Exploiting Cross-Domain Visual Feature Generation for Disordered Speech Recognition.
Shansong LiuXurong XieJianwei YuShoukang HuMengzhe GengRongfeng SuShi-Xiong ZhangXunying LiuHelen MengPublished in: INTERSPEECH (2020)
Keyphrases
- speech recognition
- cross domain
- feature generation
- text categorization
- hidden markov models
- transfer learning
- language model
- word sense disambiguation
- speech signal
- text classification
- pattern recognition
- knn
- information extraction
- knowledge transfer
- visual features
- document classification
- visual information
- k nearest neighbor
- low level
- feature selection
- semantic features
- neural network
- nearest neighbor
- multi modal
- principal component analysis
- artificial intelligence