Enhancing Far-Field Speech Recognition with Mixer: A Novel Data Augmentation Approach
Recent advancements in end-to-end (E2E) modeling have notably improved automatic speech recognition (ASR) systems; however, far-field speech recognition (FSR) remains challenging due to signal degradation from factors such as low signal-to-noise ratio, reverberation, and interfering sounds. This req...
Enregistré dans:
| Auteurs principaux: | , , , |
|---|---|
| Format: | Artigo |
| Langue: | Inglês |
| Publié: |
MDPI AG
2025-04-01
|
| Collection: | Applied Sciences |
| Sujets: | |
| Accès en ligne: | https://www.mdpi.com/2076-3417/15/7/4073 |
| Tags: |
Pas de tags, Soyez le premier à ajouter un tag!
|