Robust speaker recognition: a feature-based approach
Citations
1,686 citations
Cites background from "Robust speaker recognition: a featu..."
...Speaker-recognition systems can be made somewhat robust against noise and channel variations [33], [49], ordinary human changes (e.g., time-of-day voice changes and minor head colds), and mimicry by humans and tape recorders [22]....
[...]
1,433 citations
Cites background or methods from "Robust speaker recognition: a featu..."
...The graph structure was motivated by invariance against the affine feature distortion model for cepstral features (e.g. Mak and Tsang, 2004; Mammone et al., 1996)....
[...]
...It is more desirable to design a feature extractor which is itself robust (Mammone et al., 1996), or to normalize the features before feeding them onto the modeling or matching algorithms....
[...]
...Linear prediction (LP) (Makhoul, 1975; Mammone et al., 1996) is an alternative spectrum estimation method to DFT that has good intuitive interpretation both in time domain (adjacent samples are correlated) and frequency domain (all-pole spectrum corresponding to the resonance structure)....
[...]
846 citations
554 citations
References
8,442 citations
"Robust speaker recognition: a featu..." refers background or methods in this paper
...The first derivative of the cepstrum (also known as the delta cepstrum) is defined as [ 17 ]...
[...]
...n n=1,2, ..., L 0 otherwise ' w(n) = and bandpass lifering (BPL) [ 17 ,26] where...
[...]
...The autocorrelation method and the covariance method are two standard methods of solving for the predictor coefficients [12, 17 ]....
[...]
...the unit circle and the gain is 1, the causal LP cepstrum cg(n) of H(z) is given by [ 17 ,23,24]....
[...]
4,862 citations
3,134 citations
"Robust speaker recognition: a featu..." refers methods in this paper
...Methods such as probabilistic optimum filtering [6], Gaussian mixture model (GMM) [ 7 ], and hidden Markov model (HMM) adaptation [SI all fall into this category....
[...]
3,103 citations