BlendGAN: Learning and Blending the Internal Distributions of Single Images by Spatial Image-Identity Conditioning
This work addresses the limitation of single-image generative models for multi-image tasks, offering a novel framework for image editing and synthesis.
The authors tackled the problem of training generative models on multiple images simultaneously, enabling applications like morphing and texture fusion that single-image models cannot support, by introducing BlendGAN which learns and blends internal distributions of several images using spatial image-identity conditioning.
Training a generative model on a single image has drawn significant attention in recent years. Single image generative methods are designed to learn the internal patch distribution of a single natural image at multiple scales. These models can be used for drawing diverse samples that semantically resemble the training image, as well as for solving many image editing and restoration tasks that involve that particular image. Here, we introduce an extended framework, which allows to simultaneously learn the internal distributions of several images, by using a single model with spatially varying image-identity conditioning. Our BlendGAN opens the door to applications that are not supported by single-image models, including morphing, melding, and structure-texture fusion between two or more arbitrary images.