Diversifying Semantic Image Synthesis and Editing via Class- and Layer-wise VAEs
This work addresses the need for richer multimodal image synthesis and editing, offering incremental improvements over existing methods by handling individual object appearance factors.
The paper tackled the problem of limited diversity in semantic image synthesis by proposing a class- and layer-wise VAE extension to learn multiple latent spaces for flexible control over object styles, resulting in more plausible and diverse images compared to state-of-the-art methods across three domains.
Semantic image synthesis is a process for generating photorealistic images from a single semantic mask. To enrich the diversity of multimodal image synthesis, previous methods have controlled the global appearance of an output image by learning a single latent space. However, a single latent code is often insufficient for capturing various object styles because object appearance depends on multiple factors. To handle individual factors that determine object styles, we propose a class- and layer-wise extension to the variational autoencoder (VAE) framework that allows flexible control over each object class at the local to global levels by learning multiple latent spaces. Furthermore, we demonstrate that our method generates images that are both plausible and more diverse compared to state-of-the-art methods via extensive experiments with real and synthetic datasets inthree different domains. We also show that our method enables a wide range of applications in image synthesis and editing tasks.