Multi-resolution stacking for speech separation based on boosted DNN.
Xiao-Lei ZhangDeLiang WangPublished in: INTERSPEECH (2015)
Keyphrases
- multiresolution
- speech recognition
- coarse to fine
- speech signal
- wavelet transform
- spoken language
- hierarchical representation
- endpoint detection
- audio visual
- subband
- training process
- wavelet coefficients
- speech synthesis
- ensemble learning
- random forests
- combining multiple
- automatic speech recognition
- dialogue system
- wavelet domain
- audio stream
- broadcast news
- data sets
- recognition engine
- spontaneous speech
- vocal tract
- multi modal
- face detection