Augmented Cepstral Normalization for Robust Speech Recognition

@inproceedings{Acero2000AugmentedCN,
  title={Augmented Cepstral Normalization for Robust Speech Recognition},
  author={Alex Acero},
  year={2000}
}
We proposed an augmented cepstral mean normalization algorithm that differentiates noise and speech during normalization, and computes a different mean for each. The new procedure reduced the error rate slightly for the case of sameenvironment testing, and significantly reduced the error rate by 25% when an environmental mismatch exists over the case of standard cepstral mean normalization.