Exploiting Cross Domain Acoustic-to-articulatory Inverted Features For Disordered Speech Recognition.
Shujie HuShansong LiuXurong XieMengzhe GengTianzi WangShoukang HuMingyu CuiXunying LiuHelen MengPublished in: CoRR (2022)
Keyphrases
- speech recognition
- speech recognition systems
- cross domain
- speech signal
- language model
- speech synthesis
- speaker independent
- mel frequency cepstral coefficients
- speech recognizer
- hidden markov models
- speech recognizers
- speaker identification
- automatic speech recognition
- noisy environments
- pattern recognition
- acoustic models
- knowledge transfer
- transfer learning
- sentiment classification
- noisy speech
- feature extraction
- text categorization
- feature set
- feature vectors
- training data