Distribution-Preserving Steganography Based on Text-to-Speech Generative Models.
Kejiang ChenHang ZhouHanqing ZhaoDongdong ChenWeiming ZhangNenghai YuPublished in: IEEE Trans. Dependable Secur. Comput. (2022)
Keyphrases
- generative model
- text to speech
- probabilistic model
- mixture model
- speech synthesis
- discriminative learning
- hierarchical hidden markov models
- discriminative models
- prior knowledge
- conditional random fields
- prosodic features
- topic models
- em algorithm
- word processing
- text to speech synthesis
- random variables
- probability distribution
- object categories
- hidden variables
- semi supervised
- maximum entropy principle
- discriminative methods
- generative and discriminative models