An End-to-End Deep Learning Approach to Simultaneous Speech Dereverberation and Acoustic Modeling for Robust Speech Recognition.
Bo WuKehuang LiFengpei GeZhen HuangMinglei YangSabato Marco SiniscalchiChin-Hui LeePublished in: IEEE J. Sel. Top. Signal Process. (2017)
Keyphrases
- speech recognition
- end to end
- deep learning
- speech recognition systems
- speaker independent
- noisy environments
- speech recognizers
- speech signal
- acoustic models
- noisy speech
- speech recognizer
- speech synthesis
- automatic speech recognition
- speech processing
- hidden markov models
- language model
- pattern recognition
- speaker dependent
- mel frequency cepstral coefficients
- machine learning
- acoustic features
- isolated word
- speaker recognition
- speech recognition errors
- unsupervised learning
- supervised learning
- data mining
- cepstral coefficients
- neural network
- multiscale