Most previous private matching work are based on the secure multi-party computation (SMC). There are two mainstreams of approaches to solve the private profile-based friending problem. The first category is based on private set intersection (PSI) and private cardinality of set intersection (PCSI). Early work in this category mainly address the private set operation problem in database research, studied the problem of private k nearest neighbor search. Later on some work treat a user’s profile as multiple attributes chosen from a public set of attributes and provide well-designed protocols to privately match users’ profiles based on PSI and PCSI. The second category is based on private vector dot product. Considers a user’s profile as vector which represents his/her social coordinate, and the social proximity between two uses as the matching metric. It calculates the metric by private dot product. A trusted central server is requited to precompute users social coordinates and generate certifications and keys. Improves these work with a fine-grained private matching by associating a user-specific numerical value with every attribute to indicate the level of interest. Most these approaches lacks a specific definition of matching user. symmetric-encryption based privacy-preserving profile matching and secure communication channel establishment mechanism in decentralized social networks without any presetting or trusted third party. We take advantage of the common attributes between matching users to encrypt a secret message with a channel key in it. Several protocols were proposed for achieving different levels of privacy. Our lightweight privacy preserving flexible profile matching in decentralized mobile social networks without any presetting or trust third party. A secure communication channel is constructed between matching users.
You are here: Home / ieee projects 2013-2014 / Privacy preserving profile matching and Secure communication