Noise-Robust Voice Conversion by Conditional Denoising Training Using Latent Variables of Recording Quality and Environment.
Takuto IgarashiYuki SaitoKentaro SekiShinnosuke TakamichiRyuichi YamamotoKentaro TachibanaHiroshi SaruwatariPublished in: CoRR (2024)
Keyphrases
- latent variables
- denoising
- structural svms
- noisy environments
- structured prediction
- probabilistic model
- conditionally independent
- gaussian noise
- denoising methods
- noisy images
- latent variable models
- posterior distribution
- denoising algorithm
- approximate inference
- random variables
- noise level
- prior knowledge
- variational bayes
- image denoising
- hidden variables
- probabilistic principal component analysis
- total variation
- structural svm
- elastic net
- gaussian process
- max margin
- generative model
- knowledge discovery
- training set
- latent space
- information retrieval
- signal to noise ratio
- information extraction