English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

In this paper, we propose a speaker-verification system based on maximumlikelihood linear regression (MLLR) super-vectors, for which speakers arecharacterized by m-vectors. These vectors are obtained by a uniformsegmentation of the speaker MLLR super-vector using an overlapped slidingwindow. We consider three approaches for MLLR transformation, based on theconventional $1$-best automatic transcription, on the lattice wordtranscription, or on a simple global universal background model (UBM). Sessionvariability compensation is performed in a post-processing module withprobabilistic linear discriminant analysis (PLDA) or the eigen factor radial(EFR). Alternatively, we propose a cascade post-processing for the MLLRsuper-vector based speaker-verification system.  In this case, the m-vectors or MLLR super-vectors are first projected onto alower-dimensional vector space generated by linear discriminant analysis (LDA).Next, PLDA session variability compensation and scoring is applied to thereduced-dimensional vectors. This approach combines the advantages of bothtechniques and makes the estimation of PLDA parameters easier. Experimentalresults on telephone conversations of the NIST 2008 and 2010 speakerrecognition evaluation (SRE) indicate that the proposed m-vector systemperforms significantly better than the conventional system based on the fullMLLR super-vectors. Cascade post-processing further reduces the error rate inall cases. Finally, we present the results of fusion with a standard i-vectorsystem in the feature, as well as in the score domain, demonstrating that them-vector system is both competitive and complementary with it.

Sub-vector Extraction and Cascade Post-Processing for Speaker Verification Using MLLR Super-vectors