Disentangling the Spatial Structure and Style in Conditional VAE
This work addresses the challenge of interpretable and controllable generation in conditional generative models, but it is incremental as it builds on existing methods like SPADE and AdaIN.
The paper tackles the problem of disentangling latent space in conditional VAEs into spatial structure and style codes, with one being label-relevant and the other irrelevant, and demonstrates effectiveness through experiments on two datasets.
This paper aims to disentangle the latent space in cVAE into the spatial structure and the style code, which are complementary to each other, with one of them $z_s$ being label relevant and the other $z_u$ irrelevant. The generator is built by a connected encoder-decoder and a label condition mapping network. Depending on whether the label is related with the spatial structure, the output $z_s$ from the condition mapping network is used either as a style code or a spatial structure code. The encoder provides the label irrelevant posterior from which $z_u$ is sampled. The decoder employs $z_s$ and $z_u$ in each layer by adaptive normalization like SPADE or AdaIN. Extensive experiments on two datasets with different types of labels show the effectiveness of our method.