Cargando…
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...
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: | |
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 |
Ejemplares similares
-
Structured neuronal encoding and decoding of human speech features
por: Tankus, Ariel, et al.
Publicado: (2012) -
Knowledge Tracing via Attention Enhanced Encoder-Decoder
por: Zhang, Kai, et al.
Publicado: (2022) -
Speaker-independent auditory attention decoding without access to clean speech sources
por: Han, Cong, et al.
Publicado: (2019) -
GPS Trajectory Completion Using End-to-End Bidirectional Convolutional Recurrent Encoder-Decoder Architecture with Attention Mechanism
por: Nawaz, Asif, et al.
Publicado: (2020) -
Attention-Based Personalized Encoder-Decoder Model for Local Citation Recommendation
por: Yang, Libin, et al.
Publicado: (2019)