Learning Disentangled Prompts for Compositional Image Synthesis
This addresses the problem of efficient style adaptation in image generation for AI and creative applications, but it appears incremental as it builds on existing prompt tuning methods.
The paper tackles domain-adaptive image synthesis by teaching pretrained generative models new styles from as few as one image, using a framework with disentangled prompts for semantic and domain information to synthesize images in target styles.
We study domain-adaptive image synthesis, the problem of teaching pretrained image generative models a new style or concept from as few as one image to synthesize novel images, to better understand the compositional image synthesis. We present a framework that leverages a pretrained class-conditional generation model and visual prompt tuning. Specifically, we propose a novel source class distilled visual prompt that learns disentangled prompts of semantic (e.g., class) and domain (e.g., style) from a few images. Learned domain prompt is then used to synthesize images of any classes in the style of target domain. We conduct studies on various target domains with the number of images ranging from one to a few to many, and show qualitative results which show the compositional generalization of our method. Moreover, we show that our method can help improve zero-shot domain adaptation classification accuracy.