EUVIMEDEUVIMED entdecken
Crossref

Limited Data Speaker Verification using Different Modelling Techniques

Jayanthi Kumari T R, Anita R, Rajesh L, Navya V

International Journal of Emerging Technologies and Innovative Research · 10.04.2026 · 44

DOI: 10.48175/ijetir-9210 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Emerging Technologies and Innovative Research, online erschienen am 10.04.2026. Verfasst von Jayanthi Kumari T R und 3 weiteren Autor:innen. Die Arbeit nennt 12 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Most modeling techniques in speaker verification (SV) systems rely on sufficient data to effectively model the speaker, resulting in good performance. However, studies have shown that performance significantly degrades when the available training data is limited, due to inadequate speaker modeling. Additionally, insufficient speech data during testing can lead to unreliable verification decisions. Therefore, the objective of this work is to achieve reliable and accurate performance in SV systems even under data-limited conditions. Several modeling techniques have been employed in SV systems, including CVQ, FVQ, SOM, LVQ, GMM, and GMM-UBM

Zitieren

APA

R, J. K. T., R, A., L, R., & V, N. (2026). Limited Data Speaker Verification using Different Modelling Techniques. International Journal of Emerging Technologies and Innovative Research, 44. https://doi.org/10.48175/ijetir-9210

Vancouver

R JKT, R A, L R, V N. Limited Data Speaker Verification using Different Modelling Techniques. International Journal of Emerging Technologies and Innovative Research. 2026:44. doi:10.48175/ijetir-9210

BibTeX
@article{R2026,
  author = {R, Jayanthi Kumari T and R, Anita and L, Rajesh and V, Navya},
  title = {Limited Data Speaker Verification using Different Modelling Techniques},
  journal = {International Journal of Emerging Technologies and Innovative Research},
  year = {2026},
  pages = {44},
  doi = {10.48175/ijetir-9210}
}
Kostenlos starten →