Privacy preserving multimedia content identification for cloud based bag-of-feature architectures

In this paper we consider privacy-preserving multimedia content identification for a cloud based Bag-of-Feature (BoF) framework. We analytically model how geometric information can be used as a shared secret and derive the tradeoff between identification capability, privacy and computational load. In addition we suggest a descriptor ambiguization method that introduces uncertainty to the server with respect to the true interest of the data users.