Knowledge Distillation-Based Training of Speech Enhancement for Noise-Robust Automatic Speech Recognition.
Geon Woo LeeHong Kook KimDuk-Jo KongPublished in: IEEE Access (2024)
Keyphrases
- noisy environments
- speech enhancement
- automatic speech recognition
- speech signal
- speech recognition
- noise reduction
- spectral subtraction
- background noise
- hidden markov models
- speaker identification
- prior knowledge
- signal to noise ratio
- additive noise
- broadcast news
- information retrieval
- multi modal
- edge detection
- acoustic features
- vocal tract
- speech retrieval
- neural network