期刊名称：International Journal of Computer Science and Information Technologies
电子版ISSN：0975-9646
出版年度：2016
卷号：7
期号：2
页码：780-782
出版社：TechScience Publications
摘要：Photo sharing is an attractive feature whichpopularizes Online Social Networks (OSNs). Unfortunately, itmay leak users’ privacy if they are allowed to post, comment,and tag a photo freely. In this paper, we attempt to addressthis issue and study the scenario when a user shares a photocontaining individuals other than himself/herself (termed cophotofor short). To prevent possible privacy leakage of aphoto, we design a mechanism to enable each individual in aphoto be aware of the posting activity and participate in thedecision making on the photo posting. For this purpose, weneed an efficient facial recognition (FR) system that canrecognize everyone in the photo. However, more demandingprivacy setting may limit the number of the photos publiclyavailable to train the FR system[1]. To deal with this dilemma,our mechanism attempts to utilize users’ private photos todesign a personalized FR system specifically trained todifferentiate possible photo co-owners without leaking theirprivacy. We also develop a distributed consensus basedmethod to reduce the computational complexity and protectthe private training set. We show that our system is superiorto other possible approaches in terms of privacy usingencryption algorithm and opensource. Our mechanism isimplemented as a proof of concept Android application onFace book’s platform.
关键词：online social networks; FR system; open social;privacy; homomorphic encryption
