Fast algorithm for speech recognition using speaker cluster HMM.
Masayuki YamadaYasuhiro KomoriTetsuo KosakaHiroki YamamotoPublished in: EUROSPEECH (1997)
Keyphrases
- speech recognition
- learning algorithm
- hidden markov models
- cost function
- matching algorithm
- optimization algorithm
- computational cost
- forward backward
- segmentation algorithm
- worst case
- np hard
- optimal solution
- significant improvement
- k means
- dynamic programming
- search space
- detection algorithm
- objective function
- computational complexity
- speaker adaptation
- automatic speech recognition
- recognition algorithm
- cluster centers
- data clustering
- expectation maximization
- probabilistic model
- feature vectors
- preprocessing
- clustering algorithm