Are audio or textual training data more important for ASR in less-represented languages?
Thomas PellegriniLori LamelPublished in: SLTU (2008)
Keyphrases
- training data
- multimedia
- expressive power
- automatic speech recognition
- decision trees
- test data
- visual information
- prior knowledge
- signal processing
- speech recognition
- language identification
- classification models
- test set
- broadcast news
- manually constructed
- text summarization
- keywords
- textual information
- spontaneous speech
- noisy environments
- audio video
- learning algorithm
- language independent
- audio visual
- visual data
- training process
- cross lingual
- training samples
- classification accuracy
- training set
- natural language
- feature selection