Mapping Mel sub-band energies using Deep belief network for robust speech recognition.
Mojtaba GholamipourBabak NasersharifPublished in: IST (2016)
Keyphrases
- speech recognition
- subband
- cepstral coefficients
- noisy environments
- hidden markov models
- language model
- discrete wavelet transform
- wavelet transform
- speech signal
- wavelet coefficients
- speech recognition systems
- image compression
- pattern recognition
- multiresolution
- high frequency
- frequency domain
- filter bank
- wavelet decomposition
- wavelet domain
- face recognition
- bit rate
- graphical models
- frequency band
- speaker identification
- feature vectors
- energy function
- graph cuts
- probabilistic model
- linear prediction
- deep belief networks
- machine learning