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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Text-dependent speaker verification is becoming popular in the speakerrecognition society. However, the conventional i-vector framework which hasbeen successful for speaker identification and other similar tasks worksrelatively poorly in this task. Researchers have proposed several new methodsto improve performance, but it is still unclear that which model is the bestchoice, especially when the pass-phrases are prompted during enrollment andtest. In this paper, we introduce four modeling methods and compare theirperformance on the newly published RedDots dataset. To further explore theinfluence of different frame alignments, Viterbi and forward-backwardalgorithms are both used in the HMM-based models. Several bottleneck featuresare also investigated. Our experiments show that, by explicitly modeling thelexical content, the HMM-based modeling achieves good results in thefixed-phrase condition. In the prompted-phrase condition, GMM-HMM andi-vector/HMM are not as successful. In both conditions, the forward-backwardalgorithm brings more benefits to the i-vector/HMM system. Additionally, wealso find that even though bottleneck features perform well fortext-independent speaker verification, they do not outperform MFCCs on the mostchallenging Imposter-Correct trials on RedDots.

Comparison of Multiple Features and Modeling Methods for Text-dependent Speaker Verification