A Maximum Entropy Based Hierarchical Model for Automatic Prosodic Boundary Labeling in Mandarin.
Fangzhou LiuHuibin JiaJianhua TaoPublished in: ISCSLP (2008)
Keyphrases
- hierarchical model
- latent variables
- hierarchical models
- speech recognition
- prosodic features
- human body
- image segmentation
- object boundaries
- semi automatic
- data mining
- information theory
- text to speech synthesis
- fully automatic
- input image
- hidden markov models
- prior knowledge
- natural language
- training data
- learning algorithm
- machine learning