Improved automatic speech recognition system using sparse decomposition by basis pursuit with deep rectifier neural networks and compressed sensing recomposition of speech signals.
Mihai GavrilescuPublished in: COMM (2014)
Keyphrases
- compressed sensing
- basis pursuit
- speech signal
- neural network
- speech recognition
- image reconstruction
- automatic speech recognition systems
- automatic speech recognition
- vocal tract
- random projections
- additive noise
- sparse representation
- speech quality
- natural images
- non stationary
- pattern recognition
- hidden markov models
- speech sounds
- denoising
- speech enhancement
- machine learning
- fourier domain
- convex optimization
- signal processing
- multiscale
- image processing
- language model
- higher order