We propose a method of voice passphrase variability calculation. To calculate the variability, we transform an acoustic passphrase into a sequence of formants, then into a multi-dimensional histogram. We then compute a value which characterizes the entropy of the sequence. We provide a computer simulation and conclude that using this variability for passphrase creation (e.g. during the enrollment process) helps to significantly increase the performance of speaker verification and speaker identification systems.
CITATION STYLE
Sukhmel, V., Aleinik, S., & Shchemelinin, V. (2015). Voice passphrase variability evaluation for speaker recognition. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8915, pp. 3–9). Springer Verlag. https://doi.org/10.1007/978-3-319-20125-2_1
Mendeley helps you to discover research relevant for your work.