VQMIVC: Vector Quantization and Mutual Information-Based Unsupervised Speech Representation Disentanglement for One-shot Voice Conversion.
Disong WangLiqun DengYu Ting YeungXiao ChenXunying LiuHelen MengPublished in: CoRR (2021)
Keyphrases
- entropy constrained
- vector quantization
- vector quantizer
- speaker recognition
- image compression
- successive approximation
- input vector
- text to speech
- image registration
- codebook design
- image processing
- distortion measure
- mutual information
- supervised learning
- fractal image compression
- fractal image coding
- neural gas
- emotion recognition
- reduced complexity
- speech recognition
- image representation
- unsupervised learning
- semi supervised
- multiscale