PINtext: A framework for secure communication based on context

Abstract

We propose a framework for secure device-pairing based on contextual information. In contrast to related work, we utilise fuzzy cryptography schemes to establish a common secret among devices and to account for noise in the contextual features. The framework is implemented for ambient audio and its feasibility is demonstrated in a case study. Additionally, we discuss properties of fingerprints from ambient audio and estimate the entropy of fingerprints in statistical tests. © 2012 ICST Institute for Computer Science, Social Informatics and Telecommunications Engineering.