Speaker Adaptation Using Spectro-Temporal Deep Features for Dysarthric and Elderly Speech Recognition.
Mengzhe GengXurong XieZi YeTianzi WangGuinan LiShujie HuXunying LiuHelen MengPublished in: IEEE ACM Trans. Audio Speech Lang. Process. (2022)
Keyphrases
- speech recognition
- speaker adaptation
- speech recognition systems
- speech recognizer
- speaker independent
- speaker dependent
- hidden markov models
- speech synthesis
- language model
- automatic speech recognition
- pattern recognition
- speech signal
- feature vectors
- maximum likelihood
- feature space
- noisy environments
- speaker identification
- neural network
- mel frequency cepstral coefficients
- low level
- feature extraction