Figure 2 shows the receiver operating characteristics (ROC) curve of proposed speech and speaker recognition
References [12-14] have used it to enhance the performance of speaker recognition
In the linear combination, since both PSVM and GMM methods are inherently capable of accurate speaker verification, each can be individually trained and a linear combination of them be applied for speaker recognition
. In this case, better results shall be obtained.
Kishore, "Source and system features for speaker recognition
using AANN models", in Proc.
Previous work on security-based speaker recognition
systems largely falls within the domain of dealing with interspeaker variables, like channel and background noise.
Commonly, thirteen MFCC are used in most applications of speech recognition and speaker recognition
. In order to verify whether this number provides good classification accuracy and find a number of LSF coefficients, we conducted an experiment with a large variation in the number of extracted coefficients.
Waibel, "Applications of LDA to Speaker Recognition
," presented at the International Conference on Speech and Language Processing, Beijing, China, October 2000; retrieved February 10, 2002, from http://www.is.cs.cmu.edu/mie/.
The seven leading biometric technologies include facial recognition, fingerprint recognition, hand geometry, iris recognition, retina recognition, signature recognition, and speaker recognition
. To evaluate the effectiveness of biometrics in border control, it is important to recognize that the use of biometric technology would be but one component of the decision to support systems that determine who is allowed to enter the United States and who is not.
* Automated speaker recognition
systems that incorporate Pashtu, Urdu, Farsi, Arabic dialects, and other Middle Eastern and Central Asian languages.
Technologies include video enhancement, time line analysis software and speaker recognition
A possible performance breakthrough by a NIST scientist involving significant reduction in error rates was reported at the 2001 Speaker Recognition
Workshop, held in May 2001, in Linthicum, MD.