<->B W& 0,
,'$*) &'-& &'( :<024/$*) -<3$+0&'2 +(/@.(, &'( R+(<-&0J(
%-&4%-.&$+S -)/? '().(? 02>+$J(, &'( A@(+5 >(+%$+2-).( $%
&'( 2(&+0. &+((B
M >+(<020)-+5 J(+,0$) $% &'0, *$+D *-, >+(,()&(/ -&
#78C YVVV FYZIB K(+(? *( -<,$ /(,.+01( &*$ )(* &$$<,B C'(
%0+,& &$$< 0, - J0,@-<0N-&0$) -<3$+0&'2 &'-& A@0.D<5 2->, &'(
$1L(.&, $% &'( :<024&+(( $)&$ <$*4/02(),0$)-< >$0)&,? &+50)3
&$ >+(,(+J( &'( /0,&-).(,B E0&' &'0, &$$<? $)( .-) J0,@-<0N(
'$* R3$$/S &'( &+(( 0,? -, *(<< -, /$ J0,@-< /-&- 20)0)3
6(B3B? /(&(.&0)3 .<@,&(+, -)/ $@&<0(+,;B C'( ,(.$)/ &$$< 0, -
%$+2@<- &'-& (,&02-&(, &'( -J(+-3( )@21(+ $% /0,D -..(,,(,
)((/(/ &$ -),*(+ +-)3( A@(+0(,B C'0, %$+2@<- &-D(,
-/J-)&-3( $% &'( R-1,$<@&( %-&4%-.&$+?S ,$ &'( (,&02-&( .-)
1( >(+%$+2(/ $) -)5 &+(( -)/ )$& $)<5 $) 0/(-< $)(,B
C'( +(2-0)/(+ $% &'( >->(+ 0, ,&+@.&@+(/ -, %$<<$*,O W)
&'( )(=& ,(.&0$)? *( %0+,& 30J( - 1+0(% '0,&$+5 $% 9M9,?
0).<@/0)3 - .$).0,( /(,.+0>&0$) $% &'( /-&- ,(&, *( @,(/ 0)
$@+ (=>(+02()&,B :(.&0$) [ 0)&+$/@.(, &'( :<024&+((? -)/
:(.&0$) \ >+(,()&, 0&, )(* ,><0&&0)3 -<3$+0&'2 1-,(/ $)
20)02-< ,>-))0)3 &+((,B :(.&0$) H 0)&+$/@.(, &'( R-1,$<@&(
%-&4%-.&$+S -)/ &'( R+(<-&0J( %-&4%-.&$+S -)/ :(.&0$) ]
>+(,()&, &'( J0,@-<0N-&0$) D0& /(J(<$>(/ %$+ &'( :<024&+((B
C'( :<024/$*) -<3$+0&'2 0, /(,.+01(/ 0) :(.&0$) ^? *'0<(
:(.&0$) Z >+(,()&, &'( /(J(<$>2()& $% - %$+2@<- *'0.'
-<<$*, &'( (,& 02-&0$) $% ,(<(.&0J0&5 %$+ +-)3( A@(+0(,B
:(.&0$) _ >+(,()&, - >(+%$+2-).( (J-<@-&0$) $% &'( :<024
&+(( -)/ :(.&0$) GV 30J(, &'( .$).<@,0$), $% &'0, >->(+B
674189: !03 3!-! 79-7
C'( /(,03) $% (%%0.0()& -..(, , 2(&'$/, '-, 0)&(+(,&( /
+(,(-+.'(+, %$+ 2$+( &'-) &'+(( /(.-/(,B M) (=.(<<()&
,@+J(5 $% 2@<&0/02(),0$)-< -..(,, 2(&'$/, .-) 1( %$@)/ 0)
FGHIB K$*(J(+? 2$,& $% &'(,( -..(,, 2(&'$/, ->><5 $)<5 &$ -
$)( $+ 2@<&0/02(),0$)-< J(.&$+ /-&- ,(&B
9(&+0. /-&- ,(&, '-J( -&&+-.&(/ &'( -&&()&0$) $% +(,(-+.'4
(+,? 2-0)<5 +(3-+/0)3 &'( >+$1<(2 $% ,@>>$+&0)3 )(-+(,&
)(03'1$+ -)/ +-)3( A@(+0(,B C'( >0$)((+0)3 *$+D $%
8@+D'-+/ -)/ `(<<(+ F^I >+$J0/(/ $&'(+ 0)&(+(,&0)3 &(.')04
A@(, %$+ >-+&0&0$)0)3 - 2(&+0. /-&- ,(& 0) - +(.@+,0J( %-,'0$)
*'(+( &'( +(.@+,0J( >+$.(,, 0, 2-&(+0-<0N(/ -, - &+((B C'(
%0+,& &(.')0A@( >-+&0&0$), - /-&- ,(& 15 .'$$,0)3 -
+(>+(,()&-&0J( %+$2 &'( ,(& -)/ 3+$@>0)3 &'( (<(2()&, *0&'
+(,>(.& &$ &'(0+ /0,&-).( %+$2 &'( +(>+(,()&-&0J(B C'( ,(.$)/
&(.')0A@( >-+&0&0$), &'( $+030)-< ,(& 0)&$ - %0=(/ )@21(+ $%
,@1,(&, -)/ .'$$,(, - +(>+(,()&-&0J( %+$2 (-.' $% &'(
,@1,(&,B C'( +(>+(,()&-&0J( -)/ &'( 2-=02@2 /0,&-).(
%+$2 &'( +(>+(,()&-&0J( &$ - >$0)& $% &'( .$++(,>$)/0)3
,@1,(& -+( -<,$ 2-0)&-0)(/B C'( 2(&+0. &+(( $% X'<2-)) FY_I
-)/ &'( a-)&-3(4>$0)& &+(( 6J>4&+((; $% b-)0<$, F[YI -+(
,$2(*'-& ,020<-+ &$ &'( %0+,& &(.')0A@( $% F^I -, &'(5
>-+&0&0$) &'( (<(2()&, 0)&$ &*$ 3+$@>, -..$+/0)3 &$ -
+(>+(,()&-&0J(? .-<<(/ - RJ-)&-3( >$0)&BS W) F[YI? &'( J>4&+((
'-, -<,$ 1(() 3()(+-<0N(/ &$ - 2@<&0*-5 &+(( -)/? 0) FZI? 0&
*-, %@+&'(+ 02>+$J(/ &'+$@3' @,0)3 - 1(&&(+ -<3$+0&'2 &$
.'$$,( &'( J-)&-3( >$0)&, -)/ &$ -),*(+ )(-+(,& )(03'1$+
A@(+0(,B W) $+/(+ &$ +(/@.( &'( )@21(+ $% /0,&-).(
.-<.@<-&0$),? 8-(N-4b-&(, (& -<B FGI ,@33(,&(/ @,0)3 &'( ,-2(
J-)&-3( >$0)& 0) -<< )$/(, &'-& 1(<$)3 &$ &'( ,-2( <(J(<B
C'()? - 10)-+5 &+(( /(3()(+-&(, 0)&$ - ,02><( <0,& $% J-)&-3(
>$0)&,B M)$&'(+ 2(&'$/ $% X'<2-)) FY_I 0, &'( 3()(+-<0N(/
'5>(+4><-)( &+(( 63'4&+((;B C'( 3'4&+(( >-+&0&0$), &'( /-&-
,(& 0)&$ &*$ 15 >0.D0)3 &*$ $1L(.&, -, +(>+(,()&-&0J(, -)/
-,,03)0)3 &'( +(2-0)0)3 &$ & '( .<$,(,& +(>+(,()&-&0J(B
8$ND-5- -)/ !
c
N,$5$3<@ F\I? FHI >+$>$,(/ -) (=&(),0$) $%
&'( J>4&+(( .-<<(/ &'( R2@<&0J-)&-3(4>$0)& &+((S 62J>4&+((;?
*'0.' .-+(%@<<5 .'$$,(, ; J-)&-3( >$0)&, %$+ - )$/( *'0.'
'-, - %-)$@& $% )
"
B C'( d($2(&+0. "(-+ M..(,, C+((
6d"MC; $% 8+0) F]I .-) 1( J0(*(/ -, - +(%0)(2()& $% &'(
,(.$)/ &(.')0A@( >+(,()&(/ 0) F^IB W) -//0&0$) &$ &'(
+(>+(,()&-&0J( -)/ &'( 2-=02@2 /0,&-).(? 0& 0, ,@33(,&(/
&'-& &'( /0,&-).(, 1(&*(() >-0+, $% +(>+(,()&-&0J(, 1(
,&$+(/B C'(,( /0,&-).(, .-) 1( @,(/ &$ >+@)( &'( ,(-+.'
,>-.( @,0)3 &'( &+0-)3<( 0)(A@-<0&5B M<,$? -) ->>+$-.' &$
(,&02-&( /0,&-).(, 1(&*(() $1L(.&, @,0 )3 >+(.$2>@&(/
/0,&-).(, $) ,(<(.&(/ $1L(.&, $% &'( /-&- ,(& 0, >+$>$,(/
15 :'-,'- -)/ E-)3 0) FYHIB
M<< 2(&'$/, >+(,()&(/ -1$J( -+( ,&-&0. 0) &'( ,(),( &'-&
&'(5 /$ )$& ,@>>$+& 0),(+&0$), -)/ /(<(&0$),B C'( 94&+(( $%
Q0-..0-? (& -<B FGGI $J(+.$2(, &'0, /(%0.0().5 -<<$*0)3 %@+&'(+
0),(+&0$),B C'( 94&+(( 0, - '(03'&41-<-).(/ &+(( *'(+( &'(
/-&- (<(2()&, -+( ,&$+(/ 0) &'( <(-J(,B M) 0)&(+)-< )$/( 0, -
,(& $% ()&+0(, *'(+( (-.' ()&+5 .$),0,&, $% - >$0)&(+ &$ -
,@1&+((? &'( +$@&0)3 $1L(.&? -)/ &'( .$J(+0)3 +-/0@, % $% &'(
,@1&+((B #-.' )$/( 0, %0<<(/ *0&' -& <(-,& > -)/ -& 2$,&
5 ()&+0(,? *'(+( * & +,"B "$&0.( &'-& &'( 94&+(( ,@>>$+&,
0),(+&0$), ,020<-+ &$ T4&+((, FG^IB #=>(+02()&-< +(,@<&,
.$2>-+0)3 &'( 94&+(( -)/ &'( 2J>4&+(( F_I -)/ &'( 94&+((
-)/ &'( Te4&+(( FGVI -+( >+$J0/(/? -, *(<< -, - .$,& 2$/(<
1-,(/ $) Wf! -)/ /0,&-).( /0,&+01@&0$) %$+ &'( 94&+(( FGYIB
:N?%S? R: ?V=_ !?6: %S5R`%SW ?S5 ,%67?V%a?:%TS T! 2R:N%; 5?:? 6R:6 76%SW 6V%29:NRR6 OQb
Authorized licensed use limited to: Carnegie Mellon Libraries. Downloaded on February 21,2010 at 14:58:48 EST from IEEE Xplore. Restrictions apply.

C'( :<024&+(( -, *(<< -, &'( 94&+(( -+( /5)-20. 9M9,
0) &'( ,(),( &'-& &'(5 -<<$* /5)-20. 0),(+&0$),B 7(<(&0$),
-+( )$& ,@>>$+&(/ 15 :<024&+(( )$+ 15 94&+(( 5(&B "(J(+4
&'(<(,,? /(<(&0$), .-) 1( '-)/<(/ 0) - ,020<-+ *-5 -, 0& 0, 0)
,$2( J-+0-&0$), $% 8g4&+((, *'(+( &'( $1L(.&, 0) &'( 0)/(=
)$/(, .$@</ 1( 2-+D(/ -, R/(<(&(/S FG_IB
E0&' +(,>(.& &$ J0,@-<0N-&0$)? &'(+( -+( ,(J(+-< &$$<, %$+
J0,@-<0N0)3 0)/(= &+((, 0) &'( <0&(+-&@+( FGHIB M) 0)&(3+-&(/
%+-2(*$+D %$+ J0,@-< /(1@330)3 $% -..(,, 2(&'$/, 0,
>+(,()&(/ 0) FY\IB W) &'-& *$+D? &'( &$$< R-2/1S -<<$*, &'(
@,(+ &$ J0,@-<0N( &'( ,&+@.&@+( $% &'( *'$<( &+((? &'( )$/(,?
-)/ (J() &'( /-&- /0,&+01@&0$)B K$*(J(+? 0& /$(, )$&
,@>>$+& 2(&+0. /-&- ,(&,B M) ->>+$-.' &$ J0,@-<0N0)3 -
2(&+0. ,(& 0, &$ @,( - /0,&-).(4>+(,(+J0)3 2->>0)3
-<3$+0&'2B C'0, D0)/ $% -<3$+0&'2 &-D(, - /-&- ,(& -)/ -
/0,&-).( %@).&0$) -)/ 2->, (-.' $1L(.& 0) &'( /-&- ,(& &$ -
>$0)& 0) - Y4/02(),0$)-< ,>-.(? &+50)3 &$ >+(,(+J( &'(
/0,&-).(,B C'(+( -+( 2-)5 /0,&-).(4>+(,(+J0)3 2->>0)3
-<3$+0&'2, 0) <0&(+-&@+(? ,@.' -, &'( RU-,&9->S FG\I? Q$%(
FGZI? *'0.' *-, /(J(<$>(/ &$ 2-> >+$&(0) /-&- ,(&,? -)/
&'( R9(&+0.49->S F[GI? -020)3 &$*-+/ T"M /-&-B W) &'0,
*$+D? *( @,(/ &'( RU-,&9->S -<3$+0&'2 L@,& %$+ J0,@-<0N-4
&0$) 1(.-@,( 0& >(+%$+2, *(<< -, - 3()(+0. 2->>0) 3
-<3$+0&'2B K$*(J(+? -)5 $% &'( $&'(+ -<3$+0&'2, .$@</
-<,$ 1( (2><$5(/B
W) $+/(+ &$ 0<<@,&+-&( &'( >(+%$+2-).( $% &'( /0%%(+()&
9M9,? *( @,( ,0= ,5)&'(&0. -)/ +(-< /-&- ,(&, &'+$@3'$@&
&'( >->(+B C'( /0,&-).( /0,&+01@&0$) -)/ 0)&+0),0. /02()4
,0$)-<0&5 $% &'(,( /-&- ,(&, -+( >+(,()&(/ 0) FY]IB C'( /-&-
,(&, -+( &'( %$<<$*0)3O
! R:0(+>0),D5Sh- ,5)&'(&0. ,(& $% GV]?YZZ >$0)&, 0) -
&*$4/02(),0$)-< ,>-.( %+$2 &'( :0(+>0),D5 &+0-)3<(
6,(( U03B G-;? - *(<<4D)$*) /-&- ,(&? *'0.' 0, ,(<%4
,020<-+ 6R%+-.&-<S;B
! R9dQ$@)&5Sh- ,(& $% GH?HH_ 3($3+->'0.-< >$0)&, 0)
- &*$4/02(),0$)-< ,>-.( /(,.+010)3 &'( .$$+/0)-&(,
$% &'( +$-/ 0)&(+,(.&0$), 0) 9$)&3$2(+5 Q$@)&5?
9-+5<-)/ 6,(( U03B G1;B
! R#03()%-.(,Sh- ,(& $% GG?_VV %-.( J(.&$+, 6G]4
/02(),0$),; %+$2 &'( W)%$+2(/0- i+$L(.& F[VI -&
Q-+)(30( 9(<<$) X)0J(+,0&5B
! RU-.(W&Sh- /-&- ,(& .$),&+@.&(/ *0&' - /0,&-).(
2-&+0= 30J() 15 &'( U-.(W&! !"#$%&'() *('!+", -./0.
1&2(3$ +! & 2"44('2+&5 6'"782$ #'"4 9+!+",+2!
:"'6"'&$+", &,7 $;(+' 7+!$&,2( #8,2$+", %&! '1.
!,->C1-)!B C'( ,(& $% G?VH] %-.(, &'-& 3()(+-&( &'0,
/0,&-).( 2-&+0= *-, -<,$ 30J() 15 &'( W)%$+2(/0-
i+$L(.&B "$&0.( &'-& &'( /0,&-).( 2-&+0= .$++(,>$)/,
&$ - 2(&+0. %@).&0$)B
! R#)3<0,'E$+/,Sh- ,(& $% YH?G\[ $1L(.&, %+$2 &'(
#)3<0,' <-)3@-3( /0.&0$)-+5 6f@,+f,'-+(f<01f/0.&
%+$2 X)0= ,5,&(2,;B
! Ri$+&@3@(,(E$+/,Sh- ,(& $% \Y_?\[\ $1L(.&, %+$2
&'( i$+&@3@(,( <-)3@-3( /0.&0$)-+5 FYGIB
"$&( (,>(.0-<<5 &'-&? %$+ &'( U-.(W& /-&- ,(&? *( '-J( @,(/ -
/0,&-).( %@).&0$) %+$2 - .$22(+.0-< ,$%&*-+( >+$/@.&B W)
3()(+-<? *( '-J( @,(/ &'( -
"
2(&+0. %$+ &'( J(.&$+ /-&- ,(&,
-)/ -<< $% &'(2 -+( )$+2-<0N(/ 0) - @)0& .@1(B U$+ &'(
R#)3<0,'E$+/,S -)/ Ri$+&@3@(,(E$+/,S /-&- ,(&? &'(
j(J(),'&(0)? $+ ,&+0)3 (/0& /0,&-).( 6-
.!/0
;? *-, @,(/B
-
.!/0
!## $" 0, - 2(&+0. %@).&0$) *'0.' .$@)&, &'( 20)02-<
)@21(+ $% ,521$<, &'-& '-J( &$ 1( 0),(+&(/? /(<(&(/? $+
,@1,&0&@&(/ &$ &+-),%$+2 ,&+0)3 " 0)&$ ,&+0)3 $ 6(B3B? -
.!/0
6R'(-/?S R'$115S; k \ ( &' +(( ,@1,&0&@&0$), -)/ $)(
0),(+&0$);B E( -+( @,0)3 &'( ,-2( @)0& .$,& %$+ (-.' $>(+-&0$)?
-<&'$@3' /0%%(+()& .$,&, .-) 1( -,,$.0-&(/ *0&' (-.' (/0&
$>(+-&0$) *0&'$@& -)5 .'-)3( 0) $@+ >+$>$,(/ 2(&'$/,B
OQc %RRR :N?S6?;:%TS6 TS USTJVR5WR ?S5 5?:? RSW%SRRN%SW> ,TV= XQ> ST= O> 2?N;YZ?[N%V OPPO
!*+= X= :@( $F09'*8(&$*0&". I(4$03 '"$" #($# -#(' *& $@( ()E(3*8(&$#= d"e 6*(3E*&#KC $3*"&+.(= dDe 2W;0-&$C=
Authorized licensed use limited to: Carnegie Mellon Libraries. Downloaded on February 21,2010 at 14:58:48 EST from IEEE Xplore. Restrictions apply.

; -<9 7=(>?-199 @ !0 (>A12893 A91B21>!059
>9-1(5 -199
C'( :<024&+(( 0, - 1-<-).(/ -)/ /5)-20. &+(( &'-& 3+$*,
1$&&$24@> %+$2 &'( <(-J(, &$ &'( +$$&B j0D( $&'(+ 2(&+0.
&+((,? &'( $1L(.&, $% &'( /-&- ,(& -+( 3+$@>(/ 0)&$ %0=(/ ,0N(
/0,D >-3(,? (-.' >-3( .$++(,>$)/0)3 &$ - &+(( )$/(B C'(
$1L(.&, -+( ,&$+(/ 0) &'( <(-J(,B C'( 2-0) 0)&()& 0, &$
$+3-)0N( &'( $1L(.&, 0) - '0(+-+.'0.-< ,&+@.&@+( @,0)3 -
%)3%)-)'.&.,@) -, &'( .()&(+ $% (-.' 20)02@2 1$@)/0)3
+(30$) *'0.' .$J(+, &'( $1L(.&, 0) - ,@1&+((B C'( :<024&+((
'-, &*$ D0)/, $% )$/(,? /-&- )$/(, 6$+ <(-J(,; -)/ 0)/(=
)$/(,B M, &'( ,0N( $% - >-3( 0, %0=(/? (-.' &5>( $% )$/( '$</,
- >+(/(%0)(/ 2-=02@2 )@21(+ $% $1L(.&, 5B U$+ ,02><0.0&5?
*( -,,@2( &'-& &'( .->-.0&5 5 $% &'( <(-J(, 0, (A@-< &$ &'(
.->-.0&5 $% &'( 0)/(= )$/(,B C-1<( G ,@22-+0N(, &'(
,521$<, @,(/ 0) &'0, >->(+B
C'( <(-% )$/(, '$</ -<< $1L(.&, ,&$+(/ 15 &'( :<024&+((?
-)/ &'(0+ ,&+@.&@+( 0,
12&3'4!2 )%&&%$ '( % 5/!
/
#!!6
/
#6
728
"#6
/
9*#
*'(+( 5/!
/
0, &'( 0/()&0%0(+ $% &'( $1L(.& 6
/
-)/ !!6
/
#6
728
" 0,
&'( /0,&-).( 1(&*(() &'( $1L(.& 6
/
-)/ &'( +(>+(,()&-&0J(
$1L(.& $% &'0, <(-% )$/( 6
728
B C'( ,&+@.&@+( $% -) 0)/(= )$/( 0,
/'!2"'4!2 )%&&%$ '( %6
/
#:
/
#!!6
/
#6
728
"#
;07!<6
/
"# =.'07/2>!; 07!<6
/
"" 9*#
*'(+( 6
/
D((>, &'( $1L(.& &'-& 0, &'( +(>+(,()&-&0J( $% &'(
,@1&+(( >$0)&(/ 15 ; 07!<6
/
" -)/ :
/
0, &'( .$J(+0)3 +-/0@,
$% &'-& +(30$)B C'( /0,&-).( 1(&*(() 6
/
-)/ &'( +(>+(,()4
&-&0J( $% &'0, )$/( 6
728
0, D(>& 0) !!6
/
#6
728
"B C'( >$0)&(+
;07!<6
/
" >$0)&, &$ &'( +$$& )$/( $% &'( ,@1&+(( +$$&(/ 15
6
/
B C'( )@21(+ $% ()&+0(, 0) &'( )$/( >$0)&(/ &$ 15
;07!<6
/
" 0, '(</ 15 =2'07/2>!;07!<6
/
""B U03B Y 3+->'0.-<<5
,'$*, &'( :<024&+(( ,&+@.&@+( 6U03B Y-; -)/ -) (=-2><( $% -
&+(( *0&' ,(J() R*$+/S $1L(.&, @,0)3 &'( -
.!/0
/0,&-).(
%@).&0$) 6U03B Y1;B
C'( +(30$), &'-& .$++(,>$)/ &$ (-.' )$/( $% &'( :<024&+((
.-) $J(+<-> (-.' $&'(+B C'( 0).+(-,( $% $J(+<-> -<,$
0).+(-,(, &'( )@21(+ $% >-&', &$ 1( &+-J(+,(/ *'() -
A@(+5 0, 0,,@(/? -, *(<< -, &'( )@21(+ $% /0,&-).(
.-<.@<-&0$), &$ -),*(+ A@(+0(,B C'( :<024&+(( *-, /(J(<4
$>(/ &$ +(/@.( &'( $J(+<-> 1(&*(() +(30$), 0) (-.' <(J(<B
:N?%S? R: ?V=_ !?6: %S5R`%SW ?S5 ,%67?V%a?:%TS T! 2R:N%; 5?:? 6R:6 76%SW 6V%29:NRR6 OQ]
:?AVR X
6-88"3C 01 6C8D0.# "&' 5(1*&*$*0&#
!*+= O= d"e W3"E@*4". 3(E3(#(&$"$*0& 01 $@( #$3-4$-3( 0& *&'() "&' .("1 &0'(# 01 $@( 6.*89$3((= dDe ?& ()"8E.( 01 " #.*89$3(( #$03*&+ #(I(& LF03'#M
-#*&+ $@( -
.!/0
'*#$"&4( 1-&4$*0&#=
Authorized licensed use limited to: Carnegie Mellon Libraries. Downloaded on February 21,2010 at 14:58:48 EST from IEEE Xplore. Restrictions apply.

;C/ DEFG*F)H $I+ 7GF.?-%++
!1L(.&, -+( 0),(+&(/ 0) - :<024&+(( 0) &'( %$<<$*0)3 *-5O
:&-+&0)3 %+$2 &'( +$$& )$/(? &'( -<3$+0&'2 &+0(, &$ <$.-&( -
)$/( &'-& .-) .$J(+ &'( )(* $1L(.&B W% )$)( A@-<0%0(,? ,(<(.&
&'( )$/( *'$,( .()&(+ 0, )(-+(,& &$ &'( )(* $1L(.&B W% 2$+(
&'-) $)( )$/( A@-<0%0(,? (=(.@&( &'( 5/11-)=+0.%)) -<3$+0&'2
&$ ,(<(.& $)( $% &'(2B C'0, >+$.(,, 0, +(.@+,0J(<5 ->><0(/ %$+
-<< <(J(<, $% &'( &+((B E'() - )$/( ) $J(+%<$*,? - )(* )$/(
)
+
0, -<<$.-&(/ -& &'( ,-2( <(J(< -)/ &'( $1L(.&, -+(
/0,&+01@&(/ -2$)3 &'( )$/(,B E'() &'( +$$& )$/( ,><0&,? -
)(* +$$& 0, -<<$.-&(/ -)/ &'( &+(( 3+$*, $)( <(J(<B
C'( :<024&+(( >+$J0/(, &'+(( $>&0$), %$+ &'( 5/11-)=+0.%))
-<3$+0&'2O
! !"#$%&hT-)/$2<5 .'$$,( $)( $% &'( A@-<0%50)3
)$/(,B
! &'#$'()hQ'$$,( &'( )$/( &'-& '-, &'( 20)02@2
/0,&-).( %+$2 &'( )(* $1L(.& -)/ &'( +(>+(,()&-&0J(
6.()&(+; $% &'( )$/(B
! &'#%**+,hQ'$$,( &'( )$/( &'-& '-, &'( 20)02@2
$..@>-).5 -2$)3 &'( A@-<0%50)3 $)(,B C'0, 0, &'(
/(%-@<& 2(&'$/ /@( &$ 0&, 1(&&(+ >(+%$+2-).(? -, *(
,(( )(=&B
C'( )@21(+ $% ()&+0(, 0) (-.' .'0</ )$/( ]2O'.%,)-; 0,
2-0)&-0)(/ 0) 0&, ,'!)"'1!)4 C'( %0(</ 2O'.%,)- 0, 0)&()/(/ &$
1( @,(/ 15 &'( ;,'1>>+3 5/11-)=+0.%)) -<3$+0&'2B M<&'$@3'
0& @,(, ,$2( 2(2$+5 ,>-.( 0) (-.' ,'!)"'1!)? &'0, 0, @,@-<<5
- ,2-<< >+$>$+&0$) $% &'( &$&-< 2(2$+5 @,(/ %$+ (-.' ()&+5
6$)( 15&( 0, @,@-<<5 ()$@3';? 1@&? -, *( *0<< ,'$* <-&(+? &'0,
5/11-)=+0.%)) -<3$+0&'2 3()(+-&(/ &+((, *0&' '03'(+ )$/(
$..@>-&0$) +-&(,? <(-/0)3 &$ - ,2-<<(+ )@21(+ $% /0,D
-..(,,(,B W& -<,$ '(<>, 0) &'( :<024/$*) -<3$+0&'2? -,
>+(,()&(/ <-&(+B
C'( ,><0&&0)3 -<3$+0&'2, %$+ &'( :<024&+(( -+(O
! !"#$%&hC'( &*$ )(* .()&(+ $1L(.&, -+( +-)/$2<5
,(<(.&(/ -)/ &'( (=0,&0)3 $1L(.&, -+( /0,&+01@&(/
-2$)3 &'(2B #-.' $1L(.& 0, ,&$+(/ 0) &'( )(* )$/(
*'$,( .()&(+ 0, .<$,(,& &$ &'0, $1L(.&B C'0, 0, )$& -
*0,( ,&+-&(35? 1@& 0& 0, J(+5 %-,&B
! &'#-".hM<< >$,,01<( >-0+, $% $1L(.&, -+( .$),0/4
(+(/ -, >$&()&0-< +(>+(,()&-&0J(,B U$+ (-.' >-0+? -
<0)(-+ -<3$+0&'2 -,,03), &'( $1L(.&, &$ $)( $% &'(
+(>+(,()&-&0J(,B C'( >-0+ *'0.' 20)020N(, &'( .$J(+4
0)3 +-/0@, 0, .'$,()B C'( .$2><(=0&5 $% &'( -<3$+0&'2
0, )!+
*
"? @,0)3 )!+
"
" /0,&-).( .-<.@<-&0$),B C'0,
-<3$+0&'2 '-, -<+(-/5 1(() @,(/ %$+ &'( 94&+(( -)/ 0&
*-, %$ @) / & $ 1( &'( 2$,& > +$ 20 ,0 )3 ,><0 && 0)3
-<3$+0&'2 +(3-+/0)3 A@(+5 >(+%$+2-).( FGVIB
! -/0hC'( 20)02-< ,>-))0)3 &+(( FYVI $% &'( $1L(.&,
0, 3()(+-&(/ -)/ $)( $% &'( <$)3(,& -+., $% &'( &+(( 0,
/+$>>(/B C'0, -<3$+0&'2 0, $)( $% &'( .$)&+01@&0$),
$% &'0, >->(+B W& >+$/@.(, :<024&+((, -<2$,& -, 3$$/
-, &'( 20)9-= -<3$+0&'2 0) - %+-.&0$) $% &'( &02(B
C'( /(%-@<& -<3$+0&'2, &$ 1@0</ - :<024&+(( -+(O R;,'1>>+3S
%$+ &'( 5/11-) =+0.%)) -<3$+0&'2? &'( 9:C ,&+-&(35 %$+ ,><0&,?
&'( )$/( >&3&>,.$ 5 0, ]V %$+ J(.&$+ 6-
"
/0,&-).(; -)/ *$+/
/-&- ,(&, 6-
2!/0
/0,&-).(; -)/ YH %$+ U-.(W& /-&- ,(&B "(=&? &'(
)(* 9:C ,><0& -<3$+0&'2 0, /(,.+01(/B
J-<9 7A=(--(0K !=K21(-<> D!793 20 >(0(>!=
7A!00(0K -199
C'0, ,(.&0$) -//+(,,(, &'( %$<<$*0)3 >+$1<(2B d0J() - ,(& $%
5 $1L(.&, 0) - )$/( &$ 1( ,><0&? A@0.D<5 /0J0/( &'(2 0) &*$
3+$@>, ,$ &'-& &'( +(,@<&0)3 :<024&+(( <(-/, &$ <$* ,(-+.'
&02(,B E( >+$>$,( - ,><0& -<3$+0&'2 1-,(/ $) &'( 20)02-<
,>-))0)3 &+(( 69:C; *'0.' '-, 1(() ,@..(,,%@<<5 @,(/ 0)
.<@,&(+0)3 FYVIB E( .$),0/(+ &'( %@<< 3+->' .$),0,&0)3 $%
5 $1L(.&, -)/ +!+ , +" (/3(,? *'(+( &'( *(03'& $% &'( (/3(,
0, &'( /0,&-).( 1(&*(() &'( .$))(.&0)3 $1L(.&,B C'@,? *(
>+$.((/ *0&' &'( ,&(>, -, ,'$*) 0) U03B [B X)%$+&@)-&(<5?
&'0, -<3$+0&'2 /$(, )$& 3@-+-)&(( &'-& (-.' 3+$@> *0<<
+(.(0J( - 20)02@2 >(+.()&-3( $% $1L(.&,B C$ $1&-0) 2$+(
(J() /0,&+01@&0$)? *( .'$$,( &'( 2$,& ->>+$>+0-&( (/3(
%+$2 -2$)3 &'( <$)3(,& $)(,B W% )$)( (=0,&, 6-, 0) - ,&-+4
,'->(/ ,(&;? &'( @)(J() ,><0& 0, -..(>&(/ -)/ &'( <-+3(,&
(/3( 0, +(2$J(/B
U03B \ 0<<@,&+-&(, $@+ ->>+$-.' ->><0(/ &$ - J(.&$+ ,>-.(B
C'( )$/( &$ 1( ,><0& 0, >+(,()&(/ 0) U03B \-B M%&(+ 1@0</0)3
&'( 9:C 6U03B \1;? &'( (/3( 1(&*(() $1L(.&, 8 -)/ O *0<< 1(
/(<(&(/ -)/ $)( )$/( *0<< D((> &'( $1L(.&, 8# M# 5# -)/ :?
'-J0)3 M -, &'( +(>+(,()&-&0J(B C'( $&'(+ )$/( *0<< '-J( &'(
$1L(.&, O# T# X# -)/ W? '-J0)3 T -, &'( +(>+(,()&-&0J(B U03B \.
>+(,()&, &'( &*$ +(,@<&0)3 )$/(, -%&(+ &'( ,><0& 15 &'( 9:C
->>+$-.'B
#=>(+02()&, $) +(-< /-&- ,(&, ,'$* &'-& &'( )(* 9:C4
,><0&&0)3 2(&'$/ 0, .$),0/(+-1<5 %-,&(+ &'-) &'( 20)9-=4
,><0&&0)3 2(&'$/? *'0.' '-, 1(() .$),0/(+(/ &'( 1(,& %$+ &'(
94&+(( FGGI? F_I? *'0<( &'( A@(+5 >(+%$+2-).( 0, 20)02-<<5
-%%(.&(/B U$+ - )$/( *0&' .->-.0&5 5? &'( +@)&02( $% &'(
20)9-=4,><0&&0)3 2(&'$/ 0, 5!+
*
"? *'(+(-, &'( +@)&02( $%
&'( 9:C4,><0&&0)3 2(&'$/ 0, 5!+
"
14? +"B C'( >(+%$+2-).(
/0%%(+().( 0, +(%<(.&(/ 0) $@+ (=>(+02()&,? *'(+( &'( &02( &$
1@0</ - :<024&+(( @,0)3 &'( 9:C4,><0&&0)3 2(&'$/ 0, 2@.'
%-,&(+ &'-) @,0)3 &'( 20)9-=4,><0&&0)3 2(&'$/B 8$&'
,><0&&0)3 2(&'$/, +(,@<& 0) :<024&+((, *0&' -<2$,& &'( ,-2(
A@(+5 >(+%$+2-).(B E( $1,(+J(/ &'-& &'( A@(+5 >(+%$+4
2-).( ,@%%(+(/ - <0&&<( 15 @,0)3 &'( 9:C4,><0&&0)3 2(&'$/?
1@& $)<5 %$+ ,2-<< )$/( .->-.0&0(, 6<(,, &'-) YV;B
OQf %RRR :N?S6?;:%TS6 TS USTJVR5WR ?S5 5?:? RSW%SRRN%SW> ,TV= XQ> ST= O> 2?N;YZ?[N%V OPPO
!*+= \= ?.+03*$@8 103 #E.*$$*&+ " &0'( -#*&+ " 8*&*8". #E"&&*&+ $3((=
Authorized licensed use limited to: Carnegie Mellon Libraries. Downloaded on February 21,2010 at 14:58:48 EST from IEEE Xplore. Restrictions apply.

Citations
More filters
Journal ArticleDOI

Index-driven similarity search in metric spaces (Survey Article)

TL;DR: This article focuses on methods for similarity search that make the general assumption that similarity is represented with a distance metric d, and presents algorithms for common types of queries that operate on an arbitrary "search hierarchy."
Journal Article

Content-Based Image Retrieval: Theory and Applications.

TL;DR: The problems and challenges with the creation of CBIR systems are introduced, the existing solutions and appl ications are described, and the state of the art of the existing research in this area is presented.
Journal ArticleDOI

Contour salience descriptors for effective image retrieval and analysis

TL;DR: This work exploits the resemblance between content-based image retrieval and image analysis with respect to the design of image descriptors and their effectiveness by proposing contour saliences and segment saliences.
Journal ArticleDOI

DisC diversity: result diversification based on dissimilarity and coverage

TL;DR: In this article, the authors propose a new, intuitive definition of diversity called DisC diversity, which is defined as a subset of a query result such that each object in the result is represented by a similar object in a diverse subset and the objects in the diverse subset are dissimilar to each other.
References
More filters
Proceedings ArticleDOI

R-trees: a dynamic index structure for spatial searching

TL;DR: A dynamic index structure called an R-tree is described which meets this need, and algorithms for searching and updating it are given and it is concluded that it is useful for current database systems in spatial applications.
Journal ArticleDOI

On the shortest spanning subtree of a graph and the traveling salesman problem

TL;DR: Kurosh and Levitzki as discussed by the authors, on the radical of a general ring and three problems concerning nil rings, Bull Amer Math Soc vol 49 (1943) pp 913-919 10 -, On the structure of algebraic algebras and related rings.
Proceedings ArticleDOI

The R*-tree: an efficient and robust access method for points and rectangles

TL;DR: The R*-tree is designed which incorporates a combined optimization of area, margin and overlap of each enclosing rectangle in the directory which clearly outperforms the existing R-tree variants.
Proceedings Article

M-tree: An Efficient Access Method for Similarity Search in Metric Spaces

TL;DR: The results demonstrate that the Mtree indeed extends the domain of applicability beyond the traditional vector spaces, performs reasonably well in high-dimensional data spaces, and scales well in case of growing files.
Journal ArticleDOI

Multidimensional access methods

TL;DR: The class of point access methods, which are used to search sets of points in two or more dimensions, are presented and a discussion of theoretical and experimental results concerning the relative performance of various approaches are discussed.
Related Papers (5)
Frequently Asked Questions (1)
Q1. What are the contributions in this paper?

In this paper, the authors proposed a method to improve the quality of the data collected by the data collection system.