De-reverberation using DNN for Non-Reference Reverberant Speech Intelligibility Estimation.
Kazushi NakazawaKazuhiro KondoPublished in: GCCE (2018)
Keyphrases
- speech signal
- speech recognition
- automatic speech recognition
- sound source
- vocal tract
- pattern recognition
- speech synthesis
- noisy environments
- estimation accuracy
- estimation algorithm
- automatic speech recognition systems
- spoken language
- linear prediction
- accurate estimation
- maximum likelihood estimation
- audio visual
- computer vision
- training process
- signal to noise ratio
- non stationary
- parameter estimation
- human computer interaction
- denoising
- facial expressions
- probabilistic model
- hidden markov models
- multiscale
- multimedia