• 1. A Semantic Context-AwarePrivacy Model forFaceBlockPrimal Pappachan, Roberto Yus, Prajit KumarDas, Tim Finin, Eduardo Mena, and AnupamJoshihttp://face-block.me

2. Cameras are Ubiquitous 3. Social Networks 4. Invisible Cameras 5. Technology v/s PrivacyNews Paper articleshttp://www.youtube.com/watch?v=ClvI9fZaz6M 6. Solutions Really? 7. Introducing FaceBlockhttp://www.youtube.com/watch?v=IseoIWNWiR8 8. Privacy-Aware Pictures 9. How it works? 10. All in or nothing?A persons preferences would depend on hercontext (e.g., time, place, or activity)ExamplesI am okay being photographed by peopleI know at a private eventI do not like to be photographed when Iam at public places. 11. Semantic WebTechnologiesUnderstand the semantics of conceptssuch as public place, people I knowor private eventSemantically represent privacy policiesbased on conceptsDynamically infer user preferencesabout pictures based on context 12. Context-Aware[...] any information that can be used tocharacterize the situation of an entity Deyand Abowd 13. Privacy PoliciesFor expressing userpreference on picturesConstraints based on usercontext modelSemantic Web RuleLanguage (SWRL) 14. Context PiecesLocation basedActivity basedUnique IDTime 15. Example Policydo not allow my social network colleaguesgroup (identity context) to take pictures ofme (identity context) at parties (activitycontext) held on weekends (time context) at the beachhouse (location context) 16. Glass UserWishes to take pictures at the partyRuns FaceBlock in the backgroundReceives face identifiers and policiesDetects, recognizes and obscures thefaces as necessary 17. OthersWishes to protect his privacy at the partyGenerates face identifierSpecifies context constraints using rulesRuns FaceBlock in the background 18. ProtocolExchange IdentityShare Face IdentifierI:L: At T:BeachColleagueHouseContext RecognitionA: PartyWeekendPolicy TriggeredPrimalID, FaceBlock: True 19. Other scenarios 20. ChallengesImageFace Recognition / Detection /Identifier GenerationCommunicationMalicious Policies 21. ChallengesContext and PolicyImprecise contextPolicies - Generation, Conflict Resolution, ValidityGeneralPrivacy LossEnforcement or IncentivizingEnergy Cost 22. Take awaysUsers are defenseless against loss of privacy inpicturesNovel approach for taking privacy-awarepicturesSemantic Web technologies makes FaceBlocksmarterProof-of-concept implementationhttp://face-block.me Thank you NSF and SWSA