Biometric based unique key generation for authentic audio watermarking

Abstract

This paper proposes a method of generating pseudorandom number sequences based on biometric templates of iris image. These sequences are found to be unique in nature. Such sequences can be stored in a database for distinct identification of the extracted keys and they can act as secret keys for audio watermarking with a stamp of ownership unlike arbitrary pseudorandom number sequences and chaotic sequences. Correlation scores achieved under signal processing attacks is more than 0.9 that is significant for identification. © 2009 Springer-Verlag Berlin Heidelberg.