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

Speaker verification aims to verify whether an input speech corresponds tothe claimed speaker, and conventionally, this kind of system is deployed basedon single-stream scenario, wherein the feature extractor operates in fullfrequency range. In this paper, we hypothesize that machine can learn enoughknowledge to do classification task when listening to partial frequency rangeinstead of full frequency range, which is so called frequency selectiontechnique, and further propose a novel framework of multi-stream ConvolutionalNeural Network (CNN) with this technique for speaker verification tasks. Theproposed framework accommodates diverse temporal embeddings generated frommultiple streams to enhance the robustness of acoustic modeling. For thediversity of temporal embeddings, we consider feature augmentation withfrequency selection, which is to manually segment the full-band of frequencyinto several sub-bands, and the feature extractor of each stream can selectwhich sub-bands to use as target frequency domain. Different from conventionalsingle-stream solution wherein each utterance would only be processed for onetime, in this framework, there are multiple streams processing it in parallel.The input utterance for each stream is pre-processed by a frequency selectorwithin specified frequency range, and post-processed by mean normalization. Thenormalized temporal embeddings of each stream will flow into a pooling layer togenerate fused embeddings. We conduct extensive experiments on VoxCelebdataset, and the experimental results demonstrate that multi-stream CNNsignificantly outperforms single-stream baseline with 20.53 % of relativeimprovement in minimum Decision Cost Function (minDCF).

Multi-stream Convolutional Neural Network with Frequency Selection for Robust Speaker Verification