Mixture factorized auto-encoder for unsupervised hierarchical deep factorization of speech signal.
Zhiyuan PengSiyuan FengTan LeePublished in: CoRR (2019)
Keyphrases
- noisy images
- speech signal
- matrix factorization
- edge detection
- unsupervised learning
- automatic speech recognition
- vocal tract
- mixture model
- background noise
- spectral analysis
- linear prediction
- automatic speech recognition systems
- deep belief networks
- speech recognition
- speaker identification
- noisy environments
- measurement matrix
- fundamental frequency
- blind source separation
- speaker recognition
- speech enhancement
- machine learning
- acoustic features
- gaussian mixture model
- visual features
- probabilistic model