As an effective and low-dimension rep-resentation for speech utterances with different lengths, i-vector method has drawn considerable attentions in speaker verification. Training a Total variability space (TVS) is one of the key parts in the i-vector method. How-ever, the traditional training method only explores the re-lationship between different mean supervectors, ignoring priori category information of speakers, which results in a lack of discrimination. In the proposed method, a dis-criminative TVS based on Partial least squares (PLS) is estimated, in which both the correlation of intra-class and the distinction of inter-class are fully utilized due to us-ing speaker labels, and the proposed method can achieve a better performance.
展开▼
