SE-MelGAN - Speaker Agnostic Rapid Speech Enhancement.
Luka ChkhetianiLevan BejanidzePublished in: CoRR (2020)
Keyphrases
- speech enhancement
- vocal tract
- noisy environments
- speaker verification
- speech signal
- noise reduction
- speech recognition
- speaker identification
- automatic speech recognition
- single channel
- signal to noise ratio
- speaker recognition
- speech synthesis
- noisy speech
- image acquisition
- smoothing algorithm
- linear prediction
- audio visual
- hidden markov models
- background noise
- multi channel
- acoustic features
- wiener filter
- mel frequency cepstral coefficients
- image processing
- spectral analysis
- speaker diarization
- super resolution
- language model