Causal speech enhancement using dynamical-weighted loss and attention encoder-decoder recurrent neural network

Speech enhancement (SE) reduces background noise signals in target speech and is applied at the front end in various real-world applications, including robust ASRs and real-time processing in mobile phone communications. SE systems are commonly integrated into mobile phones to increase quality and i...

Descripción completa

Detalles Bibliográficos
Autores principales: Peracha, Fahad Khalil, Khattak, Muhammad Irfan, Salem, Nema, Saleem, Nasir
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Public Library of Science 2023
Materias:
Research Article
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10174555/
https://www.ncbi.nlm.nih.gov/pubmed/37167227
http://dx.doi.org/10.1371/journal.pone.0285629

Internet

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10174555/
https://www.ncbi.nlm.nih.gov/pubmed/37167227
http://dx.doi.org/10.1371/journal.pone.0285629

Ejemplares similares