1. Vector: Embedding, Latent Representation, Latent Code
image.png
2. Binary Classifier 评估 Encoder
image.png
image.png
3. Feature Disentangle 特征拆解
image.png
image.png
3.1 声音变声
image.png
image.png
image.png
3.2 IN & AdaIN
IN = Instance Normalization (remove global information)
AdaIN = Adaptive Instance Normalization (only influence global information)
image.png
4. Discrete Representation
image.png
Binary vector (参数较少，还可以识别没有见到的样本)
image.png
image.png
参考文献
Machine Learning (2019,Spring)
Voice Conversion
网友评论