Minimum classification error training of hidden Markov models for handwriting recognition.
Alain BiemPublished in: ICASSP (2001)
Keyphrases
- minimum classification error
- handwriting recognition
- hidden markov models
- speech recognition
- discriminative training
- statistical pattern recognition
- classification error
- character recognition
- maximum likelihood estimation
- handwritten text
- word recognition
- handwritten words
- gaussian mixture
- word segmentation
- machine learning
- automatic speech recognition
- noisy environments
- speech signal
- posterior probability
- parameter estimation
- training set