Share Email Print
cover

Proceedings Paper

Significance test with data dependency in speaker recognition evaluation
Author(s): Jin Chu Wu; Alvin F. Martin; Craig S. Greenberg; Raghu N. Kacker; Vincent M. Stanford
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

To evaluate the performance of speaker recognition systems, a detection cost function defined as a weighted sum of the probabilities of type I and type II errors is employed. The speaker datasets may have data dependency due to multiple uses of the same subjects. Using the standard errors of the detection cost function computed by means of the two-layer nonparametric two-sample bootstrap method, a significance test is performed to determine whether the difference between the measured performance levels of two speaker recognition algorithms is statistically significant. While conducting the significance test, the correlation coefficient between two systems’ detection cost functions is taken into account. Examples are provided.

Paper Details

Date Published: 23 May 2013
PDF: 13 pages
Proc. SPIE 8734, Active and Passive Signatures IV, 87340I (23 May 2013); doi: 10.1117/12.2014536
Show Author Affiliations
Jin Chu Wu, National Institute of Standards and Technology (United States)
Alvin F. Martin, National Institute of Standards and Technology (United States)
Craig S. Greenberg, National Institute of Standards and Technology (United States)
Raghu N. Kacker, National Institute of Standards and Technology (United States)
Vincent M. Stanford, National Institute of Standards and Technology (United States)


Published in SPIE Proceedings Vol. 8734:
Active and Passive Signatures IV
G. Charmaine Gilbreath; Chadwick Todd Hawley, Editor(s)

© SPIE. Terms of Use
Back to Top