Multi-speaker voice cryptographic key generation

Summary form only given. In this work we present a procedure for generating binary vectors, which can be used as keys for cryptographic purposes. This research is based on automatic speech recognition technology and support vector machines. Keys bits are produced by making a distinction among the phoneme features of the users employing hyperplanes. The implementation of the method is described and the statistics are computed for several numbers of users. The results obtained show that the proposed method is sufficiently robust to reliably regenerate the key.