Despite its noninvasive nature, subject identification by voice is not as popular as other biometric procedures (i.e. fingerprinting). In part, this is due to the difficulty of establishing how close is close enough when comparing spectral features. In this work, we address this issue by showing how to characterize spectra by means of sets of integers, borrowing topological tools used in the theory of dynamical systems. On the other hand, we report an empirical result: within a relatively small bank of speakers, there are subsets of integers that seem to strenghten the speakers' identity information. These results suggest a new direction in the identification of subjects by voice: one in which arrangements of integers define voiceprints that stand on their own, despite any acceptance/rejection thresholds. © 2004 Elsevier B.V. All rights reserved.
Documento: | Artículo |
Título: | Topological voiceprints for speaker identification |
Autor: | Trevisan, M.A.; Eguia, M.C.; Mindlin, G.B. |
Filiación: | Departamento de Física, FCEyN, Univ. de Buenos Aires Cd. Univ., Pab. 1 CI428EGA Buenos Aires, Argentina Centro de Estudios e Imestigaciones, Universidad Nacional de Quilmes, Roque Sáenz Peña 180, Bernai, B1876BXD Buenos Aires, Argentina |
Palabras clave: | Biometrics; Speaker recognition; Topological indexes; Deformation; Ergonomics; Integer programming; Modulation; Oscillations; Pressure effects; Cross-counting algorithms; Power spectrum; Speech segments; Voiceprints; Spectrum analysis |
Año: | 2005 |
Volumen: | 200 |
Número: | 1-2 |
Página de inicio: | 75 |
Página de fin: | 80 |
DOI: | http://dx.doi.org/10.1016/j.physd.2004.09.008 |
Título revista: | Physica D: Nonlinear Phenomena |
Título revista abreviado: | Phys D Nonlinear Phenom |
ISSN: | 01672789 |
CODEN: | PDNPD |
Registro: | https://bibliotecadigital.exactas.uba.ar/collection/paper/document/paper_01672789_v200_n1-2_p75_Trevisan |