Speech dereverberation for enhancement and recognition using dynamic features constrained deep neural networks and feature adaptation.
Xiong XiaoShengkui ZhaoDuc Hoang Ha NguyenXionghu ZhongDouglas L. JonesEng Siong ChngHaizhou LiPublished in: EURASIP J. Adv. Signal Process. (2016)
Keyphrases
- neural network
- speech recognition
- pattern recognition
- artificial neural networks
- fuzzy logic
- image processing
- spoken language
- image enhancement
- cepstral features
- speech signal
- neural nets
- back propagation
- genetic algorithm
- image features
- multi layer
- deep learning
- fault diagnosis
- sufficient conditions
- feed forward
- fuzzy systems
- audio visual
- network architecture
- hidden markov models
- emotion recognition
- similarity measure
- speech synthesis