Jaeyoung Choi
Gerald Friedland
ABSTRACT
Today's geo-location estimation approaches are able to infer the location of a target image using its visual content alone. These approaches typically exploit visual matching techniques, applied to a large collection of background images with known geo-locations. Users who are unaware that visual analysis and retrieval approaches can compromise their geo-privacy, unwittingly open themselves to risks of crime or other unintended consequences. This paper lays the groundwork for a new approach to geo-privacy of social images: Instead of requiring a change of user behavior, we start by investigating users' existing photo-sharing practices. We carry out a series of experiments using a large collection of social images (8.5M) to systematically analyze how photo editing practices impact the performance of geo-location estimation. We find that standard image enhancements, including filters and cropping, already serve as natural geo-privacy protectors. In our experiments, up to 19% of images whose location would otherwise be automatically predictable were unlocalizeable after enhancement. We conclude that it would be wrong to assume that geo-visual privacy is a lost cause in today's world of rapidly maturing machine learning. Instead, protecting users against the unwanted effects of pixel-based inference is a viable research field. A starting point is understanding the geo-privacy bonus of already established user behavior.
- Shane Ahern, Dean Eckles, Nathaniel S Good, Simon King, Mor Naaman, and Rahul Nair. 2007 Over-exposed?: privacy patterns and considerations in online and mobile photo sharing. Inbibinfobooktitle Proceedings of the SIGCHI conference on Human factors in computing systems. ACM, pages357--366. Google Scholar
Digital Library
- Saeideh Bakhshi, David A Shamma, Lyndon Kennedy, and Eric Gilbert. 2015 Why We Filter Our Photos and How It Impacts Engagement.. In ICWSM. pages12--21.Google Scholar
- Julia Bernd, Blanca Gordo, Jaeyoung Choi, Bryan Morgan, Nicholas Henderson, Serge Egelman, Daniel D Garcia, and Gerald Friedland. 2015 Teaching Privacy: Multimedia Making a Difference. IEEE MultiMedia volume 22, number1 (2015), pages12--19.Google Scholar
- Andrew Besmer and Heather Richter Lipford. 2010 Moving beyond untagging: photo privacy in a tagged world. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. ACM, pages 1563--1572. Google ScholarDigital Library
- Finn Brunton and Helen Nissenbaum. Obfuscation: A user's guide for privacy and protest. MIT Press. Google ScholarDigital Library
- Serdar Çiftçi, Pavel Korshunov, Ahmet Akyüz Oğuz, and Touradj Ebrahimi. 2015. Using False Colors to Protect Visual Privacy of Sensitive Content. In SPIE Human Vision and Electronic Imaging XX. pages93941L-1-93941L-13.Google Scholar
- David M Chen, Georges Baatz, Kevin Köser, Sam S Tsai, Ramakrishna Vedantham, Timo Pylvänäinen, Kimmo Roimela, Xin Chen, Jeff Bach, Marc Pollefeys, and others. 2011. City-scale landmark identification on mobile devices. In proceedings of the 2011 IEEE Conference on Computer Vision and Pattern Recognition. IEEE, pages737--744. Google ScholarDigital Library
- Jaeyoung Choi, Martha Larson, Xinchao Li, Gerald Friedland, and Alan Hanjalic. 2016. Where to be wary: The impact of widespread photo-taking and image enhancement practices on users' geo-privacy. journal arXiv preprint arXiv:1603.01335 (2016).Google Scholar
- Andrea Costanzo, Irene Amerini, Roberto Caldelli, and Mauro Barni. 2014. Forensic Analysis of SIFT Keypoint Removal and Injection. journal IEEE Transactions on Information Forensics and Security volume9, number9 (July 2014), pages 1450--1464. Google ScholarDigital Library
- Thanh-Toan Do, Ewa Kijak, Teddy Furon, and Laurent Amsaleg. 2010. Deluding Image Recognition in Sift-based Cbir Systems. In Proceedings of the 2Nd ACM Workshop on Multimedia in Forensics, Security and Intelligence (MiFor '10). pages7--12. Google ScholarDigital Library
- Hazan E. Singer Y. Duchi, J. 2011. Adaptive Subgradient Methods for Online Learning and Stochastic Optimization. JMLR volume12 (2011), pages 2121--2159. Google ScholarDigital Library
- Quan Fang, Jitao Sang, and Changsheng Xu. 2013. GIANT: Geo-informative Attributes for Location Recognition and Exploration. In Proceedings of the ACM International Conference on Multimedia (MM '13). pages13--22. Google ScholarDigital Library
- Gerald Friedl and and Robin Sommer. 2010. Cybercasing the Joint: On the Privacy Implications of Geo-tagging. In Proceedings of the 5th USENIX Conference on Hot Topics in Security (HotSec'10). USENIX Association, Berkeley, CA, USA, pages1--8. Google ScholarDigital Library
- Petr Gronat, Guillaume Obozinski, Josef Sivic, and Tomas Pajdla. 2013. Learning and calibrating per-location classifiers for visual place recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition. pages907--914. Google ScholarDigital Library
- Claudia Hauff, Bart Thomee, and Michele Trevisiol. 2013. Working Notes for the Placing Task at MediaEval 2013. In Working Notes Proceedings of the MediaEval 2013 Workshop.Google Scholar
- James Hays and Alexei A Efros. 2008. IM2GPS: estimating geographic information from a single image. In Computer Vision and Pattern Recognition, 2008. CVPR 2008. IEEE Conference on. IEEE, pages1--8.Google ScholarCross Ref
- Chao-Yung Hsu Chun-Shien Lu, and Soo-Chang Pei. 2010. Secure and robust sift with resistance to chosen-plain attack. In Image Processing (ICIP), 2010 17th IEEE International Conference on. pages997--1000.Google Scholar
- Yannis Kalantidis, Giorgos Tolias, Yannis Avrithis, Marios Phinikettos, Evaggelos Spyrou, Phivos Mylonas, and Stefanos Kollias. 2011. VIRaL: Visual Image Retrieval and Localization. Multimedia Tools Appl. volume 51, number2 (Jan. 2011), pages555--592. Google ScholarDigital Library
- Erica Klarreich. 2016. Learning Securely. Commun. ACM 59, 11 (Oct. 2016), 12--14. Google ScholarDigital Library
- Bart P. Knijnenburg, Alfred Kobsa, and Hongxia Jin. 2013. Preference-based Location Sharing: Are More Privacy Options Really Better?. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (CHI '13). pages2667--2676. Google ScholarDigital Library
- Xinchao Li, Martha A Larson, and Alan Hanjalic. 2016. Geo-distinctive Visual Element Matching for Location Estimation of Images. arXiv preprint arXiv:1601.07884 (2016).Google Scholar
- matveyco. 2013. python-tilt-shift. https://github.com/matveyco/python-tilt-shift. (2013).Google Scholar
- Anh Nguyen, Jason Yosinski, and Jeff Clune. 2015. Deep Neural Networks Are Easily Fooled: High Confidence Predictions for Unrecognizable Images. In Conference on Computer Vision and Pattern Recognition CVPR'15.Google Scholar
- Andreas Poller, Martin Steinebach, and Huajian Liu 2012. Robust image obfuscation for privacy protection in web 2.0 applications. In IS&T/SPIE Electronic Imaging. International Society for Optics and Photonics, pages 830304--830304.Google Scholar
- Bradford W. Reyns, Billy Henson, and Bonnie S. Fisher. 2012. Stalking in the Twilight Zone: Extent of Cyberstalking Victimization and Offending Among College Students. Deviant Behaviror 31, 1 (2012), pages 1--25.Google Scholar
- Olga Russakovsky, Jia Deng, Hao Su, Jonathan Krause, Sanjeev Satheesh, Sean Ma, Zhiheng Huang, Andrej Karpathy, Aditya Khosla, Michael Bernstein, Alexander C. Berg, and Li Fei-Fei. 2015. ImageNet Large Scale Visual Recognition Challenge. International Journal of Computer Vision (IJCV) 115, 3 (2015), pages 211--252. Google ScholarDigital Library
- Eleftherios Spyromitros-Xioufis, Symeon Papadopoulos, Adrian Popescu, and Yiannis Kompatsiaris. 2016. Personalized privacy-aware image classification. In Proceedings of the 2016 ACM International Conference on Multimedia Retrieval. ACM, pages 71--78. Google ScholarDigital Library
- Christian Szegedy, Wei Liu, Yangqing Jia, Pierre Sermanet, Scott Reed, Dragomir Anguelov, Dumitru Erhan, Vincent Vanhoucke, and Andrew Rabinovich. 2015. Going Deeper with Convolutions. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. pages 1--9.Google ScholarCross Ref
- Yaniv Taigman, Ming Yang, Marc'Aurelio Ranzato, and Lars Wolf. 2014. Deepface: Closing the gap to human-level performance in face verification. In In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. IEEE, pages 1701--1708. Google ScholarDigital Library
- Michele Trevisiol, Hervé Jégou, Jonathan Delhumeau, and Guillaume Gravier. 2013. Retrieving geo-location of videos with a divide & conquer hierarchical multimodal approach. In Proceedings of the 2013 ACM conference on International conference on multimedia retrieval. ACM, pages 1--8. Google ScholarDigital Library
- Tobias Weyand, Ilya Kostrikov, and James Philbin. 2016. PlaNet - Photo Geolocation with Convolutional Neural Networks. In European Conference on Computer Vision (ECCV).Google Scholar
- Shicai Yang. 2016. Towards Good Practices for Recognition and Detection. (2016). http://image-net.org/challenges/talks/2016/Hikvision_at_ImageNet_2016.pdf.Google Scholar
- Sergej Zerr, Stefan Siersdorfer, and Jonathon Hare. 2012. Picalert!: A System for Privacy-aware Image Classification and Retrieval. In Proceedings of the 21st ACM International Conference on Information and knowledge management. ACM, pages 2710--2712. Google ScholarDigital Library
- Bolei Zhou, Aditya Khosla, Agata Lapedriza, Antonio Torralba, and Aude Oliva. 2016. Places: An Image Database for Deep Scene Understanding. arXiv preprint arXiv:1610.02055 (2016).Google Scholar
Index Terms
- The Geo-Privacy Bonus of Popular Photo Enhancements
Recommendations
Protecting patient geo-privacy via a triangular displacement geo-masking method
GeoPrivacy '14: Proceedings of the 1st ACM SIGSPATIAL International Workshop on Privacy in Geographic Information Collection and AnalysisProtecting patient geo-privacy while allowing for valid geographic analyses of the data is a major challenge [1]. As a consequence, a variety of methods have been introduced to mask patients' locational information, also called geo-masking methods [2]. ...
Preserving location and absence privacy in geo-social networks
CIKM '10: Proceedings of the 19th ACM international conference on Information and knowledge managementOnline social networks often involve very large numbers of users who share very large volumes of content. This content is increasingly being tagged with geo-spatial and temporal coordinates that may then be used in services. For example, a service may ...
Show me how you move and I will tell you who you are
SPRINGL '10: Proceedings of the 3rd ACM SIGSPATIAL International Workshop on Security and Privacy in GIS and LBSDue to the emergence of geolocated applications, more and more mobility traces are generated on a daily basis and collected in the form of geolocated datasets. If an unauthorized entity can access this data, it can used it to infer personal information ...
Comments