Speech enhancement approach for body-conducted unvoiced speech based on Taylor-Boltzmann machines trained DNN.
C. KarthikeyanT. Rajesh KumarD. Vijendra BabuM. BaskarRamkumar JayaramanMohammad ShahidPublished in: Comput. Speech Lang. (2024)
Keyphrases
- speech enhancement
- speech signal
- boltzmann machines
- speech recognition
- noisy environments
- vocal tract
- noise reduction
- linear prediction
- noisy speech
- single channel
- signal to noise ratio
- automatic speech recognition
- background noise
- learning tasks
- non stationary
- boltzmann machine
- spectral analysis
- hidden markov models
- multi channel
- sound source
- wiener filter
- information retrieval
- language model
- neural network