Sciweavers

Rapid Speaker Adaptation Using Clustered Maximum-Likelihood Linear Basis With Sparse Training Data