ASR2K: Speech Recognition for Around 2000 Languages without Audio.
Xinjian LiFlorian MetzeDavid R. MortensenAlan W. BlackShinji WatanabePublished in: INTERSPEECH (2022)
Keyphrases
- speech recognition
- speaker identification
- automatic speech recognition
- language identification
- speech processing
- broadcast news
- speech recognition technology
- speech signal
- hidden markov models
- noisy environments
- language independent
- speech synthesis
- language model
- acoustic features
- cepstral coefficients
- pattern recognition
- speech retrieval
- speaker recognition
- word error rate
- mel frequency cepstral coefficients
- speech recognizer
- multimedia
- conversational speech
- audio visual speech recognition
- handwriting recognition
- audio signal
- visual information
- image processing
- speech recognizers
- english text
- spontaneous speech
- speaker independent
- signal processing
- feature selection
- speaker verification
- multimedia information
- cross lingual
- speaker dependent
- isolated word