We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Sound

Title: A text-independent speaker verification model: A comparative analysis

Abstract: The most pressing challenge in the field of voice biometrics is selecting the most efficient technique of speaker recognition. Every individual's voice is peculiar, factors like physical differences in vocal organs, accent and pronunciation contributes to the problem's complexity. In this paper, we explore the various methods available in each block in the process of speaker recognition with the objective to identify best of techniques that could be used to get precise results. We study the results on text independent corpora. We use MFCC (Melfrequency cepstral coefficient), LPCC (linear predictive cepstral coefficient) and PLP (perceptual linear prediction) algorithms for feature extraction, PCA (Principal Component Analysis) and tSNE for dimensionality reduction and SVM (Support Vector Machine), feed forward, nearest neighbor and decision tree algorithms for classification block in speaker recognition system and comparatively analyze each block to determine the best technique
Comments: presented and accepted by 2017 International Conference on Intelligent Computing and Control (I2C2)
Subjects: Sound (cs.SD); Audio and Speech Processing (eess.AS)
Cite as: arXiv:1712.00917 [cs.SD]
  (or arXiv:1712.00917v1 [cs.SD] for this version)

Submission history

From: Karthik R [view email]
[v1] Mon, 4 Dec 2017 06:09:02 GMT (628kb)

Link back to: arXiv, form interface, contact.