Attention-based latent features for jointly trained end-to-end automatic speech recognition with modified speech enhancement.
Da-Hee YangJoon-Hyuk ChangPublished in: J. King Saud Univ. Comput. Inf. Sci. (2023)
Keyphrases
- end to end
- automatic speech recognition
- speech signal
- noisy environments
- speech enhancement
- speech recognition
- conversational speech
- feature set
- feature vectors
- feature extraction
- congestion control
- machine learning
- language model
- signal to noise ratio
- noise reduction
- feature space
- bayesian networks
- vocal tract
- computer vision