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

Recently, more and more personalized speech enhancement systems (PSE) withexcellent performance have been proposed. However, two critical issues stilllimit the performance and generalization ability of the model: 1) Acousticenvironment mismatch between the test noisy speech and target speakerenrollment speech; 2) Hard sample mining and learning. In this paper, dynamicacoustic compensation (DAC) is proposed to alleviate the environment mismatch,by intercepting the noise or environmental acoustic segments from noisy speechand mixing it with the clean enrollment speech. To well exploit the hardsamples in training data, we propose an adaptive focal training (AFT) strategyby assigning adaptive loss weights to hard and non-hard samples duringtraining. A time-frequency multi-loss training is further introduced to improveand generalize our previous work sDPCCN for PSE. The effectiveness of proposedmethods are examined on the DNS4 Challenge dataset. Results show that, the DACbrings large improvements in terms of multiple evaluation metrics, and AFTreduces the hard sample rate significantly and produces obvious MOS scoreimprovement.

Dynamic Acoustic Compensation and Adaptive Focal Training for Personalized Speech Enhancement