Comparison of Tied-Mixture and State-Clustered HMMs with Respect to Recognition Performance and Training Method.
Hiroyuki SegiKazuo OnoeShoei SatoAkio KobayashiAkio AndoPublished in: J. Inf. Technol. Res. (2014)
Keyphrases
- significant improvement
- training phase
- computational cost
- training process
- high accuracy
- data sets
- training procedure
- reinforcement learning
- pattern recognition
- handwritten digits
- pairwise
- cost function
- learning stage
- detection method
- clustering method
- mixture model
- facial expressions
- dynamic programming
- preprocessing
- similarity measure
- feature extraction