Spoken Web Search using an Ergodic Hidden Markov Model of Speech.
Asif AliMark ClementsPublished in: MediaEval (2013)
Keyphrases
- hidden markov models
- speech recognition
- web search
- automatic speech recognition
- speech signal
- speech synthesis
- speech recognizer
- search engine
- phoneme recognition
- markov chain
- broadcast news
- speaker independent
- multi stream
- markov models
- speech recognition systems
- spoken language
- markov model
- viterbi algorithm
- noisy speech
- conditional random fields
- speaker dependent
- web pages
- hidden state
- speaker identification
- gesture recognition
- handwriting recognition
- discriminative training
- vocal tract
- handwritten word recognition
- noisy environments
- sequential data
- steady state
- acoustic features
- speaker recognition
- speaker diarization
- recognition engine
- dynamic probabilistic networks