!"#$%&#'()%#*%+&#,(% -.-/0#12% 34+$5#%6".)0#% 37)"%8#)%9#1&#% :#;#5%<#&+4% =#*%>(%-1(4+?$%>"#$%>+00%@(?.&(% -.-/0#12% !"#$%&'($)*$+,-'./0 !"#$%#$&'()$*(+,-"#& !"#$%&'($)*$+,-'./0 .)%/0%&)%/%!-"#1& !"#$%&'($)*$+,-'./0 2'3%*-)(#4& !"#$%&'($)*$+,-'./0 !"#$%#$& '()$*(+,-"#& .)%/0%& )%/%!-"#1& 2'3%*-)(#4& 3('%"& ),552*(62-"#& 5"'(78(#4& 9"9,/2*($8& ,#'%*)$2#'(#4& ,)%*&+%:23("*& Related work • Social networks - Gladwell, 2002 - Rogers, 2003 - Romero et al, 2011 • Video popularity - Pinto et al, 2013 - Shamma et al, 2011 - Nwana et al, 2013 • High-­‐level aJributes - Dhar et al, 2011 - Isola et al, 2011 - PaJerson et al, 2012 - Khosla et al, 2013 • Text popularity - Petrovic et al, 2011 - Hong et al, 2011 Overview • What is image popularity? • Predict popularity using image content? • How does this compare to social factors? • What makes an image popular? • Conclusions and future work Overview • What is image popularity? • Predict popularity using image content? • How does this compare to social factors? • What makes an image popular? • Conclusions and future work !",*$)($)+,7'$8&892,.)*#/0 !",*$)($)+,7'$8&892,.)*#/0 !",*$)($)+,7'$8&892,.)*#/0 !",*$)($)+,7'$8&892,.)*#/0 What is image popularity? • Context dependent Facebook newsfeed Flickr search results Flickr user :&892,.)*#$%,*,('*0 K2$2)%$T&@UC&5(//("#&S/(!=*&(524%)& 3(%<)& :&892,.)*#$;&<*'=*(0 W#%I9%*I,)%*& %U4U&S/(!=*&)%2*!:&*%),/$)& V)%*I5(O& %U4U&S2!%+""=&#%<)7%%'& V)%*I;9%!(0!& %U4U&(#'(3(',2/&,)%*)& Popularity metric Popularity metric Popularity metric Task: predict log normalized views Measure: rank correla7on (ρ) Szabo and Huberman, 2010 Overview • What is image popularity? • Predict popularity using image content? • How does this compare to social factors? • What makes an image popular? • Conclusions and future work :.'%);><7$8&892,.)*#$9()<7$)+,7'$;&<*'<*0 $*2(#(#4&(524%)& ? 1.2 2.7 0.7 4.5 3.6 5.3 X& 7& @UC& :.'%);><7$8&892,.)*#$9()<7$)+,7'$;&<*'<*0 •! ;(59/%&(524%&7%2$,*%T&(524%&(#$%#)($8& normalized views 18 1.2 9 0 0 0.5 intensity mean 1 AUDC& :.'%);><7$8&892,.)*#$9()<7$)+,7'$;&<*'<*0 •! ;(59/%&(524%&7%2$,*%T&(524%&(#$%#)($8& normalized views 18 1.2 3.6 9 0 0 0.5 intensity mean 1 AUDC& AUYC& X& :.'%);><7$8&892,.)*#$9()<7$)+,7'$;&<*'<*0 •! ;(59/%&(524%&7%2$,*%T&(524%&(#$%#)($8& normalized views rank corr = -0.00 18 1.2 3.6 9 0 0 0.5 intensity mean 1 AUDC& AUYC& X& :.'%);><7$8&892,.)*#$9()<7$)+,7'$;&<*'<*0 •! ;(59/%&(524%&7%2$,*%T&(524%&(#$%#)($8& rank corr = 0.01 18 18 9 0 0 0.5 intensity mean 1 20 9 0 0 rank corr = -0.02 normalized views normalized views normalized views rank corr = -0.00 0.125 0.25 intensity variance 8 -5 -50 50 150 intensity skewness :.'%);><7$8&892,.)*#$9()<7$)+,7'$;&<*'<*0 •! ;(59/%&(524%&7%2$,*%)T&H;E&!"/"*&)92!%& Predic>ng popularity using image content • Simple image features: HSV color space rank corr = -0.05 normalized views 20 normalized views 10 0 rank corr = 0.01 20 normalized views 20 rank corr = 0.02 10 0 0.5 mean hue 1 0 10 0 0.5 mean saturation 1 0 0 0.5 mean value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redic>ng popularity using image content rank correla-on 0.4 0.3 0.2 0.1 0 User-­‐specific Predic>ng popularity using image content rank correla-on 0.4 0.3 0.2 0.1 One-­‐per-­‐user User-­‐mix User-­‐specific 0 Predic>ng popularity using image content Predic>ng popularity using image content Overview • What is image popularity? • Predict popularity using image content? • How does this compare to social factors? • What makes an image popular? • Conclusions and future work Predic>ng popularity using social cues • User informa-on • Mean views • Photo count • Contacts • Groups • Group members • Member dura-on • Is pro • Image informa-on • Tags • Title length • Descrip-on length Predic>ng popularity using social cues 0.8 0.7 rank correla-on 0.6 0.5 0.4 One-­‐per-­‐user 0.3 User-­‐mix 0.2 User-­‐specific 0.1 0 -­‐0.1 Predic>ng popularity using image + social cues rank correla-on 0.6 0.4 One-­‐per-­‐user User-­‐mix User-­‐specific 0.2 0 Content only Social only Combined Predic>ng popularity using image + social cues 1 rank correla-on 0.8 0.6 One-­‐per-­‐user User-­‐mix 0.4 User-­‐specific 0.2 0 Content only Social only Combined Predic>ng popularity using image + social cues 1 rank correla-on 0.8 0.6 One-­‐per-­‐user User-­‐mix 0.4 User-­‐specific 0.2 0 Content only Social only Combined Overview • What is image popularity? • Predict popularity using image content? • How does this compare to social factors? • What makes an image popular? • Conclusions and future work R"/"*&9%*!%#$24%& !",*$+,4'($,<$)+,7'$8&892,./0 32#&'%&P%(_%*&%$&2/?&REFB&@AAc& !",*$+,4'($,<$)+,7'$8&892,./0 I*-/$%% I&#,(% =.0.1% "+)$.,1#&% 9/--.1$% N(?$.1% :(,1())+.*% DUZ& I&#,(%-.-/0#1+$5% color importance What makes an image popular? colors !",*$+,4'($,<$)+,7'$8&892,./0 •! F*%'(!-#4&"+_%!$&!2$%4"*(%)& L524%& W+_%!$&& !2$%4"*8& M*(6:%3)=8&%$&2/?&GLF;&@AD@& What makes an image popular? • tench, Tinca -nca • jay • goldfish, Carassius auratus • magpie • great white shark • chickadee • -ger shark, Galeocerdo cuvieri • water ouzel, dipper • hammerhead, hammerhead shark • kite • electric ray, crampfish, numbfish, torpedo • bald eagle, American eagle, Haliaeetus leucocephalus • s-ngray • vulture • cock • great grey owl, great gray owl, Strix nebulosa • hen • European fire salamander, Salamandra salamandra • ostrich, Struthio camelus • common newt, Triturus vulgaris • brambling, Fringilla mon-fringilla • eg • goldfinch, Carduelis carduelis • spoJed salamander, Ambystoma maculatum • house finch, linnet, Carpodacus mexicanus • axolotl, mud puppy, Ambystoma mexicanum • junco, snowbird • bullfrog, Rana catesbeiana • indigo bun-ng • tree frog, tree-­‐frog • robin, American robin • tailed frog, bell toad, ribbed toad, tailed toad, Ascaphus trui • bulbul • loggerhead, loggerhead turtle, CareJa careJa !",*$+,4'($,<$)+,7'$8&892,./0 I*-/$%% I&#,(% J@O(?$%% 0+'(0+"..4% 9/--.1$% N(?$.1% :(,1())+.*% DUZ& I&#,(%-.-/0#1+$5% !",*$+,4'($,<$)+,7'$8&892,./0 a%'(,5&9")(-3%&(592!$& & ,+#*$%-#*4#% 0#45@/,% @#)'($@#00% -0.>% ?"(($#"% 00#&#% !",*$+,4'($,<$)+,7'$8&892,./0 ;$*"#4&9")(-3%&(592!$& & & @1#))+(1(% &#+00.$% 1(D.0D(1% @+'+*+% &+*+)'+1$% ?/-% !",*$+,4'($,<$)+,7'$8&892,./0 G%42-3%&(592!$& & & )-#$/0#% -0/*,(1% 0#-$.-% !",*$+,4'($,<$)+,7'$8&892,./0 ;$*"#4&9")(-3%& & •! •! •! •! •! •! 5(#()=(*$& 52(//"$& +(=(#(& !,9& +*2))(%*%& *%3"/3%*& a%'(,5&9")(-3%& & •! •! •! •! •! •! !:%%$2:& 4(2#$&92#'2& +2)=%$+2//& //252& 9/"<& /2'8+,4& •! •! •! •! •! G%42-3%& & )92$,/2& 9/,#4%*& /29$"9& 4"/7!2*$& )92!%&:%2$%*& Overview • What is image popularity? • Predict popularity using image content? • How does this compare to social factors? • What makes an image popular? • Conclusions and future work Conclusion • Image popularity can be predicted to a reasonable extent using content alone! • Both image content and social cues are important for popularity predic-on • Content/social cues help to a different degree in different contexts • We iden-fy some factors that make an image popular Future work • Predic-ng other no-ons of popularity e.g. shareability, likeability, comments • Temporal evolu-on of popularity • Effect of high-­‐level aJributes such as emo-ons • Automa-cally modifying image popularity hPp://popularity.csail.mit.edu Thank you!