Context-dependent output densities for hidden Markov models in speech recognition.
Georg StemmerViktor ZeißlerChristian HackerElmar NöthHeinrich NiemannPublished in: INTERSPEECH (2003)
Keyphrases
- speech recognition
- context dependent
- hidden markov models
- automatic speech recognition
- natural language
- speech processing
- speech signal
- low level
- noisy environments
- speech recognizer
- speech synthesis
- speaker independent
- speaker identification
- keyword spotting
- conditional random fields
- high level
- handwriting recognition
- viterbi algorithm
- speech recognition systems
- feature selection
- speaker dependent
- broadcast news
- discriminative training
- language model
- face recognition