Training HMMs using a minimum recognition error approach.
Horacio FrancoAntónio Joaquim SerralheiroPublished in: ICASSP (1991)
Keyphrases
- chinese named entity recognition
- low error rates
- hidden markov models
- recognition rate
- recognition accuracy
- object recognition
- handwritten text recognition
- error rate
- training process
- discriminative training
- training error
- error bounds
- recognition algorithm
- training algorithm
- face recognition
- speech recognition
- test set
- action recognition
- visual recognition
- automatic recognition
- recognition process
- activity recognition
- pattern recognition
- training samples
- training set
- multi stream
- feature extraction
- baum welch
- supervised learning
- learning stage
- training procedure
- error detection